Conference Digest

2021 IEEE International Conference on Mechatronics and Automation

IEEE ICMA 2021

Takamatsu, Japan

August 8-10, 2021

Cosponsored by

IEEE Robotics and Automation Society Kagawa University, Japan Beijing Institute of Technology, China

Technically cosponsored by

The Robotics Society of Japan The Japan Society of Mechanical Engineers Japan Society for Precision Engineering The Society of Instrument and Control Engineers National Natural Science Foundation of China, China Chinese Mechanical Engineering Society Chinese Association of Automation State Key Laboratory of Robotics and System (HIT) The Institute of Advanced Biomedical Engineering System, BIT Key Laboratory of Convergence Medical Engineering System and Healthcare Technology, the Ministry of Industry and Information Technology, Beijing Institute of Technology, China Guangxi University of Science and Technology Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems; Tianjin Key Laboratory for Advanced Mechatronics System Design and Intelligent Control; Tianjin University of Technology, China Tianjin International Joint Research and Development Center, Tianjin University of Technology, China Harbin Engineering University University of Electro-Communications University of Electronic Science and Technology of China Optics and Precision Engineering

IEEE ICMA 2021 PROCEEDINGS

Additional copies may be ordered from:

IEEE Service Center 445 Hoes Lane Piscataway, NJ 08854 U.S.A.

IEEE Catalog Number: CFP21839-PRT ISBN 978-1-6654-4100-1 IEEE Catalog Number (CD-ROM): CFP21839-CDR ISBN (CD-ROM): 978-1-6654-4098-1

Copyright and Reprint Permission: Abstracting is permitted with credit to the source. Libraries are permitted to photocopy beyond the limit of U.S. copyright law for private use of patrons those articles in this volume that carry a code at the bottom of the first page, provided the per-copy fee indicated in the code is paid through Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923. For reprint or republication permission, email to IEEE Copyrights Manager at [email protected]. All rights reserved. Copyright ©2021 by IEEE.

The papers in this book comprise the proceedings of the meeting mentioned on the cover and title page. They reflect the author’s opinions and, in the interest of timely dissemination, are published as presented and without change. Their inclusion in this publication does not necessarily constitute endorsement by the editors or the Institute of Electrical and Electronics Engineers, Inc.

The Institute of Electrical and Electronics Engineers, Inc.

Foreword

On behalf of the IEEE ICMA 2021 Conference Organizing Committee, it is our great pleasure, an honor, and a privilege to welcome you to Takamatsu for the 2021 IEEE International Conference on Mechatronics and Automation. This conference reflects the growing interests in the broad research areas of mechatronics, robotics, sensors and automation.

IEEE ICMA 2021 marks the 18th edition of the IEEE ICMA annual conference series. We are proud to announce that a high number of 390 papers were submitted from 28 countries and regions, including 376 contributed papers, 14 papers for organized sessions, and 253 papers were accepted for oral presentation at the conference after a rigorous full-paper review process, achieving an acceptance rate of less than 65%. Presentations at IEEE ICMA 2021 are organized in 6 parallel tracks, for a total of 40 sessions, taking place during the two and half conference days. We are fortunate to be able to invite four distinguished speakers to deliver plenary talks and keynote speech.

We are very glad that you are joining us at IEEE ICMA 2021 in Takamatsu to live this unique experience. The main objective of IEEE ICMA 2021 is to provide a forum for researchers, educators, engineers, and government officials involved in the general areas of mechatronics, robotics, sensors and automation to disseminate their latest research results and exchange views on the future research directions of the related fields. IEEE ICMA 2021 promises to be a great experience for participants from all over the world, with an excellent technical program as well as social activities.

We would like to express our most sincere appreciation and thanks to all our sponsoring societies and organizations and to all the individuals who have contributed to the organization of this conference. Our special thanks are extended to our colleagues in the Program Committee for their thorough review of all the submitted papers, which is vital to the success of this conference. We must also extend our thanks to our Organizing Committee and our volunteers who have dedicated their time toward ensuring the success of this conference. Last but not least, we thank all the contributors for their support and participation in making this conference a great success. Finally, we wish you a great conference and enjoyable stay in Takamatsu, Japan.

Shuxiang Guo Keisuke Morishima Qinxue Pan Hideyuki Hirata General Chair Program Chair Program Chair Organizing Chair

Welcome Remarks

Welcome to the 2021 IEEE International Conference on Mechatronics and Automation (IEEE ICMA 2021). As one of the Advisory Council Chairs, it is my great pleasure and honor to welcome you to what promises to be its most successful conference to date.

Kagawa University delighted to be one of the hosts of the Conference which marked as the 18th edition of the IEEE ICMA among the annual conference series. The Conference reflects the growing interests in the broad research areas of mechatronics, robotics sensors and automation.

IEEE ICMA 2021 is dedicated to provide a forum for researchers, educators, engineers and government officials involved in these areas to disseminate the latest research results and exchange ideas on the future research directions.

As the founding university of the IEEE ICMA conference, Kagawa University is a research-oriented university, with science, engineering and research as its core. We have established our own unique programs related to the field of mechatronics and automation that are well known in Japan.

Kagawa University appreciates all of the sponsoring societies, organizations and all the individuals contributed to the organization of the conference, as well as all the authors, sessions organizers，plenary speakers, exhibitors for their interests and contribution to make IEEE ICMA 2021 a successful and fruitful event.

Wish IEEE ICMA 2021 conference a complete success!

Yoshiyuki Kakehi, Medical Doctor President of Kagawa University, Japan

IEEE ICMA 2021 Conference Digest Table of Content

Foreword Welcomg Remarks

IEEE ICMA 2021 Conference ...... i Organizing Committees ...... i International Program Committee ...... iv IEEE ICMA 2021 Conference Cosponsors ...... xi General Information ...... xii Conference Information ...... xviii Conference Registration ...... xx Social Events ...... xx Technical Program Keynote Speech and Plenary Talks ...... xxi Program at a Glance ...... xxix Technical Session Schedule ...... xxx Technical Sessions Monday, 9 August 2021 ...... 1 Tuesday, 10 August 2021 ...... 25 Index of Session chairs ...... 46 Index of Authors ...... 48 IEEE ICMA 2022 CFP & IJMA Journal CFP

IEEE ICMA 2021 Conference

International Scientific Advisory Board

Advisory Council Honorary Chairs Toshio Fukuda Meijo University, Japan T. J. Tarn Washington University, USA Advisory Council Chairs Yoshihiro Suenaga Kagawa University, Japan Tianyou Chai Northeastern University, China Kazuhiro Kosuge Tohoku University, Japan A.A. Goldenberg University of Toronto, Canada Paolo Dario Scuola Superiore Sant'Anna, Italy Masayoshi Tomizuka UC Berkeley, USA Mario A. Rotea University of Massachusetts, USA Ju-Jang Lee KAIST, Korea Ren C. Luo National Taiwan University, Taiwan Yu Yao Harbin Engineering University, China Yanrong Li UESTC, China Huadong Yu CUST, China Simin Li Guangxi Univ. of Science and Technology, China Qingxin Yang Tianjin University of Technology, China

Organizing Committees

General Chairs

Shuxiang Guo Kagawa University, Japan

General Co-Chairs

William R. Hamel University of Tennessee, USA

Tatsuo Arai Osaka University, Japan Darwin G. Caldwell Italian Institute of Technology, Italy Tim Lueth TUM, Germany James K. Mills Univ. of Toronto, Canada Enzeng Dong Tianjin Univ. of Technology, China Yuxin Zhao Harbin Engineering University, China Xueshan Gao Beijing Institute of Technology, China Masao Arakawa Kagawa University, Japan

Organizing Committee Chairs

Hideyuki Hirata Kagawa University, Japan Aiguo Ming U. of Electro-Communications,, Japan

Tutorial/Workshop Chairs

Guangjun Liu Ryerson University, Canada Keisuke Suzuki Kagawa University, Japan Chaoyang Shi Tianjin University Qiquan Quan Harbin Institute of Technology, China

Invited/Organized Session Chairs

Kotaro Kimura Osaka University, Japan Kazuhito Yokoi AIST, Japan Paul Wen Univ. of Southern Queensland, Australia Hideyuki Sawada Sawada University, Japan Dapeng Tian CIOMP, CAS, China Yang Shi University of Victoria, Canada Shuoyu Wang Kouchi Institute of Technology, Japan Shoichi Maeyama Kagawa University, Japan

Awards Committee Co-Chairs

James K. Mills University of Toronto, Canada Xinkai Chen Shibaura Institute of Technology, Japan Liwei Shi Beijing Institute of Technology, China

Publications Chairs

Jin Guo Beijing Institue of Technology, China

Publicity Chairs

Jinjun Shan York University, Canada Yangmin Li Hong Kong Polytechnic Univ., China Xiufen Ye Harbin Engineering University, China

Finance Chair

Jian Guo Tianjin University of Technology, China

Local Arrangement Chairs

Hidenori Ishihara Kagawa University, Japan Koh Inoue Kagawa University, Japan

Conference Secretariats

X. Jin, P. Shi, R. An, Z. Yang Kagawa University, Japan

Conference Web System Administrator

Ziyi Yang Kagawa University, Japan

Program Committees

Program Chair

Keisuke Morishima Osaka University, Japan Qinxue Pan Beijing Institute of Technology, China

Program Co-Chairs

Kevin Lynch Northwestern University, USA

Cecilia Laschi Scuola Superiore Sant’Anna, Italy

Shingo Kagami Tohoku University, Japan Hong Zhang University of Alberta, Canada Hidekuni Takao Kagawa University, Japan Yili Fu Harbin Institute of Technology, China iii

Atsushi Yamashita The University of Tokyo, Japan Stefan Byttner Halmstad University, Sweden Huosheng Hu University of Essex, U.K Hesheng Wang Shanghai Jiaotong University, China Jian Li Guangxi Univ. of Science and Tech., China Anqi Qiu National Univer. of Singapore, Singapore

International Program Committee

Adachi, Fumiyuki Addie, Ron Aiyama, Yasumichi Althoefer, Kaspar An, Ruochen Ando, Kazuaki Ang, Wei Tech Aoyama, Hisayuki Arai, Fumihito Arai, Tamio Arai, Tatsuo Araki, Kiyomichi Arzanpour, Siamak Asada, Minoru Asama, Hajime Asano, Toshio Baeg, Sang Hyeon Bai, Ou Bao, Xianqiang Ben, Ridha Ben-Tzvi, Pinhas Bi, Shusheng Bi, Zhuming Bian, Hongyu Bidaud, Philippe Bolon, Ph. Bolon, Philippe Boustany, Nada Brown, Martin Bu, Dongdong Byung, Min Cai, Lilong Caldwell, Darwin G. Cao, Maoyong Ceccarelli, Marco Chen, Chun-Ta Chen, Deyun Chen, Feng Chen, Guanlong Chen, I-Ming Chen, Ken Chen, Liguo Chen, Ting Chen, Weidong Chen, Weihai Chen, Wenhua Chen, Xin Chen, Xinkai Chen, Yangquan Chen, Zhangwei Cheng, Allen Cheng, Jianhua Cheng, Ji-Xin Cheol, Min Cheong, Joono Cho, Young-Jo Choi, Hyouk Ryeol Choi, Hyun-Taek Choi, Junho Choi, Youngjin Chou, Wusheng Chu, Jinkui Chugo, Daisuke Chui, Dehua Chung, Woojin Cui, Jianwen Dai, Jiansheng Dai, Xuefeng Dailey, Matthew Damaren, Chris Dario, Paolo Davis, Steve Dechev, Nick Deco, Gustavo Deguang, Shang Demiris, Yiannis Deng, Yulin Dillmann, Ruediger Ding, Jiexiong Ding, Xilun Do, Hyun Min Dodd, Tony Doh, Nakju Dohta, Shujiro Doi, Shunichi Dong, Enzeng Dong, Hongbin Dong, Zaili Du, Mingxing Du, Pingan Du, Xiliang Du, Zhijiang Duan, Haibin Dubay, Rickey Dufosse, Michel Fan, Jinwei

Fang, Hao Fang, Yongchun Fei, Qin Feng, Gary Feng, Weixing Fiorini, Paolo Foulloy, Laurent Fu, Bin Fu, Jiacai Fu, Mingyu Fu, Qiang Fu, Sheng Fu, Shuigen Fu, Yili FUH, Ying-His Fujino, Tadashi Fujisaki, Kiyotaka Fujisawa, Shoichiro Fujiwara, Takayuki Fujiwara, Yoshitaka Fukuda, Toshio Fukuyama, Hidenao Fung, Wai-Keung Furusho, Junji Furuta, Katsuhisa Furutani, Eiko Furutani, Katsushi Furuya, Nobuyuki Gang, Tong Gao, baofeng Gao, Guohua Gao, Hongtao Gao, Lin Gao, Shesheng Gao, Wei Gao, Xueshan Ge, Sam Ge, Sam Shuzhi Ge, Weimin Ge, Yunjian Gitchang Gofuku, Akio Gong, Haixia Gong, Yadong Gong, Zhiguo Graefe, Volker Gu, Dongbing Gu, Jason Gu, Xingzhong Guan, Yisheng Guang, Zu Guang Guglielmelli, Eugenio Guo, Jian Guo, Jin Guo, Maozu Guo, Mingliang Guo, Peng Guo, Shuxiang Guo, Yi Habibi, Saeid Haga, Yoichi Hamaguchi, Tetsuya Hamel, William R. Han, Hongbin Han, Jiqing Han, Min Han, Yujie Hane, Kazuhiro Hao, Gang Harada, Kensuke Hariri, Alireza Hasegawa, Osamu Hasegawa, Tadahiro Hasegawa, Yasuhisa Hashimoto, Hiroshi Hashimoto, Koichi Hashimoto, Minoru Hashimoto, Shuji Hata, Seiji Hattori, Tetsuo Hayashi, Jun-ichiro He, Cunfu He, Jia He, Jiping Hino, Junichi Hirai, Shigeoki Hirata, Hideyuki Hirata, Koichi Hirata, Yasuhisa Hirose, Akira Hong, Keum-Shik Hong, Seung Ho Hori, Toshio Hosoda, Kou Hou, Mingshan Hou, Shuping Hou, Xihuan Hou, Zhenguang Hu, Chao Hu, Huosheng Hu, Jiquan Hu, Jun Hu, Jwu-Sheng Hua, Jun Huang, Dagui Huang, Guo-Shing Huang, Hongzhong Huang, Ping Huang, Qiang Huang, Qingjiu Huang, Tian Huang, Yalou Huasong, Min Hwan, Chang Hwan, Dong IBA, Hitoshi Ichikawa, Akihiko Ichiki, Masatoshi Iguchi, Tetsuhiro Ikehara, Masaaki Ikeuchi, Masashi Ikuta, Koji Inomo, Hitoshi Ise, Toshifumi ishiguro, Hiroshi Ishihara, Hidenori Ishihara, Sunao Ishii, Akira Ishii, Chiharu Ishii, Kazuo Ishii, Koji Ishikawa, Junzo

Ito, Tomotaka Iwamura, Yoshiaki Iwatsuki, Nobuyuki Izuishi, Kunihiko Janabi-Shari, Farrokh Ji, Ping Ji, Yuehui Jia, Kebin Jia, Songmin Jia, Zhengyuan Jiang, Pingyu Jiang, Zhen Jiao, Shuhong Jin, Hongzhang Jin, Rencheng Jing, Wuxing Jo, Yongho Jr., Marcelo H.Ang, Jung, Kwangmok Jung, Seul K.H.Pang, Grantham Kagami, Shingo Kagawa, Koji Kamata, Minoru Kamiya, Yoshitsugu Kanamori, Chisato Kaneko, Shunichi Kang, Sung Chul Kang, Taehun Karaki, Masayuki Kato, Takahisa Khajepour, Amir Kiguchi, Kazuo Kim, Byeongho Kim, Doik Kim, Jinhyun Kim, Jung Kim, Keehoon Kim, Sungshin Kim, Wheekuk Kimura, Kotaro Kitajima, Hiroyuki Kitajima, Katsuhiro Kobayashi, Tetsuo Kobayashi, Toshihiro Koh, Kyungchul Kojima, Masaru Kok, Tan Komeda, Takashi Kometani, Reo Kong, Xiawen Konyo, Masashi Koo, Ja Choon Koshimizu, Hiroyasu Kosuge, Kazuhiro Kotiw, Mike Kotosaka, Shinya Kouzani, Abbas Koyanagi, Kenichi Koyanagi, Mitsumasa Kubota, Naoyuki Kubota, Takashi Kulkarni，M. S. Kuno, Yoshinori Kurazuma, Ryo Kurisu, Masamitsu Kuroki, Nobutaka Kuwakado, Hidenori Kuwano, Hiroki Kyriakopou, Kostas J. Lai, Yongjun Lan, Hai Laschi, Cecilia Laugier, Christian Lee, Jang Myung Lee, Jihong Lee, Songjun Lee, Yikuen Li, Bin Li, Bing Li, Chunwen Li, Desheng Li, Gang Li, Haisen Li, Hongsheng Li, Howard Li, Jianfeng Li, Jiangang Li, Jin Li, Li Li, Mantian Li, Maoxun Li, Qin Li, Wenfeng Li, Xiaoshan Li, Xiaoyang Li, Xiukun Li, Yangmin Li, Youfu Li, Yuhua Li, Yun Li, Yunhua Li, Zhang Li, Zhiyi Li, Zhiyong Li, Zhongjian Lian, Feng-Li Liang, Guolong Liang, Yan Liao, Hongen Liao, Wei-Hsin Lin, Chyi-Yeu Lin, Hsiung-Cheng Lin, Ming-Tzong Liu, Bo Liu, Da Liu, Dikai Liu, Fang Liu, Fanming Liu, Guangjun Liu, Guangyu Liu, Haibo Liu, Honghai Liu, Hugh Liu, Jiming Liu, Jindong Liu, Jinguo Liu, Jun Liu, Liqiang Liu, Ming Liu, Peter X. Liu, Qing

Liu, Rong Liu, Xiangdong Liu, Xiaoping Liu, Xin-Jun Liu, Yongguang Liu, Yugang Liu, Yunhui Liu, Zhen Long, Guilu Lottin, Jacques Lou, Yunjiang Loureiro, Rui Low, Kin-Huat Lu, BaoLiang Lu, Shengfu Lu, Tien-Fu M.Chen, Ben M.Gupta, Madan Ma, Chunguang Ma, Shugen Ma, Xu Ma, Xudong Mae, Yasushi Magnenat-Thalm, Nadia Mao, Jianqin Maruyama, Hisataka Masaki, Yamakita Masek, Vlastimil Mashatan, Vahid Mashec, Vlasitimi Masuda, Tadashi Matsuhisa, Hiroshi Matsunaga, Saburo Matsuno, Fumitoshi Matsuno, Takayuki Matsushige, Kazumi Maxwell, Andrew Melchiorri, Claudio Meng, Max Q.-H. Meng, Yan Mills, James K. Mills, James K. Minami, Hirotsugu Minato, Kotaro Minemura, Kiyoshi Ming, Aiguo Mitsuishi, Mamoru Miyanaga, Yoshikazu Miyauchi, Satoru Mo, Hongwei Mo, Shuhua Mochiyama, Hiromi Morii, Masakatsu Morikawa, Hiroyuki Morishige, Koichi Morishima, Keisuke Morita, Noboru Morita, Yoshifumi Murakami, Toshiyuki Muscato, Giovanni Nagata, Fusaomi Nagatani, Keiji Nagato, Keisuke Nakajima, Masahiro Nakamura, Akio Nakamura, Hikaru Nakao, Masayuki Nakatani, Akihiro Nakauchi, Yasushi Nanayakkara, Thrish Nefti-Meziani, Samia Nelson, Bradley J. Ni, Jinping Ni, Zhonghua Ning, Hui Nohmi, Masahiro Nojima, Toshio Oana, Hidehiro Obara, Minoru Ogose, Shigeaki Ohara, Kenichi Ohsawa, Yasuharu Ohtake, Hiroshi Oka, Koichi Okada, Eiji Oki, Eiji Okuma, Masaaki Omichi, Takeo Oohira, Fumikazu Osumi, Hisashi Otake, Mihoko Otsuka, Akimasa Ouezdou, Fathi Ben Ouyang, Puren P.Miller, David Pan, Yajun Pang, Muye Park, Jong Hyeon Park, Jooyoung Park, Sangdok Payande, Sharam Perez, Ruben Pobil, Angel P. del Prassler, Erwin Qi, Guangyun Qi, Hairong Qi, Naiming Qiao, Gang Qiao, Hong Qiu, Anqi Qiu, Hua Radermacher, Klaus Rao, Wenbi Ren, Carolyn Ren, Wei Ren, Xiangshi Rhim, Sungsoo Ridha Ben Mrad Roh, Segon Rong, Weibin Ru, Changhai Ryeol, Dong Ryu, Jee-Hwan Sabatier, Jocelyn Sabti, Ali Saito, Takashi Sakaguchi, Masamichi Sakai, Shuichi Sakka, Sophie Sakurai, Yoshio Salman, Shaaban Ali Sampei, Seiichi

vii

Sandini, Giulio Sawada, Hideyuki Sekiyama, Kousuke Semini, Claudio Shaaban Ali Salman Shan, Jinjun Shao, Keyong Shao, Xinjian Shen, Yantao Shen, Yueshi Sheng, Jie Sheng, Weihua Shi, Chaoyang Shi, Guangfan Shi, Haizhang Shi, Jichuan Shi, Liwei Shi, Zhen Shibata, Takanori Shimojo, Makoto Shimotsu, Masateru Shoham, Moshe Shriaki, Wataru Singh, Akash Soar, Jeffrey Son, Jaebum Song, Jae-Bok Song, Kai-Tai Song, Quanjun Song, Yu Song, Zhibin Stasse, Olivier Stein, Cathryne Su, Chanmin Q. Su, Chanmin Q. Su, Chun-yi Su, Liying Sugar, Tom Sugita, Naohiko Suh, Il Hong Sun, Baoyuan Sun, Daqian Sun, Dong Sun, jinwei Sun, Kangning Sun, Xiaojun Sun, Yong Sun, Yu Sun, Zhaowei Suzuki, Keisuke Suzuki, Minoru Suzuki, Takahiro Suzuki, Yuji Tadakuma, Kenjiro Takahashi, Ryoichi Takahashi, Satoru Takahashi, Satoshi Takahashi, Tatsuro Takaiwa, Masahiro Takamasu, Kiyoshi Takanobu, Hideaki Takasaki, Masaya Takeda, Takashi Takeda, Yukio Takesue, Naoyuki Takubo, Tomohito Tan, Jeffrey Tan, Jindong Tan, Lihai Tan, Min Tan, Zhenfan Tanaka, Mami Tanaka, Takayuki Tang, Mo Tang, Yike Tanikawa, Tamio Tanji, Yuichi Tao, Nongjian Tarumi, Hiroyuki Terada, Hidetsugu Tian, Yantao Tomita, Naohide Tonet, Oliver Torii, Akihio Touge, Tetsuo Tsagarakis, Nikos Tsai, Ching-Chih Tsuji, Toshio Tsukada, Toshihiko Tsukamoto, Hiroshi Tsumaki, Yuichi Tsunoda, Okitoshi Tung, Steve Ueno, Satoshi Vachkov, Gancho Vai, Ming-I Vanderborght, Bram Vernon, Brent Vlacic, Ljubo Voos, Holger Wada, Masayoshi Wada, Osami Wada, Takahiro Wan, Feng Wan, Xinhua Wang, Baikun Wang, Cheng Wang, DongMei Wang, Gang Wang, Guoli Wang, Hua Wang, Joseph Wang, Keqi Wang, Lizhen Wang, Ludan Wang, Michael Yu Wang, Peishan Wang, Pu Wang, Shuchen Wang, Shuxing Wang, Tongyu Wang, Wen Wang, Xianghong Wang, Xiaoyun Wang, Xin Wang, Xinsong Wang, Yafeng Wang, Yuechao Wang, Yuezong Wang, Zhidong Wang, Zhuo Wang, Zongyi

viii

Wang, Zuobin Warisawa, Shin-ichi Watanabe, Keigo Watanabe, Mutsumi Wei, Wei Wen, Bangchun Wen, Paul Wong, Pak-Kin Wu, Changhua Wu, Gang Wu, Jinglong Wu, Lei Wu, Qiong Wu, Shijing Wu, Xiaofeng Wu, Xiaojun X.Yang, Simon Xi, Jeff Xi, Zhihong Xiang, Zhengrong Xiao, Jizhong Xiao, Lan Xiao, Nan Xie, Lihua Xie, Ming Xie, Shane Xie, Shaorong Xin, Ming Xing, Huiming Xiong, Caihua Xu, Bo Xu, Chunquan Xu, De Xu, Dingjie Xu, Fen Xu, Honghai Xu, Jianan Xu, Lixin Xu, Mengguo Xu, Qingsong Xu, Shijie Xu, Yaoqun Xue, AnKe Xue, Dingyu Yakou, Takao Yamada, Takayoshi Yamaguchi, Tomomi Yamamoto, Manabu Yamamoto, Motoji Yamamoto, Yoshio Yamashita, Atsushi Yamaura, Hiroshi Yan, Shaoze Yan, Shengyuan Yan, Zhao Yanagihara, Mamoru Yang, Cheng Yang, Enxia Yang, Erfu Yang, Fang Yang, Guiliin Yang, Hyun Suck Yang, Jianwu Yang, Jing Yang, Kwangjin Yang, Qingsheng Yang, Wu Yang, Xiukun Yang, Yong Yang, Yousheng Yang, Zhaojun Yano, Masafumi Yao, Yiyu Ye, Cang Ye, Changlong Ye, Shujiang Ye, Xiufen Yi, Byung-Ju Yi, Chuanyun Yi, Jianqiang Yin, Guofu Yin, Zhengsheng Yin, Zhouping Ying, Lixia Ying, Xianghua Yokoi, Kazuhito Yokokohji, Yasuyoshi Yokota, Sho Yoshida, Shunichi You, Bo Young, Nak Yu, Dejie Yu, Huadong Yu, Jie Yu, Junzhi Yu, Qiang Yu, Shui Yu, Xiaoyang Yu, Yong Yu, Yueqing Yu, Zhangguo Yuan, Jianjun Yuan, Juntang Yuan, Libo Yuan, Xiaobu Yue, Chunfeng Yue, Dong Yue, Yong Yun, Chao Yuta, Shinichu Zeng, Chunnian Zha, Hongbin Zhang, Baida Zhang, Chengjin Zhang, Dan Zhang, Dianlun Zhang, Hong Zhang, Jianpei Zhang, Jianwei Zhang, Jinxiu Zhang, Lei Zhang, Lijun Zhang, Lixun Zhang, Mingjun Zhang, Rubo Zhang, Sen Zhang, Songyuan Zhang, Xianmin Zhang, Xiaolong Zhang, Xiaoyu Zhang, Xinming Zhang, Xuping

Zhang, Yanhua Zhang, Yi Zhang, Yimin Zhang, Yong Zhang, Yongde Zhang, Yonggang Zhang, Youmin Zhang, Yunong Zhang, Zhaohui Zhang, Zhe Zhao, Cangwen Zhao, Chunhui Zhao, Lin Zhao, Qing Zhao, Xin Zhao, Xinhua Zhao, Yuxin Zhao, Zhijun Zheng, Fei Zheng, Guibin Zheng, Jinyang Zheng, Liang Zheng, Lingling Zheng, Yuanfang Zhong, Ning Zhou, Wei Zhou, Xunyu Zhu, Chi Zhu, Chunbo Zhu, George Zhu, Jianguo Zhu, Qidan Zhu, Xiangyang Zhu, Xiaorui Zhu, Xilin Zhu, Yu

Zu, Jean Zyada, Zakarya

IEEE ICMA 2021 Conference Cosponsors

Cosponsored by

IEEE Robotics and Automation Society Kagawa University, Japan Beijing Institute of Technology, China

Technically cosponsored by

General Information

Takamatsu

Takamatsu City is located facing the calm, island-dotted Seto Inland Sea. The lives of the Takamatsu people have always been strongly tied economically and culturally to the sea. Takamatsu is both the capital of the prefecture and the central management city for Shikoku. The temperature is relatively stable year-round, and there is very little rainfall. Takamatsu was founded in the Kamakura period (1185-1333), and in 1588, Toyotomi Hideyoshi's retainer, Ikoma Chikamasa built a castle on the Tamamo Coast and named it Takamatsu Castle. This is how Takamatsu got its name. The Ikoma family ruled the town for four enerations (54 years) and the Matsudaira family did so for eleven generations (220 years). During the Meiji Restoration, the feudal system was abolished, and Takamatsu was made the capital of Kagawa prefecture. Takamatsu was incorporated as a municipality on February 15, 1890, becoming the 40th incorporated city in Japan.

Since the 1910s, there have been eight municipal mergers, and now Takamatsu City stretches from the Seto Inland Sea in the north, to the Tokushima prefectural border in the south. It has become a wide municipal area blessed with the ocean, mountains, and rivers, a lively city center, and slow-paced rural districts. The city has become a place where urban functionality and natural resources are in balance, providing a good life for its citizens.

Thanks to geographical and other natural factors, Takamatsu has always been the central management city for Shikoku, but in particular because of the Seto Ohashi Bridge which was built in 1988, the New Takamatsu Airport which was built in 1989, and the Takamatsu Expressway which reached Takamatsu in 1992, in April of 1999, the city was designated a core city.

Now, Takamatsu is working at becoming an even more convenient and green city, making use of the unique qualities that each region of the city has to offer, creating a compact and sustainable city.

xii

• Map of Accommodations

➢ It takes 15-20 mins on foot from IEEE ICMA 2021 Conference site to

accommodation ⑨

xiii

Attractions

• Ritsurin Garden Among the gardens in Japan designated as National Special Scenic Beauty, Ritusrin Garden is the largest. Construction started around 1625 by Takatoshi Ikoma, the feudal lord of Takamatsu, and took about 100 years with successive feudal lords to complete in 1745. The garden has six ponds and thirteen mounds strategically placed to use Mt Shiun as a background. Different flowers bloom all year round, changing the scenery as you walk. “One step, one scenery.” The garden also has an excellent reputation overseas. It was given three stars as the highest-rated, worth-visiting place for sightseeing in the Michelin Green Guide Japan in 2009.

• Tamamo Park (Ruins of Takamatsu Castle) There is a song about Takamatsu Castle that goes, “You can see the Takamatsu Castle above the sea in Sanuki.” The castle is also called Tamamo Castle because Kakinomoto Hitomaro used the word “Tamamo yoshi” as a pillow word of Sanuki in Manyoshu. Takamatsu Castle was built by Chikamasa Ikoma, the first feudal lord of Takamatsu. Taking several years to build, the castle has outer, middle and inner moats. It is one of only three castles surrounded by water moats in Japan. At that time, Takamatsu Castle had a magnificent view of the surrounding area with a unique Western-style, three-story, five-layered, castle tower. However, it was removed as it became old. The Ushitora tower, Moon-watch tower as well as Mizutegomon gate still exists which reflects ancient times. A recreational area with old pine trees and other beautiful trees attracts many residents and tourists.

xiv

• Hiunkaku with romantic Taisho atmosphere Hiunkaku was built as a second house for Yorinaga Matsudaira, the 12th feudal lord of them Matsudaira family. Construction started in the second year of the Taisho period and took about three years to complete. There is a room called the “big study room” (142 tatami mats) and there are seven rooms named after the view each room has. For example: the “cycadophyte room,” the “pine tree room” and the “wave room”. Nowadays, tea ceremonies and concerts are held with a romantic Taisho atmosphere-a fusion of Japanese tradition and Western skills.

• Beautiful cherry blossom of Sakura no baba (turf of cherry blossom) “Sakura no baba” is located in the southern part of the castle tower ruins. As its names shows, this was the place used to train horses in ancient times, but now has become one of the best places to see cherry blossoms with as many as 90 cherry trees blooming in spring.

• A Well-known Genpei Battlesite: Yashima The peak of Yashima is flat and resembles a roof, which is the origin of the name “Yashima,” which means “roof island.” There are three viewpoints located along a walking path atop the mountain, named Dankorei, Shishi no Reigan, and Yu-kakutei. From there, the entire Seto Inland Sea National Park can be seen, as well as the island-dotted Seto Inland Sea, and the cityscape of Takamatsu City. A battle site from the Genpei War can also be seen on the mountain. In addition, the 84th temple in the Shikoku 88 Temple Pilgrimage, Yashima-ji, is located atop the mountain, as well as the Shin-Yashima Aquarium, which is said to be located highest above sea level of all the aquariums in Japan.

• Konpira Shrine (Kotohira-cho) Long long passage leading to the shrine is lined up a variety of shops. Hard to step up but fan to see on the way to famous shrine. Kanemaru-za on the way is the oldest theater in Japan and an important cultural asset.

Weather

Takamatsu has a humid subtropical climate with hot summers and cool winters. Some rain falls throughout the year, but the months from May to September have the heaviest rain. Month JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

°F 48.7 49.3 55.2 66.2 74.5 80.1 87.3 89.1 81.7 72 62.2 53.4 °C 9.3 9.6 12.9 19.0 23.6 26.7 30.7 31.7 27.6 22.2 16.8 11.9

Transportation

All the registrants should make their own local transportation in the city. In Takamatsu, it is necessary to book a taxi excepting some dedicated places, such as the JR Shikoku Takamatsu Station and the

Takamatsu Airport. The taxi price contains the base price JPY620 below 1.5km and extra JPY80 for per

314m.

It takes about 30 minutes by taxi from the Takamatsu Airport to the Takamatsu Symbol Tower, the taxi fare is about JPY5000 (approx. US＄51). Whenever you arrive at the airport, there are always many taxis waiting at the airport to pick up passengers. The other method is by bus. It also takes about 30 minutes from the Takamatsu Airport to the JR Shikoku Takamatsu Station which is very near to the

Takamatsu Symbol Tower. You can find the bus at the No.2 bus stand, the bus fare is JPY740 (approx.

US＄7). Please ask for a receipt with the taxi or the bus.

xvi

Useful Information

• Language: Official language is Japanese. English can be understood by many young people and is used in hotels and big restaurants. In all tourist hotels, staff can speak in Japanese, English and other languages. They can also write down addresses or instructions in Japanese for taxi drivers or others.

• Currency: The Japanese yen is the official currency of Japan. It is the third most traded currency in the foreign exchange market after the United States dollar and the euro. It is also widely used as a reserve currency after the U.S. dollar, the euro and the pound sterling. Money exchanges by cash or traveler’s cheques can be made at the branches of Bank of Japan at Takamatsu Airport, hotels and tourist stores. Please remember to keep the receipt to exchange back to foreign currency when leaving Japan..

• Credit Cards: Visa, Master Card and American Express are the most commonly used in Japan. Cards can be used in most middle to top-range hotels, department stores, but they cannot be used to finance your transportation costs.

• Time: GMT + 9 hours (the whole of Japan is set to Tokyo time)

• Electricity: Electricity is 100 Volts, which is different from North America (120V), Central Europe (220V) and most other regions of the world. Japanese electrical plugs have two, non-polarized pins.

• Water: Bottled mineral water can easily be bought in all stores and automatic vending machine for JPY150. And sometimes hotels provide it free of charge. Furthermore, potable water is only available in a few 4 to 5 star hotels, while water in thermos flasks in rooms is usually non-potable tap water.

• Measurement: Japan uses centimeters and meters, kilogram, liter, km/h as SI units. But measurement tools and measure tapes sold in Japan often include inch as well

• Tipping: On the rare occasion that you actually need to give a tip in Japan, do so by putting the money inside of a tasteful, decorative envelope and seat it. Hand it to the recipient with a slight bow; do not expect them to open your “gift” right away. Pulling cash out of your pocket in full view of the recipient is the worst way to give a tip in Japan

• Attention: Smoking is prohibited in public places in Takamatsu, such as hospitals, office buildings, theatres, cinemas, museums, planes, and trains.

• Hotlines: 110 - Police 119 – Fire and Ambulance

xvii

Conference Information

Conference Venue

Due to the influence of the Covid-19 pandemic, the IEEE ICMA 2021 will be a full virtual conference using Zoom software, Located in Takamatsu, kagawa, Japan.

JR Hotel Clement Takamatsu which serves as the official conference hotel is a 20-storey luxury hotel located in Sunport Takamatsu. The hotel offers guest rooms commanding a view over the Seto Inland Sea and the city of Takamatsu, and boasts a range of facilities including multi-purpose banquet halls, making it the ideal choice for business, sightseeing or international conferences. JR Hotel Clement Takamatsu has a range of restaurants including Chinese, Western and Japanese to cater to every taste. The map of the Conference Venue and Awards Banquet Venue

xviii

Japanese Address Cards

JR Hotel Clement Takamatsu

JR ホテルクレメント高松

住所：香川県高松市浜ノ町 1-1 760-0011

Tel: (81-87) 811-1111 Fax: (81-87) 811-1100

xix

Conference Registration

A conference registration desk will be opened online from 13:30 on August 8 to 11:00 on August 10 in Online Conference Main Room.

August 8, 2021 13:30~18:30 (Online Conference Main Room) August 9, 2021 07:30~18:30 (Online Conference Main Room) August 10, 2021 08:00~18:00 (Online Conference Main Room)

Social Events

The social events organized by the IEEE ICMA 2021 include the conference reception, the awards banquet, the conference registration, the farewell party, etc.

Awards Banquet and Farewell Party

The Awards Banquet will be held online from 18:30 to 21:00 on August 10, 2021 in Online Conference Main Room.. All the conference participants are welcome to join this event.

IEEE ICMA 2021 Conference

Keynote Speech

Intelligent Robots and Moon Shot Program

Toshio Fukuda, Ph.D.

Professor Meijo University, Japan 1 Chome-501 Shiogamaguchi, Tempaku Ward, Nagoya, Japan Postcode: 486-8502, Japan E-mail: [email protected] http://www.mein.nagoya-u.ac.jp

xxi

Abstract:

Recent robot technology (RT) has made remarkable progress in both manufacturing and service sectors. Because of this RT advanced technology, there are growing demands to make robots work more friendly and flexible coordinated with human for service. There are many researches and developing works undergoing for robot and human interaction, such as assistance and supports of human by robots in manufacturing, inspection and maintenance, entertainment, education, bio-medical applications, rehabilitation and techno-care of aged people. Robot is required to have the more flexibility and adaptation control to human behavior, more friendly robot and human interface, and estimation capability of human intention some way to make more proactive motion. There are a lot of problems to solve them with robotic sensor, actuator, control, communication and interface with human. This talk is an overview of the Multi-scale robotics, based on the Cellular Robotics System, which is the basic concept of the emergence of intelligence, coevolution and self-organization in the multi-scale way from Organizational Level, Distributed Robotics to Biological Cell Engineering and Nanorobotics, which was proposed more than three decades ago. It consists of many elements how the system can be structured from the individual to the group/society levels in analogy with the biological system. It covers with the wide range of challenging topics, but intelligent robotics is focused on this talk. In particular, focusing on the coevolution and self-organization capabilities, I will show a new initiative on AI and Robot, one of the Moon-Shot Programs started by Japanese Government, since 2020. Based on the Society 5.0, it is a new and challenging program aiming at the AI robotic system in 2050. I will introduce some of the projects in this program for realization of the Society 5.0 by back-casting technologies from the 2050 to the current ones. Toshio Fukuda (M'83-SM'93-F'95) graduated from Waseda University, Tokyo, Japan in 1971 and received the Master of Engineering degree and the Doctor of Engineering degree both from the University of Tokyo, in 1973 and 1977, respectively. He is Professor Emeritus of Nagoya University and Professor of Meijo University and Waseda University. He is mainly engaging in the research fields of intelligent robotic system, micro and nano robotics, bio-robotic system and industry applications in robotics and automation. He was the President of IEEE Robotics and Automation Society (1998-1999), and IEEE President (2020). He was Editor-in-Chief of IEEE/ASME Trans. Mechatronics (2000-2002). He was chairs of many conferences, such as the Founding General Chair of IEEE International Confere on Intelligent Robots and Systems (IROS, 1988), IEEE Conference on Cyborg and Bionic Systems (CBS, 2017), IEEE Conference on Intelligence and Safety of Robots (ISR, 2018). He has received many awards such as IEEE Robotics and Automation Pioneer Award (2004), Technical Field Award (2010). IEEE Fellow (1995). SICE Fellow (1995). JSME Fellow (2002), RSJ Fellow (2004), VRSJ Fellow (2011). xxii

IEEE ICMA 2021 Conference

Plenary Talk 1

The Future of Robot-Assisted Surgery

Ken Goldberg, Ph.D.

Professor and Director

William S. Floyd Jr. Distinguished Chair in Engineering

Department Chair, Industrial Engineering / Operations Research (IEOR)

Director, AUTOLAB and CITRIS "People and Robots" Initiative Founding Member, Berkeley AI

Research (BAIR) Lab Joint Appointments: EECS, Art Practice, School of Information (UC Berkeley)

and Radiation Oncology (UC San Francisco Medical School).

University of California, Berkeley

E-mail: [email protected]

http://goldberg.berkeley.edu

xxiii

Abstract:

An emerging generation of robots will assist surgeons by performing tedious subtasks such as suturing and debridement to improve consistency, reduce fatigue, and open the door to long-distance tele-surgery.

Advances in AI can be applied to data collected from surgical systems such as Intuitive’ s da Vinci to learn underlying control policies. In this talk I'll present recent advances from our lab including novel hardware and software with applications to cutting, suturing, palpation, dissection, retraction, debridement and a recent result -- "Superhuman Peg Transfer".

Ken Goldberg is UC Berkeley Professor of Industrial Engineering and Operations Research with joint appointments in EECS, College of Engineering, School of Information, and Art Practice, Director,

CITRIS "People and Robots" Initiative, Co-Director, Center for Automation and Learning for Medical

Robotics (Cal-MR), and Adjunct Professor of Radiation Oncology at UCSF Medical School. He was appointed the William S. Floyd Jr Distinguished Chair in Engineering and serves as Chair of the

Industrial Engineering and Operations Research Department. He and his students pursue research in machine learning for robotics and automation in warehouses, homes, and operating rooms. Ken developed the first provably complete algorithms for part feeding and part fixturing and the first robot on the Internet. Despite agonizingly slow progress, he persists in trying to make robots less clumsy. He has over 250 peer-reviewed publications and 8 U.S. Patents. He co-founded and served as Editor-in-Chief of the IEEE Transactions on Automation Science and Engineering. Ken's artwork has appeared in 70 exhibits including the Whitney Biennial and films he has co-written have been selected for Sundance and nominated for an Emmy Award. Ken was awarded the NSF PECASE (Presidential Faculty

Fellowship) from President Bill Clinton in 1995, elected IEEE Fellow in 2005 and selected by the IEEE

Robotics and Automation Society for the George Saridis Leadership Award in 2016.

More information can be obtained in http://goldberg.berkeley.edu

Prof. Goldberg's 50 papers on surgical robotics and brachytherapy radiation delivery: https://goldberg.berkeley.edu/pubs/#MR

xxiv

IEEE ICMA 2021 Conference

Plenary Talk 2

Co-worker Robots for Industrial Applications

Kazuhiro Kosuge

Department of Electrical and Center for Transformative AI and Electronic Engineering Robotics The University of Hong Kong Tohoku University Hong Kong Japan

Abstract:

A Co-worker Robot “PaDY” (in-time Parts/tools Delivery to You robot) was developed for an automobile assembly process usually carried out only by a human worker. PaDY delivers necessary parts and tools to the worker when he/she needs them to reduce the worker’s load, to improve efficiency of the work and to prevent mistakes of the work. PaDY is a simple robot with two degrees of freedom

xxv and was easily implemented in a real assembly process without difficult issues. When we try to apply this concept to different types of tasks, we encounter several issues. An adaptive motion planning scheme has been developed for easy implementation of the co-worker robot “PaDY.” A hybrid active/passive co-worker system has been developed for assisting an assembly of a heavy object. A mechanical logic has been developed for low cost and reliable implementation of a co-worker mobile robot developed for a kitting process. In this talk, several prototype systems will be introduced based on the new concepts.

Dr. Kazuhiro Kosuge (Fellow, IEEE) received the B.S., M.S., and Ph.D. in control engineering from the Tokyo Institute of Technology, in 1978, 1980, and 1988 respectively. After having served as a

R&D Staff of the Production Engineering Department, Nippon Denso Company, Ltd., a Research

Associate at Tokyo Institute of Technology and an Associate Professor at Nagoya University, he joined

Tohoku University as Professor in 1995 and served as Distinguished Professor from 2018 to March 2021.

He is now serving as the Director of the Center for Transformative AI and Robotics, Specially

Appointed Professor of Graduate School of Engineering, Tohoku University. He has recently joined the

University of Hong Kong as Chair Professor in the Department of Electrical and Electronic Engineering.

He received Medal of Honor, Medal with Purple Ribbon, from the Government of Japan in 2018 - a national honor in recognition of his prominent contributions to academic and industrial advancements.

He also received IEEE RAS George Saridis Leadership Award in Robotics and Automation in 2021 for his exceptional vision of innovative research and outstanding leadership in the robotics and automation community through technical activity management. He is an IEEE Fellow, JSME Fellow, SICE Fellow,

RSJ Fellow, JSAE Fellow and a member of the Engineering Academy of Japan. He was the President of the IEEE Robotics and Automation Society, from 2010 to 2011, the IEEE Division X Director, from

2015 to 2016 and the IEEE Vice President for Technical Activities for 2020.

xxvi

IEEE ICMA 2021 Conference

Plenary Talk 3

Post-Da Vinci Surgical Robots: Possibility and Expectations

Max Q.-H. Meng, FIEEE, FCAE

Chair Professor and Chairman Department of Electronic and Electrical Engineering Southern University of Science and Technology Shenzhen, China E-mail: [email protected]

xxvii

Abstract:

Research on surgical robotics is attracting more and more public attention and research efforts during the past decades. Recent revolutionary development and drastic progress in robotic technology in terms of both hardware capability and software power have made it possible for researchers to redefine what surgical robotics is capable of achieving to facilitate complicated medical procedures with much less pain and surgical procedures without even external scars. In this talk, we will start with an introduction to how research on minimally invasive surgical robotics started and what the milestone achievements are, and then move onto our own research efforts on post-Da Vinci surgical robots with several case study examples. Personal thoughts and outlook on future research efforts and potentials in surgical robotics will be outlined to conclude the talk.

Max Q.-H. Meng is currently a Chair Professor and the Head of the Department of Electronic and Electrical Engineering at the Southern University of Science and Technology in Shenzhen, China, on leave from the Department of Electronic Engineering at the Chinese University of Hong Kong. He received his Ph.D. degree in Electrical and Computer Engineering from the University of Victoria, Canada, in 1992. He joined the Chinese University of Hong Kong in 2001 as a Professor and later the Chairman of Department of Electronic Engineering. He was with the Department of Electrical and Computer Engineering at the University of Alberta in Canada, where he served as the Director of the ART (Advanced Robotics and Teleoperation) Lab and held the positions of Assistant Professor (1994), Associate Professor (1998), and Professor (2000), respectively. He is an Honorary Chair Professor at Harbin Institute of Technology and Zhejiang University, and also the Honorary Dean of the School of Control Science and Engineering at Shandong University, in China. His research interests include robotics, perception and intelligence. He has published more than 750 journal and conference papers and book chapters and led more than 60 funded research projects to completion as Principal Investigator. He has been serving as the Editor-in-Chief and editorial board of a number of international journals and as the General Chair or Program Chair of many international conferences, including the General Chair of IROS 2005 and General Chair of ICRA 2021 to be held in Xi’an in June 2021. He served as an Associate VP for Conferences of the IEEE Robotics and Automation Society (2004-2007), Co-Chair of the Fellow Evaluation Committee and an elected member of the AdCom of IEEE RAS. He is a recipient of the IEEE Millennium Medal, a Fellow of IEEE, a Fellow of Hong Kong Institution of Engineers, and an Academician of the Canadian Academy of Engineering.

xxviii

IEEE ICMA 2021

Program at a Glance

Click here to download the zoom software You can click the blue hyperlink to get access to the online conference room August 8-10, 2021 Online Conference Takamatsu Symbol Tower Takamatsu, Kagawa, Japan Tokyo Time Sunday, August 8, 2021 UTC+9: 13:30 - 18:30 Registration Desk Open (Online in Conf. Main Room)

UTC+9: 16:00 - 17:00 Keynote Speech (Prof. Toshio Fukuda) (Online in Conf. Main Room)

Tokyo Time Monday, August 9, 2021 UTC+9: 08:30 - 09:00 Opening Ceremony (Online in Conf. Main Room) UTC+9: 09:00 - 09:50 Plenary Talk#1 (Prof. Ken Goldberg) (Online in Conf. Main Room) UTC+9: 09:50 - 10:40 Plenary Talk#2 (Prof. Kazuhiro Kosuge) (Online in Conf. Main Room) UTC+9: 10:40 - 11:00 Morning Break UTC+9: 11:00 - 12:00 Technical Poster Sessions MA1 (Online in Conf. Main Room) UTC+9: 12:00 - 13:30 Lunch Break UTC+9: 13:30 - 15:00 Technical Sessions MP1 (Online Conf. Room 1-Room 6) UTC+9: 15:00 - 15:15 Afternoon Break UTC+9: 15:15 - 16:45 Technical Sessions MP2 (Online Conf. Room 1-Room 6) UTC+9: 17:00 - 18:30 Technical Sessions MP3 (Online Conf. Room 1-Room 6) Tokyo Time Tuesday, August 10, 2021 UTC+9: 08:30 - 09:20 Plenary Talk #3 (Prof. Max Q.-H. Meng) (Online in Conf. Main Room) UTC+9: 09:30 - 11:00 Technical Sessions TA1 (Online Conf. Room 1-Room 6) UTC+9: 11:00 - 11:15 Morning Break UTC+9: 11:15 - 12:15 Technical Sessions TA2 (Online Conf. Room 1-Room 6) UTC+9: 12:15 - 13:30 Lunch Break UTC+9: 13:30 - 15:00 Technical Sessions TP1 (Online Conf. Room 1-Room 6) UTC+9: 15:00 - 15:30 Afternoon Break UTC+9: 15:30 - 17:00 Technical Sessions TP2 (Online Conf. Room 1-Room 6)

Award Banquet and Farewell Party in Takamatsu Symbol Tower UTC+9: 18:30 - 21:00 (Online in Conf. Main Room)

* 15 minutes (Speech: 12 minutes, Q&A:3 minutes) are scheduled for oral presentation including discussions for each paper.

*30 minutes (core time) are scheduled for poster presentation

xxix

IEEE ICMA 2021 Technical Program, Sunday, August 8, 2021 Room Room 1 Room 2 Room 3 Room 4 Room 5 Room 6 Tokyo Time UTC+9: 13:30 - 18:30 Registration Desk Open (Online in Conf. Main Room) UTC+9: 16:00 - 17:00 Keynote Speech (Prof. Toshio Fukuda) (Online in Conf. Main Room) IEEE ICMA 2021 Technical Program, Monday, August 9, 2021 Room Room 1 Room 2 Room 3 Room 4 Room 5 Room 6 Tokyo Time UTC+9: 8:30 - 9:00 Opening Ceremony (Online in Conf. Main Room) UTC+9: 9:00 - 9:50 Plenary Talk #1 (Dr. Ken Goldberg) (Online in Conf. Main Room) UTC+9: 9:50 - 10:40 Plenary Talk #2 (Dr. Kazuhiro Kosuge) (Online in Conf. Main Room) UTC+9: 10:40 - 11:00 Morning Break UTC+9: 11:00 - 12:00 Technical Sessions MA1 (Poster Session) (Online in Conf. Main Room) UTC+9: 12:00 - 13:30 Lunch Break MP1-1 MP1-2 MP1-3 MP1-4 MP1-5 MP1-6 UTC+9: 13:30 - 15:00 Modeling, Simulation Techniques Control Theory and Application Control Theory and Application Organized session: Biomimetic Signal and Image Processing (I) Signal and Image Processing (IV) and Methodologies (I) (I) (IV) Underwater Robots UTC+9: 15:00 - 15:15 Afternoon Break MP2-1 MP2-2 MP2-3 MP2-4 MP2-5 MP2-6 UTC+9: 15:15 - 16:45 Modeling, Simulation Techniques Control Theory and Application Control Theory and Application Organized session: New Mobile Signal and Image Processing (II) Vision System, Robotic Vision and Methodologies (II) (II) (V) Mechanism and Control MP3-1 MP3-2 MP3-3 MP3-4 MP3-5 MP3-6 Organized session: Biomimetic xxx UTC+9: 17:00 - 18:30 Modeling, Simulation Techniques Control Theory and Application Neuro, Fuzzy, and Intelligent Sensor Design, and Novel Sensing Signal and Image Processing (III) Measurement and Control in and Methodologies (III) (III) Control Systems Robotics IEEE ICMA 2021 Technical Program, Tuesday, August 10, 2021 Room Room 1 Room 2 Room 3 Room 4 Room 5 Room 6 Tokyo Time UTC+9: 8:30 - 9:30 Plenary Talk #3 (Dr. Max Q.-H. Meng) (Online in Conf. Main Room) TA1-1 TA1-2 TA1-3 TA1-4 TA1-5 TA1-6 UTC+9: 9:30 - 11:00 Intelligent Mechatronics and Manipulator Control and Industrial, Manufacturing Process Robot Navigation and Control Mobile Robot System (I) Biomimetic Systems Application (I) Manipulation (I) and Automation Algorithm UTC+9: 11:00 - 11:15 Morning Break TA2-1 TA2-2 TA2-3 TA2-4 TA2-5 TA2-6 UTC+9: 11:15 - 12:15 Intelligent Mechatronics and Manipulator Control and Elements, Structures, and Mobile Robot System (II) Mobile Robot System (V) Organized session: Soft Robotics Application (II) Manipulation (II) Mechanisms UTC+9: 12:15 - 13:30 Lunch Break TP1-1 TP1-2 TP1-3 TP1-4 TP1-5 TP1-6 UTC+9: 13:30 - 15:00 Medical, Biomedical and Organized session: Medical Robots Human-System Interaction and Intelligent Biomedical Instrument Mobile Robot System (III) No show Rehabiitation Systems (I) for Minimal invasive Surgery (I) Interface Technology UTC+9: 15:00 - 15:30 Afternoon Break TP2-1 TP2-2 TP2-3 TP2-4 TP2-5 TP2-6 UTC+9: 15:30 - 17:00 Medical, Biomedical and Organized session: Medical Robots Medical, Biomedical and Mobile Robot System (IV) No show No show Rehabiitation Systems (II) for Minimal invasive Surgery (II) Rehabiitation Systems UTC+9: 18:30 - 21:00 Award Banquet and Farewell Party (Online in Conf. Main Room)

Monday August 9, 2021

Morning Sessions

MA1-P Poster Session (Intelligent Mechatronics and Automation)

Monday August 9, 2021

Afternoon Sessions

MP1-1 Signal and Image Processing (I) MP1-2 Modeling, Simulation Techniques and Methodologies (I) MP1-3 Control Theory and Application (I) MP1-4 Control Theory and Application (IV) MP1-5 Organized session: Biomimetic Underwater Robots MP1-6 Signal and Image Processing (IV) MP2-1 Signal and Image Processing (II) MP2-2 Modeling, Simulation Techniques and Methodologies (II) MP2-3 Control Theory and Application (II) MP2-4 Control Theory and Application (V) MP2-5 Organized session: New Mobile Mechanism and Control MP2-6 Vision System, Robotic Vision MP3-1 Signal and Image Processing (III) MP3-2 Modeling, Simulation Techniques and Methodologies (III) MP3-3 Control Theory and Application (III) MP3-4 Neuro, Fuzzy, and Intelligent Control MP3-5 Organized session: Biomimetic Measurement and Control in Robotics MP3-6 Sensor Design, and Novel Sensing Systems

IEEE ICMA 2021 Conference Digest

MA1-P: Poster Session (Intelligent Mechatronics and Automation)

Session Chairs: Qinxue Pan, Beijing Institute of Technology Ziyi Yang, Kagawa University Online Conference Main Room, UTC+9(Tokyo Time): 11:00-12:00, Monday, 9 August 2021

MA1-P(1) 11:00-12:00 MA1-P(2) 11:00-12:00 Field Crop Extraction Based on Machine Vision The Selective Maintenance of The Multi- Yanjuan Wu,Jian Wang,Yunliang Wang,Yiwen Zhao,and Sai Zhang Component System with Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems, Tianjin University of Technology Considering Stochastic Maintenance Quality • This paper is designed from three Hui Cao, Fuhai Duan aspects: image acquisition, School of Mechanical Engineering，Dalian University of Technology, China background segmentation and crop • Reliability-based selective 1.8 extraction. maintenance for a multi- 1.5

• Image processing steps mainly component system with multiple 1.2

include HSV color space maintenance actions is studied. 0.9 conversion, mainly used for • The maintenance quality of 0.6 γ = 1 background segmentation;Erosion γ = 2 imperfect preventive Probability density function 0.3 γ = 3 algorithm, mainly used for crop maintenance is modeled by Beta 0.0 extraction. 0.0 0.2 0.4 0.6 0.8 1.0 distribution Random variable ηi • Experimental results show that this The Seedling Crops • The maintenance improvement is The probability density function method can extract effective crops related to maintenance frequency. of η (maintenance frequency) from crop images at seedling stage. i

MA1-P(3) 11:00-12:00 MA1-P(4) 11:00-12:00

AUV Path Following Formation Control Based on Solvable Criteria and Sequential Solution Method Extended Kalman Filter for the General 6R Inverse Kinematics Problem Juan Li1,2, Jiaqi Wang2, Ruikun Yuan2 Justin Zhang, Winston Zhang, and Michael Zhang 1.Science and Technology on Underwater Vehicle Technology, Harbin Engineering University University of Michigan, Ann Arbor, USA 2. College of Intelligent Systems Science and Engineering, Harbin Engineering University Harbin, 15001, China • A novel criteria for evaluating • Firstly, a formation coordination controller solvability of 6R joint structures is was designed to control the speed of the proposed using joint-angle following AUVs so that multiple AUVs superposition could complete formation control and formation maintenance. • An improvement on previous • Then, the Extended Kalman Filter (EKF) sequential joint angle solving method is used to estimate the current methods in 6R structures is position of the leader, improve the introduced accuracy of the leader position information • Both methods can help robot A 6R open-loop joint structure received by the followers, and reduce the Formation control based on is equivalent to a 7R closed- formation control error of the system. designers quickly evaluate the EKF with time-varying delay feasibility of 6R joint designs and loop structure solve for joint angles

MA1-P(5) 11:00-12:00 MA1-P(6) 11:00-12:00 Development of a Manufacturing System for An Improved Coning Error Compensation Gear Assembly using Collaborative Robots Algorithm Guoyuan Li, Erlend Holseker, Arvin Khodabandeh, Isak Gamnes Sneltvedt, Erik Bjø rnø y, Mingshuai Cui, Jianhui Zeng and Houxiang Zhang Institute of Intelligent Science and Engineering, Harbin Engineering University Norwegian University of Science and Technology Harbin, China Aalesund, Norway • The current algorithm has a defect: • A manufacturing system is proposed for When the number of sub-samples in gear assembly; each compensation cycle increases, • Two collaborative robots and a mobile the attitude update frequency will robot are used to collaborate on a decrease. planetary gear assembly task; • Improved algorithm：Reuse the angle • Eye-in-hand vision is used to detect increment information of the previous landmarks and shapes to localize parts, N cycles. and force/touch sensor is used to • The improved algorithm can increase assemble the parts smoothly; the attitude update frequency to be Coning Motion consistent with the gyro sampling • Experiment verifies the proposed system. Gear assembly experiment frequency.

IEEE ICMA 2021 Conference Digest

MA1-P: Poster Session (Intelligent Mechatronics and Automation)

Session Chairs: Qinxue Pan, Beijing Institute of Technology Ziyi Yang, Kagawa University Online Conference Main Room, UTC+9(Tokyo Time): 11:00-12:00, Monday, 9 August 2021

MA1-P(7) 11:00-12:00 MA1-P(8) 11:00-12:00 Target Detection and Tracking of Ground Mobile Quick Response Code Application for Virtual Robot Based on Improved SSD Network Model Chenfei Chen, Xiaofei Wang, Jutao Wang and Bin Li Zhongli Ma, Jiadi Li, Lili Wu *, Linshuai Zhang, Yi Kang, Yaohan Zeng Tianjin University of Technology School of Control Engineering, Chengdu University of Information Technology Chengdu, China • Developed by Android Studio • Target data establishment and • Use the ZXing toolkit for two- enhancement. dimensional code recognition. • FSSD network construction and • Three modes of operation on the testing and SSD network based model are provided through screen on deep separable convolution. touch • FSSD-based moving target • Through two-dimensional code detection and tracking. scanning, the animation of complex assembly can be played

interface of the QR app Test sample

MA1-P(9) 11:00-12:00 MA1-P(10) 11:00-12:00

Design and Research of a 36000rpm Nitrogen Analysis of Assembly Deformation of a Compressor Detachable High Speed Rotor Tie Rod Jiangang Zhu, Sheng Feng, Xiaobo Dai, Jian Zhou, Lie Yu Xiaobo Dai, Sheng Feng,Jiangang Zhu,Jiale Tian, School of Mechanical Engineering, Xi'an Jiaotong University, Xi'an, Shaanxi Jian Zhou,Lie Yu School of Mechanical Engineering, Xi'an Jiaotong University, Xi'an, Shaanxi • The nitrogen compressor is • It has the advantages of convenient designed to rotate at 36000 RPM, installation and replacement of parts, power at 150KW and mass flow saving materials and reducing costs. rate at 2.5kg/s. • By calculating the deformation of each • The aerodynamic design part on the tie rod to estimate the change in simulation mainly completed the the pre-tightening force, the assembly detailed design of the impeller, accuracy can be better improved. Detachable high-speed rotor diffuser and volute. • Finite element analysis of the rigidity of • The static strength check and Cfturbo modeling design the parts on the detachable high-speed vibration safety of the impeller rotor tie rod and its influence on the structure are analyzed. deformation of the parts.

MA1-P(11) 11:00-12:00 MA1-P(12) 11:00-12:00 Additive Attention for CNN-based Classification An Autonomous Behavior Switching Method for Xuesheng Li, Qiwei Xu* Indoor Mobile Service Robots School of Aeronautics and Astronautics Yiran Tian, Xingrun An, Xiaoqing Qiu, Xichen Xu and Sen Zhang University of Electronic Science and Tech of China School of Mechanical Engineering, Tianjin University of Technology 2006 Xiyuan Avenue, Chengdu, Sichuan, China Tianjin, China Input Frame • Build a motion planning algorithm Average Pooling library for indoor service robots • An improvement of Squeeze and based on the pluginlib of ROS. →→ x+ Excitation module: change it full 1 1 • Design a ROS-based behavior connection layers into simple linear ReLU switching method to make robots equations, learning 2 constants in →→ 2 x+ 2 adjust the planning algorithms and training. parameters by autonomously • Less parameters and a slightly better Sigmoid initiating the set_parameters performance on classification task. Dot Product service in the move_base. • Work better for a small category and • Verify the designed method in a is practical for industry. simulation environment that is The Additive Attention Module build with ROS and MORSE. Simulation Verification 2

IEEE ICMA 2021 Conference Digest

MA1-P: Poster Session (Intelligent Mechatronics and Automation)

Session Chairs: Qinxue Pan, Beijing Institute of Technology Ziyi Yang, Kagawa University Online Conference Main Room, UTC+9(Tokyo Time): 11:00-12:00, Monday, 9 August 2021

MA1-P(13) 11:00-12:00 MA1-P(14) 11:00-12:00

Trajectory and Attitude Measurement of Skier Based on multi-feature information attention based on MINS/UWB Integration for Indoor fusion for multi-modal remote sensing image Intelligent Skiing System semantic segmentation Xiaodan Cong1, Hongpeng Guo2, Jinliang Ruan3, Lianwu Guan3, and Yanbin Gao3 Chongyu Zhang 1. Institute of Intelligent Manufacturing, Heilongjiang Academy of Sciences, Harbin, China School of Information Science and Technology, University of Jinan 2. School of Physical Education Science, Harbin Normal University, Harbin, China Jinan, China 3. College of Intelligent Sciences and Engineering, Harbin Engineering University, Harbin, China • Trajectory and attitude measurement of • we propose a Dual-way Feature skier based on MINS/UWB integration attention Fusion Network (DFFNet), is proposed. which consists of two branches, optical remote sensing image branch • The proposed MINS/UWB integration and elevation feature branch. algorithm is verified by real Skier in indoor Intelligent Skiing System. • Experiment results on ISPRS Vaihingen image dataset Visualization of multimodal • This system is benefit for the training demonstrate the effectiveness of the framework results in Vaihingen of the skier without any restrictions by proposed method. dataset weather, sites, cost and so on. Ski Machine and Skier Training

MA1-P(15) 11:00-12:00 MA1-P(16) 11:00-12:00

Morphometric Analysis of Coronal Radar Target MTD 2D-CFAR Algorithm Based on Craniosynostosis Bones of Facial Cranium Compressive Detection Cong Liu, Yunqing Liu, Qi Li, Zikang Wei Based on Discrete Cosine Transform Institute of Space Optical Electronic Technology, Heng Nie, Shengmin Zhou, Bingjue Li Changchun University of Science and Technology, School of Mechanical Engineering Southeast University Nanjing, Jiangsu Province, China Changchun, China

• According to the CS theoretical Radar echo Range signal Reference unit Discrete cosine transform (DCT) is model, the compressive detection S k

Y (ij, ) p Protection unit Sparse p

e utilized to approximate nasal in the range-dimension is representation r Detecting unit  = HS aperture, left orbit, right orbit, left completed through the kk k Measurement

zygomatic arch and right zygomatic measurement matrix in the mode matrix Z Y ij,

D T

arch curves，including 43 coronal of no signal reconstruction. M Compressive T H1 Decision detection sampling H craniosynostosis and 20 normal • Then the radar target constant false 0 yS= H persons. Stepwise discriminant alarm detection on Range-Doppler k k k k analysis was performed on the distribution is constructed by using Compressive detection MTD generated coefficients. A three-dimensional skull model of a MTD and 2D-CFAR)combined 2D-CFAR algorithm normal newborn with the compressed detected

MA1-P(17) 11:00-12:00 MA1-P(18) 11:00-12:00 Recovery of Simulated Fuel Debris by Remote Vehicle Pose Estimation System Base on Pressure Sensor Control Robot for Decommissioning Work Array for Clamping Parking Robot Juzhong Zhang, Yuyi Chu, Zhisen Wang, Tingfeng Ye, Liming Cai and Hongbo Yang Naoki Igo, Shota Yamaguchi, Noriyuki Kimura, Kazuma Ueda, Kazuma Kobayashi, Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Sciences Toyoaki Tomura, Toshifumi Satake, Satoshi Mitsui Suzhou, China National Institute of Technology, Asahikawa College Hokkaido, Japan • Vehicle pose estimation is one of the most essential functions for the clamping parking robot. • The goal of the robots is to recover fuel debris. • The methods based on laser scanners are the mainstream, However, the • The robot is wired by LAN. price is very expensive, and the • The operator controls the robot using algorithm is complex. the camera images. • The method based on pressure sensor • Robot consists: main-unit and sub-unit. array is simple, effective and cheap, • The main-unit elevates and lowers the it can not only estimate the pose of a sub-unit by 3.2 meters. car, but also achieve other important Main-unit elevating sub-unit information for the parking robot. The vehicle pose estimation system

IEEE ICMA 2021 Conference Digest

MA1-P: Poster Session (Intelligent Mechatronics and Automation)

Session Chairs: Qinxue Pan, Beijing Institute of Technology Ziyi Yang, Kagawa University Online Conference Main Room, UTC+9(Tokyo Time): 11:00-12:00, Monday, 9 August 2021

MA1-P(19) 11:00-12:00 MA1-P(20) 11:00-12:00 Research on Path Planning of Mobile Robot Research on Active Disturbance Rejection Based on Improved Jump Point Search Control Technology in LCL Grid-Connected Algorithm Inverter Yunliang Wang, Sai Zhang ,Yanjuan Wu, Yiwen Zhao, and Jian Wang Xuesong Zhou, Jie Yin and Youjie Ma Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems, Tianjin School of Electrical and Electronic Engineering, Tianjin University of Technology University of Technology Tianjin, China Tianjin, China • Grid-connected system is a • This paper presents an improved jump complex system with strong V1 V3 V5 point search algorithm. L L2 i1a 1 i2a uga uncertainty and nonlinear . ua L1 L2 u • It was carried on the ROS mobile robot i1b i2b gb • A linear active disturbance Vdc u N b i for experiment. i L1 2c L2 ugc rejection control grid-connected uc 1c • The experimental results show that the V V V current controller is designed. 4 6 2 C C C path generated by the improved JPS • The anti-disturbance performance algorithm is conducive to the robot and tracking performance of the Three-phase LCL full-bridge operation. The Mobile Robot system are analyzed in detail by grid-connected inverter frequency response method.

MA1-P(21) 11:00-12:00 MA1-P(22) 11:00-12:00

Research on Roundness Detection and Sorting of Research on SLAM Algorithm and Oil Cooling Pipe Based on Machine Vision Navigation of Mobile Robot Based on ROS Li Junjie1, Wang Chunguang2, Wu Jin2*, Wang Xiaoming1 ,Zhu Yaqiao2 ,Zhang Shihui2 Bin Liu1, Zhiwei Guan2*,Bin li1, Guoqiang Wen2,Yu Zhao2 1 Tianjin University of Science and Technology, Tianjin 300384, PR China 1 School of Mechanical Engineering, Tianjin University of Technology, Tianjin 300384，PR China 2 Tianjin Sino-German University of Applied Sciences , Tianjin 300350, PR China 2 Tianjin Sino-German University of Applied Sciences, Tianjin 300350, PR China • Robot combined with Machine vision • This article first summarizes the complete the roundness detection and representation method of the sorting of the cooling tube. environment map, uses the ROS mobile • Verify the correctness of the mathematical robot to map the indoor environment, model and kinematics equations through compares the mapping effect of the Adams and Matlab. Gmapping algorithm and the Cartographer algorithm, analyzes the • By comparing the least square algorithm results of the mapping and proposes the and the Hough algorithm, it is concluded optimal mapping plan. that the least square method has higher Visual detection and detection accuracy in roundness detection sorting system • Compare Dijkstra, A* algorithm and and suitable for industrial applications. Dynamic Window Approach and Mobile Robot Autolabor choose the best navigation algorithm.

MA1-P(23) 11:00-12:00 MA1-P(24) 11:00-12:00

Control Strategy Study of Dual-fuel Gas Turbine Design of ADRC for unmanned surface vehicle Zhen Yang, Yan Wang College of Intelligent Systems Science and Engineering, Harbin Engineering University heading based on Immune Particle Swarm Harbin, China Optimization 6000 6000 Fu Mingyu, Wang Qiusu, Wang Yuchao and Xu Yujie • In order to ensure the rapidity and stability of 5000 5000 4000 4000 College of Intelligent Systems Science and Engineering,

dual-fuel gas turbine in the fuel-switching 3000 3000 Gtout_G/(kg/h)

N3/rpm Harbin Engineering University of Harbin Gtout 、 2000 2000 process, this paper proposes a dual-fuel gas Gtout_G N3 Gtout Harbin, China turbine switching control strategy. This 1000 1000 0 0 • An extended state observer is method can ensure the stable control of the 0 10 20 30 40 50 60 70 switching process of the dual-fuel unit, make t/s designed to estimate the ‘total Parametric variation curves under codirectional dynamic the unit quickly complete the transition switching control strategy of gas switching to fuel process disturbance’ and complete the process of the two kinds of fuel-switching, so 6000 6000 design of active disturbance as to realize the fast and stable control of the 5000 5000 dual-fuel unit. In this paper, the process of 4000 4000 rejection control ADRC law.

3000 3000 Gtout_G/(kg/h)

gas-oil switching of the dual-fuel engine has N3/rpm Gtout 、 2000 2000 Gtout_G • For the difficulty of parameter N3 been carried out by the digital simulation Gtout 1000 1000

software. The simulated results show that the 0 0 tuning of ADRC, the IPSO is 0 20 40 60 80 strategy can realize the fast switching control t/s combined with active disturbance between natural gas and diesel oil in an Parametric variation curves under codirectional dynamic switching control strategy of fuel-switching to gas process rejection control to optimize the emergency. The unmanned surface vehicle parameters of ADRC 4

IEEE ICMA 2021 Conference Digest

MA1-P: Poster Session (Intelligent Mechatronics and Automation)

Session Chairs: Qinxue Pan, Beijing Institute of Technology Ziyi Yang, Kagawa University Online Conference Main Room, UTC+9(Tokyo Time): 11:00-12:00, Monday, 9 August 2021

MA1-P(25) 11:00-12:00 MA1-P(26) 11:00-12:00

Research on Vibration Event's Feature Image Matching Algorithm Based on Improved Extraction Method of Φ-OTDR System FAST and RANSAC Qiongnan Yang, Chenguang Qiu，Litao Wu, Jianjun Chen Haiqiang Zhu , Baofeng Zhang ，Zhili Zhang, Huimin Gao East China Institute of Optoelectronic Integrated Devices Bengbu, Anhui Province, China Tianjin University of Technology, Tianjin Sino-German University of Applied Sciences Tianjin, China • This paper proposes a model method that • The algorithm first uses the multi- named the VMD-MPE-SVM model. level FAST algorithm to extract • Optimized the parameters in the variational corner points. modal decomposition algorithm through • The FLANN algorithm is used to different methods: decomposition times complete the rough matching of the and penalty factors. feature points. • Comparing the impact of different • Use the P-RANSAC algorithm to classifiers on the final recognition effect, it remove the part of the mismatched Image matching is found that the recognition effect of the point. svm algorithm is very well. Algorithm flowchart

MA1-P(27) 11:00-12:00 MA1-P(28) 11:00-12:00 Active noise cancellation techniques to enhance Energy Conserving Path when Chasing an audition in noisy cities Intruder by Autonomous Guards J. Alan Calderó n Ch1,2,*., John Lozano Benjamín Barriga1, Julio Tafur1, Juan Carlos Lengua1, Doron Nussbaum and Gonzalo Solano1 School of Computer Science, Carleton University, 1 Engineering Department, Mechatronic Master Program, Pontificia Universidad Cató lica del Ottawa, Ontario, Canada Perú . Lima, Perú 2 Applied Nanophysics, Institute for Physics. Technical University of Ilmenau. Ilmenau, • Chasing intruders with discrete Germany information (snapshots) • It is concluded that sensors and • Real time path planning using actuators (microphones and snapshots loudspeakers) that were based in • Optimizing locations for reducing nanostructures can enhance noise overall energy consumption attenuation Path planning when two snapshots • Path planning which reduces energy • The microphones that were based are allowed. Points p1 and p2 are the consumption path points. in nanostructures can be adapted geometrically to many intricate AAO samples places to measure noise.

MA1-P(29) 11:00-12:00 MA1-P(30) 11:00-12:00 An MR Fluid based Master Manipulator of A Novel Contact Force Measurement the Vascular Intervention Robot with Haptic Scheme For Slave Catheter Robot In Feedback Robotic Endovascular Surgery Qiang Gao, Yu Zhan, Yu Song, Junjie Liu, Jiabin Wu Qiang Gao, Jiabin Wu, Yu Song, Junjie Liu，Zemin Mao and Yu Zhan Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems Tianjin University of Technology, Tianjin, China Technical College for the Deaf Tianjin University of Technology Tianjin, China • Design based on magnetorheological fluid technology. • An intravascular robotic teleoperation system has proposed. • Haptic interface with output controllable resistance. • The catheter manipulator is designed to mimic the surgeon's action at the • The physician's experience in main end. traditional interventional procedures • The contact force between the catheter Catheter operator can be applied. Master manipulator based and the patient's vessels can be on magnetorheological fluid • Device can be disassembled for easy measured by the motor current loss- sterilization. speed-resistance model.

IEEE ICMA 2021 Conference Digest

MA1-P: Poster Session (Intelligent Mechatronics and Automation)

Session Chairs: Qinxue Pan, Beijing Institute of Technology Ziyi Yang, Kagawa University Online Conference Main Room, UTC+9(Tokyo Time): 11:00-12:00, Monday, 9 August 2021

MA1-P(31) 11:00-12:00 MA1-P(32) 11:00-12:00 Design and Development of a Bladder Volume Fixed-time Attitude Feedback Control for Determination Device Based on A-mode Quadrotor UAV based on Fixed-time Extended Ultrasound State Observer Qiang Gao, Yunfei Gao, Zemin Mao, Chunping Liu, Yu Song, Haobo Zhang, Zixuan Zhang Junfang Li , Lei Liu and *Junjie Liu Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems, Tianjin College of Electrical and Electronic Engineering ,Tianjin University of Technology University of Technology Tianjin, China Tianjin, China Pitch Angle tracking effect • The fixed-time attitude feedback 50 • In this paper, an ultrasonic urine Reference 40 LADRC control scheme based on fixed-time FXTADRC system based on A-mode ultrasound 30 is developed for measuring the urine extended state observer （FXTESO) 20 accumulation in the human bladder. is proposed. 10

0 θ/° • The accuracy of the system • Virtual controllers are introduced to -10 measurements in three sets of decouple the system. -20 -30

experiments were 99.8%, 99.65%, • The FXTESO is used to estimate the -40

-50 and 99.52%. Experimental results unknown disturbance in real time. 0 5 10 15 20 25 30 35 40 verify the effectiveness of the t/s The process of the experiment • The independent fixed-time attitude The Pitch angle tracking effect system. feedback control laws are designed

MA1-P(33) 11:00-12:00 Robust controller design based on policy iteration Gao Qiang, Yang Xuxi, Ji Yuehui, Liu Junjie Tianjin Key Laboratory for control Theory & Applications in Complicated Systems Tianjin University of Technology Tianjin, China • A robust controller design method based on online policy iterative algorithm is proposed for a class of nonlinear uncertainty systems. • Transforms the control into the optimal control for the nominal systems. • A single-layer critic network was construct to solve HJB equation, which educes the amount of The State Trajectory calculation.

IEEE ICMA 2021 Conference Digest

MP1-1: Signal and Image Processing (I)

Session Chairs: Toshio Fukuda, Meijo University Ken'ichi Koyanagi, Toyama Prefectural University Online Conference Room 1, UTC+9(Tokyo Time): 13:30-15:00, Monday, 9 August 2021

MP1-1(1) 13:30-13:45 MP1-1(2) 13:45-14:00 Hand Gesture Recognition Based on a All-Around 3D Reconstruction from Spherical Nonconvex Regularization Images with Semantic Segmentation Jing Qin¹, Joshua Ashley² and Biyun Xie² Tuulia Pennanen, Siva Ariram, Antti Tikanmäki, Juha Röning ¹Department of Mathematics, University of Kentucky, USA Biomimetics and Intelligent Systems Group, University of Oulu, Finland ²Department of Electrical and Computer Engineering, University of Kentucky, USA • Using a lightweight, affordable 360- degree camera for dense 3D reconstruction in mobile robotics. • Reconstruction based on the optical flow between a pair of spherical images. • Robustness improved by semantic segmentation. • Suitable for the outdoors and other Reconstruction of a garden environments with uniform or scene from two spherical Hand Gestures repeating textures. images

MP1-1(3) 14:00-14:15 MP1-1(4) 14:15-14:30 3D-Printed Balloon Type Pneumatic Actuator for A Back-Support Exoskeleton with a Cable-Driven Hybrid Force Display Glove Series-Parallel Elastic Actuation: Prototype Ken'ichi Koyanagi, Takumi Tamamoto, Kentaro Noda, Takuya Tsukagoshi, Toru Oshima, and Design and Operational Analysis Daisuke Takata Faculty of Engineering, Toyama Prefectural University Hugo Hung-tin Chan, Hongpeng Liao, Fei Gao, Xuan Zhao, and Wei-Hsin Liao Toyama, Japan The Chinese University of Hong Kong, Hong Kong, China • A type of pneumatic actuator for an • Exoskeleton reduces spinal stress when lifting original virtual-reality-based hand • An exoskeleton with cable-driven series rehabilitation device. parallel elastic actuator (SPEA) is proposed to • An improved actuator that features a provide lifting force new 3D-printed inner construction • Cantilever beams work in parallel with an using new material. active cable-driven series elastic actuator • The 3D-printed balloon actuator is • Possess advantages such as maintaining the for a hybrid force display glove that body degree of freedom, lowering maximum combines an actuator and a passive motor power, and reducing impedance to the Schematic of back- element. 3D-Printed Balloon Actuator user support exoskeleton

MP1-1(5) 14:30-14:45 MP1-1(6) 14:45-15:00

Research on Gait Recognition Based on Lower Automated Real-time 3D Visual Servoing Control Limb EMG Signal of Single Cell Surgery Processes Junyao Wang, Yuehong Dai, Tong Kang, and Xiaxi Si Bo Wen Yao and James K. Mills University of Electronic Science and Technology of China (UESTC) Department of Mechanical and Industrial Engineering, University of Toronto Chengdu, China Toronto, Ontario, Canada

• The EMG signals of thigh muscles and calf • Automated real-time 3D visual servoing muscles under 5 different gait are obtained by control of single cell surgery processes experiment and are input into BP network • 3D image processing program proposed after noise reduction and features extraction for Blastomere centroid localization to recognize gait. • Position based visual servoing (PBVS) • The mean recognition rate of calf EMG controller for real time feedback control signal is higher than that of thigh EMG. of surgery processes When thigh and calf are used together, the recognition rate of all kinds of gait is greatly • Deep learning-based object detection improved. algorithm for cell real-time localization Automated Single Cell Surgery EMG signal Acquisition under low microscope magnification

IEEE ICMA 2021 Conference Digest

MP2-1: Signal and Image Processing (II)

Session Chairs: Biyun Xie, University of Kentucky Ken'ichi Koyanagi, Toyama Prefectural University Online Conference Room 1, UTC+9(Tokyo Time): 15:15-16:45, Monday, 9 August 2021

MP2-1(1) 15:15-15:30 MP2-1(2) 15:30-15:45 Research on Sonar Image Denoising Method An Evaluation of RGB-Thermal Image Based on Fixed Water Area Noise Model Segmentation for Snowy Road Environment Min Chen1,2, Lei Li1, Zejie Li1 and Xiaomei Xie1 Sirawich Vachmanus, Ankit A. Ravankar, Takanori Emaru, and Yukinori Kobayashi 1.School of Aeronautics and Astronautics University of Electronic Science and Technology of Human Mechanical Systems and Design Engineering, Hokkaido University China Chengdu, Sichuan Province, China Hokkaido, Japan 2.Yangtze Delta Region Institute (Huzhou)University of Electronic Science and Technology of China Huzhou, Zhejiang Province, China • Semantics segmentation in snowy • Theoretical modeling of noise in environment. fixed waters. • Combination information of RGB • Guided filtering image denoising image and Thermal image. based on frequency-domain filtering • Comparison between daytime and preprocessing. nightime segmentation in snow. • Sonar image acquisition based on • Dual inputs network structure with COMSOL. concatenate operation. • Simulation analysis of image RGB-T semantics segmentation denoising performance. The Sonar Image after Denoising

MP2-1(3) 15:45-16:00 MP2-1(4) 16:00-16:15 Automatic Intracranial Aneurysm Segmentation Attenuation Coefficient of Tissue Based on Based on Spatial Information Fusion Feature Modified Depth-Resolved Model Huang Lin, Liu Yana, Zhang Xiao, Li Qin from 3D-RA using U-Net Department of Biomedical Engineering, School of Life Science,Beijing Institute of Technology Mengqi Cheng, Nan Xiao*, Hang Yuan, Kaidi Wang 5 South Zhongguancun Street, Beijing, China Key Laboratory of Convergence Biomedical Engineering System and Healthcare Technology, • The modified depth-resolved model is The Ministry of Industry and Information Technology, School of Life Science, used to calculate the tissue attenuation Beijing Institute of Technology, Haidian District, Beijing 100081, China • Prosed a automatic coefficient based on OCT data. intracranial aneurysm • The methods of subtracting the noise segmentation method. term and median filtering were used to • The SIF feature and reduce the influence of noise on the data augmentation are attenuation coefficient estimation result used to improve model • The DR model by subtracting the noise performance. term can characterize the hierarchical The Optical Phantom • Carried out three structure of the tissue better and make experiments for model The network structure the attenuation coefficient more accurate. verification.

MP2-1(5) 16:15-16:30 MP2-1(6) 16:30-16:45

A Zigzag Maneuver Target Tracking Algorithm with Colored Design of Convenient Structured Light Projection Glint Measurement Noise System for Microscopic Measurement Chenghao Shan and Weidong Zhou Department of Intelligent Systems Science and Engineering, Harbin Engineering University Yimeng Du, Yuezong Wang*, Youfan Peng Harbin, China Faculty of Materials and Manufacturing, Beijing University of Technology • Model glint measurement noise Beijing, China as student’s t distribution • A convenient structured light projection • Transform the problem into the system is designed for the measurement estimating problem for a non- of the morphology of small objects in linear system with white heavy- the height range from tens of microns to tailed measurement noise. hundreds of microns. • The augmentation state vector • The structured-light projection system together with the intermediate can meet more requirements of variables are simultaneously The true trajectory and measurement by editing the properties of Setup of structured-light inferred by introducing tracking trajectories of the structured-light pattern, and it is low- projection system variational Bayesian approach. zigzag maneuvering target. cost and flexible to use.

IEEE ICMA 2021 Conference Digest

MP3-1: Signal and Image Processing (III)

Session Chairs: Haiyuan Wu, Wakayama University Peng Shi, Kagawa University Online Conference Room 1, UTC+9(Tokyo Time): 17:00-18:30, Monday, 9 August 2021

MP3-1(1) 17:00-17:15 MP3-1(2) 17:15-17:30 A Super Baseline for Pedestrian Re- A feature fusion model for Person Identification Identification using Top-view Image Jiwei Zhang, Haiyuan Wu Jiwei Zhang, Haiyuan Wu Graduate School of Systems Engineering, Wakayama University Graduate School of Systems Engineering, Wakayama University Wakayama, Japan Wakayama, Japan • We propose a new baseline that • We identify the area of the person in combines and improves two types of the captured image by overlooking deep learning algorithms to achieve the camera, and fuse the extracted more accurate pedestrian re- features. identification from images taken by • First, extract 4 features from the surveillance cameras. region of interest: • We introduce some tricks into a famous 1) texture ,2) color,3) GLCM, strong baseline. We conducted ablation 4)features extracted by VGG. experiments on each trick using unlearned data and confirmed the • Then we fused these features and applied them to the SVM classifier effectiveness and stability of the Pedestrian Re-Identification Person Recognition proposed method from the results. to classify and recognize pedestrians.

MP3-1(3) 17:30-17:45 MP3-1(4) 17:45-18:00 Improving Caption Consistency to Image with Profile Fitting-based Small Target Detection in Semantic Filter by Adversarial Training Water for Side-scan Sonar Image Junlong Feng, Jianping Zhao Zhanshuo Liu, Xiufen Ye, Shuxiang Guo, Huiming Xing, Zengchao Hao, Yao Li School of Computer Science and Technology, College of Intelligent Systems Control and Engineering, Harbin Engineering University Changchun University of Science and Technology, Harbin, Heilongjiang Province, 150001 China Changchun, China • This is a two-stage detection • A semantic filter module is method, and each stage is designed to learn informative consistent with the imaging semantic knowledge. principle of side-scan sonar. • Our model takes full • This method is suitable for advantage of text vectors and distinguishing target and visual features from both speckle noise. local and global levels. • This method does not need to • Linguistic analysis verified obtain the sample information the efficacy in promoting The Semantic Filter Module of the target in advance. The Clustering of peak points caption fluency and accuracy.

MP3-1(5) 18:00-18:15 MP3-1(6) 18:15-18:30 A Lightweight Defect Detection Algorithm of Research on Recognition of Pointer Meter Based Insulators for Power Inspection on Improved East Algorithm Lei Yang1, Shouan Song1, Yong Niu2 and Yanhong Liu1 Lei Shao , Yuxiang Chen , Xiaoning Xu, Wentao Sun, Hong liLiu 1. School of Electrical Engineering, Zhengzhou University, 450001, China Tianjin Key Laboratory for Control Theory &Application in Complicated Systems, School of 2. Xining Urban Vocational & Technical College, Qinghai Province, China Electrical and Electronic Engineering Tianjin University of Technology, China • A novel lightweight defect • Uses the lightweight neural network detection algorithm is proposed MOGA to replace the backbone for insulator defects which network of East algorithm, thus fuses the advantages of shallow learning and deep learning. reducing the memory occupied by • An insulator location algorithm the algorithm. based on YOLOV3 Tiny is • The feature pyramid is added to proposed to remove the backbone network to strengthen the disturbance of backgrounds for defect detection. feature extraction. • A defect recognition method • Use NMS local sensing algorithm to with SVM and transfer learning Framework of proposed defect screen the reading box to get the is proposed to solve the small- detection algorithm Table reading interface. scale defect detection issue. scale number and position of the instrument.

IEEE ICMA 2021 Conference Digest

MP1-2: Modeling, Simulation Techniques and Methodologies (I)

Session Chairs: Hideyuki Hirata, Kagawa University Wei Zhou, Beijing Institute of Technology Online Conference Room 2, UTC+9(Tokyo Time): 13:30-15:00, Monday, 9 August 2021

MP1-2(1) 13:30-13:45 MP1-2(2) 13:45-14:00 A Rollover Strategy for Wrist Damage Reduction Automated Exploration, Capture And in a Forward Falling Humanoid Photogrammetric Reconstruction Of Interiors Dongdong Liu, Yuhang Lin, and Vikram Kapila Mechanical and Aerospace Engineering Department, NYU Tandon School of Engineering, Using An Autonomous Unmanned Aircraft Brooklyn, NY, USA Markus Lieret, Vladyslav Kogan, Christian Hofmann and Jörg Franke Department of Mechanical Engineering, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany • Design, implement, and examine a rollover strategy for a humanoid robot by • We present a fully automated system for employing a differential dynamic digitization and photogrammetric recon- programming approach struction of complex indoor environments. • Simulation and multi-trial experiment results show over 44% wrist impact • Using an autonomous unmanned aircraft equipped with a suitable sensor system, reduction using the rollover strategy The rollover strategy and intelligent algorithms for navigation and compared with bimanual fall bimanual strategy exploration a previously unknown • First paper that experimentally validates environment is autonomously explored and rollover wrist damage reduction strategy a three-dimensional model is reconstructed. Result of the photogrammetric in a humanoid robot reconstruction

MP1-2(3) 14:00-14:15 MP1-2(4) 14:15-14:30 Development of a Fixture for the Vascular A Discrete Particle Swarm Optimization for Interventional Surgical Robotic System Storage Location Assignment Problem of Retail Jian Guo1* ,Xiuqiang Shao1 and ShuxiangGuo1,2 E-Commerce 1Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems and Intelligent Robot Laboratory, Tianjin University of Technology,Tianjin, China Chaodan Zhao, Jianping Dou and Xia Zhao 2Intelligent Mechanical Systems Engineering Department Faculty of Engineering, Kagawa School of Mechanical Engineering , Southeast University University, Kagawa, Japan Nanjing, Jiangsu Province, China • Establish an optimization model of the This paper proposes a vascular storage location assignment problem interventional surgery robot (SLAP) for retail e-commerce. system with a bionic clamp. The • Propose a new discrete particle swarm robot uses a gear slider structure optimization (DPSO) algorithm. to push and retract the guide • Compare the performance of the DPSO (a) Initial cargo position. (b) Optimized cargo position. wire catheter. Its clamp is Object Function Before DPSO and the artificial fish swarm algorithm Efficiency Priority 11765.64 9777.5 mainly used to clamp and twist Shelf Stability 5.4213 3.3584 (AFSA) versus five instances. Similar Products Adjacency 517.78 319.6615 the catheter and guide wire, and • Verify the superiority and reliability of adopts a portable design. The structure of the slave manipulator Cargo position of 150 goods the DPSO in solving the SLAP.

MP1-2(5) 14:30-14:45 MP1-2(6) 14:45-15:00 Trajectory Tracking Control for a Biomimetic Robust Optimization Design of Gun Turret Spherical Robot Based on ADRC Structure for Lightweight Mugen Zhou, Shuxiang Guo, Xihuan Hou, He Yin, Ao Li, Debin Xia, Zan Li, Meng Liu Yao Ge, Longmiao Chen, Jianhui Tan and Caicheng Yue Key Laboratory of Convergence Medical Engineering System and Healthcare Technology, School of Mechanical Engineering, Nanjing University of Science and Technology Beijing Institute of Technology, Beijing, China Nanjing, China • This article addresses the problem of 2-D trajectory tracking for • Aim at the structural optimization biomimetic spherical robots (BSR) with uncertain models and problem of a gun turret. unknown disturbances in a constrained workspace. • The robust optimization model of • A robust active disturbance rejection control (ADRC) scheme and the turret structure is established by thrust allocation scheme are presented to deal with the various using 6σ robust optimization design operational constraints. method. • The RBF is used to construct the surrogate model. • The results are obtained by using the The gun turret model combinatorial optimization The trajectory tracking algorithm. control framework

IEEE ICMA 2021 Conference Digest

MP2-2: Modeling, Simulation Techniques and Methodologies (II)

Session Chairs: Kazuhiro Kosuge, Tohoku University Ruochen An, Kagawa University Online Conference Room 2, UTC+9(Tokyo Time): 15:15-16:45, Monday, 9 August 2021

MP2-2(1) 15:15-15:30 MP2-2(2) 15:30-15:45 Optimization of Pin Arrangement and Geometry in EV and HEV Heat Sink Using Genetic UR5 robot manipulation using Algorithm Coupled With CFD Matlab/Simulink and ROS A. Vivas*, J. Sabater** Weichang Yang, James K. Mills *University of Cauca, Colombia; **Miguel Hernandez University, Spain [email protected], [email protected] Department of Mechanical and Industrial Engineering, University of Toronto, Toronto, Canada • Manipulation of the UR5 robot. • Optimizing the pin position, diameter • Matlab/Simulink are widely used in and height in pin fin heat sink for EV and HEV. academic and industrial research. • The optimized pin fins exhibit better • Manipulation scheme is made with heat transfer coefficients by ROS Toolbox. 15.8%~20.3% than a linear uniformly • Coppelia Vrep is used to test the spaced pin design. trajectories. • Varying pin height and diameter are Optimized Pin Fin Heat Sink • Real UR5 robot controlled from also beneficial in improving the Pattern, Temperature Matlab/Simulink via ROS. performance of the heat sink. UR5 robot being manipulated from Contour and Flow Diagram Matlab

MP2-2(3) 15:45-16:00 MP2-2(4) 16:00-16:15

A Multi-Physics Simulation Approach to A Fast Algorithm of Normal Vector Calculation for Predict Shape Morphing of Flexible NURBS Surface Based on Taylor Expansion Devices in Magnetic Field Yinuo Zhou, Qiang Liu, Zhenshuo Yin, Sanying Zhu School of Mechanical Engineering and Automation, Beihang University Huichen Ma, Junjie Zhou Jiangxi Research Institute of Beihang University School of Mechanical Engineering, Beijing Institute of Technology Beijing, China Beijing, China • Only necessary to calculate the derivative • An approach for general multi-physics vector of the iso-parametric line of the B- simulations of multifunctional hydrogels. spline surface at each node from 0 to its • Negative thermal expansion coefficient is order in the NURBS surface fraction introduced into the magneto-thermal- expression. mechanical coupled model. • The Taylor formula can be used to • Based on multi-physics simulation platform, calculate the corresponding parameters in the analysis of flexible devices is carried the normal unit vector of the NURBS surface. out. Simulation results are close to similar The Magnetic Simulation Results reported experimental results. Bilayer Hydrogel • Is 40-50% faster than the calculation of deBoor algorithm.

MP2-2(5) 16:15-16:30 MP2-2(6) 16:30-16:45 Assessment of Large Vessel Support Capability On Combined Elastic and Based on Comprehensive Weighting Model Nonholonomic Model of a Class of Peng Shang, Yanhua Sun, Zhuo Zhou, Nieyong Huang and Jian Zhou Mobile Robots with Arc Wheels School of Mechanical Engineering, Xi’an Jiaotong University Shaanxi, China Xuchao Huang, Songxin Zhou, Zhi Yang and Yisheng Guan * Guangdong University of Technology（GDUT） • A three-level assessment indicator Subjective weight Objective weight Guangzhou, Guangdong, China. Comprehensive weight 11 system has been established to analyze 0.7 43 12 • No longer limited to the analysis and large vessel support capability. 0.6 0.5 42 21 verification of a single wheel, but from a 0.4 • A comprehensive weighting model 0.3 system perspective to model the hexapod. 0.2 based on AHP and entropy method has 41 22 0.1 • Avoid existing modeling methods cannot 0.0 been developed combing the qualitative model the characteristic of stress point of the 34 23 and quantitative characteristics of the wheel is the lowest point on the wheel. assessment indicators. 33 23 • The analytical solution of the kinematics can • The model can reduce the influence of 32 25 be expressed by a formula instead of a 31 the uncertainty of the expert group and complicated system of equations. Hexapod improve reliability of assessment . Weights of the indicators

IEEE ICMA 2021 Conference Digest

MP3-2: Modeling, Simulation Techniques and Methodologies (III)

Session Chairs: Qinxue Pan, Beijing Institute of Technology Wei Zhou, Beijing Institute of Technology Online Conference Room 2, UTC+9(Tokyo Time): 17:00-18:30, Monday, 9 August 2021

MP3-2(1) 17:00-17:15 MP3-2(2) 17:15-17:30 Study on Propagation Characteristics of Refracted Remaining Useful Life Indirect Prediction of Longitudinal Waves in Carbon Fiber Composites Lithium-ion Batteries Based on Dropout Gated Qinxue Pan, Yunmiao Zhang, Lang Xu, Rui Luo, Shuangyang Li, Wei Li, Sa Li Recurrent Unit School of Mechanical Engineering, Beijing Institute of Technology Meng Wei, Min Ye*, Qiao Wang, and Xinxin Xu Beijing, China National Engineering Laboratory for Highway Maintenance Equipment Chang‘an University • The figure shows amplitude Xi’an, China curves of refracted • RUL prediction framework based on compressional waves. the dropout Gated Recurrent Unit is • The main lobe turns toward proposed. the surface as the incident (a) (b) • The charging voltage saturation time is angle increases. 1.5M 2.5M selected as indirect health indicator. • The main lobe in CFRP is • The dropout method is proposed to closer to the surface. avoid the over-fitting. • Another main lobe can be • The Gated Recurrent Unit is proposed Electric Vehicle seen in CFRP materials. (c) (d) to address the gradient disappearance phenomenon. 2M Steel

MP3-2(3) 17:30-17:45 MP3-2(4) 17:45-18:00 A Novel State of Charge Estimation Method of Discussion and Improvement of The Lithium-ion Battery Based on NARX Neural Rectification Capability of NEL Rectifier under Network Model The Concentric Reducer Based on FLUENT Qiao Wang, Min Ye, Meng Wei, Chenguang Wu and Xinxin Xu Simulation National Engineering Laboratory for Highway Maintenance Equipment Chang 'an University Chao Dong1 ,Yuchuan Li1，Xin Shao2 ,Siqi Han2 ，Tao Wang2 Xi'an, China 1School of Electrical and Electronic Engineering, Tianjin University of Technology, Tianjin 2Intelligent Manufacturing College, Tianjin SinoGerman University of Applied Sciences, Tianjin • Use the NARX neural network to establish the battery model. • Understand the rectification effect of the NEL rectifier on • Combine with the moving window the turbulence of the flow field caused by the concentric method. reducer through the FLUENT simulation • Verify the effectiveness of the • Improve the structure of the rectifier by analyzing the flow field distribution and optimize its rectification effect proposed model by comparing with The Improved Rectifier other estimators. • Through the FLUENT simulation, the changes of the flow field distribution in the tube before and after the improvement were analyzed, and the results showed that The electric vehicles the rectification effect was greatly improved.

MP3-2(5) 18:00-18:15 MP3-2(6) 18:15-18:30 Research on Energy Consumption Forecast of Research on Line Loss Calculation of Urban Enterprise Power Equipment Based on Cable Line Algorithm Fusion Mode Y. Liu, Y. Li, C. Li Department of Kunming University of Science and Technology Ji Li , Xiang Ma , Chao Li and Lei Shao Chenggong District, Kunming, Yunnan,China School of Electrical and Electronic Engineering, Tianjin University of Technology, China • This paper proposes and investigates the • Using CEEMD to decompose power theoretical line loss calculation of urban equipment data, and using ARMA to underground cables based on the CGA forecast each component. optimized IBP model . • Established a new type of prediction • The proposed algorithm has the property model-CEEMD-ARMA model. of the proposed method has the • The prediction results of the characteristics of being easy to operate CEEMD-ARMA model and the and improving the calculation accuracy EMD-ARMA model are compared. of the theoretical line loss of urban underground cables. • The results show that the prediction Prediction Comparison Chart accuracy of the CEEMD-ARMA • Experimental results have verified the model is relatively high. efficacy of the proposed algorithm. Urban Cable Line

IEEE ICMA 2021 Conference Digest

MP1-3: Control Theory and Application (I)

Session Chairs: Hideyuki Sawada, Waseda Univesity Liang Zheng, Changchun University of Science and Technology Online Conference Room 3, UTC+9(Tokyo Time): 13:30-15:00, Monday, 9 August 2021

MP1-3(1) 13:30-13:45 MP1-3(2) 13:45-14:00 Conservatism Comparison of SUIO and Zero-sequence Circulating Current Luenberger Observer in Invariant Set-Based Suppression Strategy for Microgrid Inverters Robust Fault Detection for LPV Systems Xuesong Zhou1,2, Kairui Guo1,2,Youjie Ma1,2 and Weibao Zhong1,2 Bo Min1, Jun Yang2, Junbo Tan1, Changliang Wang3 and Xueqian Wang1 1. Tianjin Key Laboratory for Control and Application in Complicated Systems.2.School 1Center of Intelligent Control and Telescience, Tsinghua Shenzhen International Graduate of Electrical and Electronic Engineering, Tianjin University of Technology School, Tsinghua University, 518055 Shenzhen, P.R.China Tianjin, China 2 Navigation and Control Research Center, Department of Automation, Tsinghua University, 100084 Beijing, P.R.China • In order to solve the problems of 3 V V V Shanghai Academy of Spaceflight Technology, 201109 Shanghai, P.R.China 11 31 51 system loss caused by ZSCC, an R1 ia1 L1 ea ua1 C ib1 eb u ub1 N improved controller was designed. dc i e u c1 c • The stability conditions are established. The controller adds compensation for c1 V41 V61 V21 • A practical construction approach of the total disturbance estimation error. V V V the invariant sets is presented. 12 32 52 • In this way, zero-vector allocation R2 ia2 L2 ua2 ib2 • An effective comparison criterion of ub2 factor is obtained to adjust the duty i u c2 the FD conservatism is obtained by c2 cycle, eliminate the difference V42 V62 V22 using the Frobenius norm-based sizes The FD results between the inverters, and thereby of residual zonotopes. suppress ZSCC. Parallel structure diagram

MP1-3(3) 14:00-14:15 MP1-3(4) 14:15-14:30 Research on Auto Disturbance Rejection Control Vector Control of Active Disturbance Rejection Strategy of Battery Energy Storage Staggered Induction Motor based on Sliding Mode Variable Parallel Buck Converter Xuesong Zhou, Diankang Su*, Youjie Ma, Junqing Pan Structure School of Electrical and Electronic Engineering, Tianjin University of Technology Xuesong Zhou, Junqing Pan* , Youjie Ma, Diankang Su Tianjin, China School of Electrical and Electronic Engineering,Tianjin University of Technology • A linear active disturbance Tianjin, China rejection control strategy udc iL* + u0 + u • A composite control strategy Kp 1/b0 PWM  + r + for two-phase interleaved - - - based on sliding mode variable SMC - + - buck converter is proposed. Kd structure active disturbance 1 b 0 b • It is used for the constant z1 z2 z3 rejection control is proposed. 0 LESO Buck Converter z2 z1 current charging control of iL LESO Second-order Linear Active Disturbance • The controller not only retains the battery and compared i r Rejection Controller + - Battery  + + + the strong robustness of transfer K with PI control. function SVPWM PI u - i - • The linear active Constant current charging control traditional sliding mode, but s s disturbance rejection diagram of Buck converter based on also effectively weakens the control strategy has better LADRC motor chattering problem SMADRC double closed loop control through compensation anti-jamming performance.

MP1-3(5) 14:30-14:45 MP1-3(6) 14:45-15:00

MPPT Control of Photovoltaic System Modeling and Control Strategy Simulation of Combined with Correction Link and LADRC Dual Active Bridge DC-DC Converter Youjie Ma, Congcong Wei, Xuesong Zhou, and Weibao Zhong Youjie Ma, Hongyu Wen, Xuesong Zhou and Jie Yin School of Electrical and Electronic Engineering, Tianjin University of Technology Tianjin Key Laboratory of Control Theory and Application for Complex Systems Tianjin, China Tianjin University of Technology ,Tianjin, China • A maximum power point tracking (MPPT) controller combined with linear active disturbance rejection control of the correction • The small signal model of DCDC

link (LADRC-C) is designed. i1 H H i2 i0 converter is established by small 1 2 • Introduce the idea of lead and lag V y S1 S4 T M1 M 4 m u Photovoltaic perturbation method. iL MPPT k 1/b n :1 p + system correction to adjust the bandwidth U1 L vH 2 C C U2 − − • An Active Disturbance Rejection 1 v 2 R of the disturbance observation b H1 Virtual Power Control (ADR-VPC) S2 S3 M 2 M 3 channel, which improves the z z 3 strategy is proposed observation accuracy of the linear 1 Improved LESO Dual active bridge expansion state observer. • The design of the controller is DC-DC converter • Significantly improve the tracking System structure diagram derived in detail speed of the system and greatly under LADRC-C control reduce the power oscillation.

IEEE ICMA 2021 Conference Digest

MP2-3: Control Theory and Application (II)

Session Chairs: Hideyuki Sawada, Waseda Univesity Yu Song, Tianjin University of Technology Online Conference Room 3, UTC+9(Tokyo Time): 15:15-16:45, Monday, 9 August 2021

MP2-3(1) 15:15-15:30 MP2-3(2) 15:30-15:45 Error-constrained Moving Path-Following Experimental Validation of the Open-Source Control for a Stratospheric Airship DMPC Framework GRAMPC-D applied to the Liran Suna,*, Ming Zhub, Xiao Guoc, Jiace Yuana and Huabei Goua Remotely Accessible Robotarium aSchool of Aeronautic Science and Engineering, Beihang University, Beijing, China Daniel Burk, Andreas Völz, Knut Graichen, bInstitute of Unmanned System , Beihang University, Beijing, China Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany cResearch Institute for Frontier Science , Beihang University, Beijing, China • An error-constrained moving path • GRAMPC-D is an open-source software following (EMPF) for a stratospheric airship is first introduced and framework for distributed model predictive formulated. control (DMPC) in the milli second range. • An error-constrained guidance • This paper presents an experimental method based on a simple BLF and validation of the software framework. an extended LOS guidance law is designed and the error constraint • Three experiments are presented with 4, 6 requirement will never be violated. and 11 robots. • Two adaptive controllers based on • The picture shows a moment during the backstepping technique are designed formation control experiment. The Formation control for the attitude and velocity control Geometrical illustration of variables and frames formation is given by the letter “G”. under external disturbances.

MP2-3(3) 15:45-16:00 MP2-3(4) 16:00-16:15 Forceps Motion Estimation in Laparoscopic Rectifier Stage Control Strategy for Solid Surgery using Deep Learning for Robotic State Transformer Based on Sliding Mode Camera Control Active Disturbance Rejection Masahiko Minamoto, Yamato Umetani, Shigeki Hori, Tetsuro Miyazaki, and Kenji Kawashima Zhiqiang Gao, Jiafeng Liu，Xuesong Zhou, Youjie Ma Monozukuri Engineering Department, Tokyo Metropolitan College of Industrial Technology Tianjin Key Laboratory for Control Theory & Applications in Complicated Industry Tokyo, Japan Systems, Tianjin University of Technology • We propose a method to estimate the Tianjin, China forceps tip motion 0.1 seconds • SMADRC solve the problem that ahead by deep learning using the distributed generation units are segmented forceps in camera image. often in multivariable, strong coupling and strong unknown • We experimentally confirmed that disturbances. the forceps movement of 0.1 seconds ahead can be estimated • SMADRC consists of SMC and online in pick-and-place and suture LESO. tasks. the estimate position of • The effectiveness of the control method is verified by simulation. structure diagram of energy router the forceps 0.1 seconds ahead.

MP2-3(5) 16:15-16:30 MP2-3(6) 16:30-16:45

Three-Level Control of DC Side Voltage Balance A general approach to modeling of Cascaded H-Bridge APF demand-side resources Yunliang Wang,Zichen Li,Yanjuan Wu,Zhengkun Bai,and Ang Li Yunliang Wang,Zhengkun Bai,Yanjuan Wu,Zichen Li, and Ang Li Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems, Tianjin Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems, Tianjin University of Technology University of Technology Tianjin, China Tianjin, China

isa La iL ea Ra usa L e Rb b u • A three-layer voltage equalization b isb sb Nonlinear A general virtual battery model was L ec Rc isc c usc Load

control strategy. ca L cb L cc established by limiting energy scenario i i i L u ca ucb ucc method. • Three modules cascade H bridge, CB1 CC1 CA1 significantly improve APF • Optimize the electricity consumption

C C pressure level. CA2 B2 C2 plan according to the change of time- • Repetitive control is adopted to of-use price. CA2 CB3 CC3

improve the accuracy of current. A phase B phase C phase • The specific electricity consumption information of consumers is hidden to Cascaded H-bridge protect the privacy of users. Virtual battery APF

IEEE ICMA 2021 Conference Digest

MP3-3: Control Theory and Application (III)

Session Chairs: Chao Jia, Tianjin University of Technology Yu Song, Tianjin University of Technology Online Conference Room 3, UTC+9(Tokyo Time): 17:00-18:30, Monday, 9 August 2021

MP3-3(1) 17:00-17:15 MP3-3(2) 17:15-17:30 Stable Tuning of Extended State Observers Self-Imitation Learning for Robot Tasks with using PSO and Penalty Functions Sparse and Delayed Rewards Ricardo Cortez, Ruben Garrido Zhixin Chen and Mengxiang Lin* Automatic Control Deparment, Cinvestav-IPN School of Mechanical Engineering and Automation, Beihang University, Beijing, China CDMX, Mexico Nanjing Research Institute of Electronics Technology, Nanjing, China • Tuning of Extended State Observer • Standard RL would fail in real-world using PSO. robot applications with sparse and • Proposal of the optimization delayed rewards. problem in a way that the stability • A practical self-imitation learning of the observer be guaranteed. method is developed by using constant • Comparison of Under Low reward assignment instead of Constraints respect the use of immediate rewards defined manually. Penalty Functions. Tuned observer using PSO • The effectiveness of the method is • Use of Penalty Functions using the Algorithms with Under Low evaluated empirically in continuous The framework of self- roots of the polynomial such could Constraints (ULC) and robotics control of OpenAI Gym imitation be applied to a n-order system. Penalty Functions (PF) MuJoCo. Learning for robotic learning

MP3-3(3) 17:30-17:45 MP3-3(4) 17:45-18:00 Disturbance Rejection via Fuzzy Control with Deep Transfer Learning in Inter-turn Short Circuit Disturbance Observer for Active Magnetic Fault Diagnosis of PMSM Bearing System Yuefeng Fang, Manyi Wang and Liuxuan Wei School of Mechanical Engineering, NanJing University of Science and Technology Bo Wang, Haipeng Geng, Wei Zheng, Jiayi Yao and Dingchong Lyu Nanjing,China School of Mechanical Engineering, Xi 'an Jiaotong University Xian, China • proposes an efficient inter-turn • AMB system full of uncertainty short circuit fault diagnosis and interference. method based on transfer learning and 1d-CNN. • Fuzzy logic can comprehend the inaccuracies and uncertainties of • optimizing the 1d-CNN based the actual world. on L1 regularization, and cost- sensitive loss function strategy. • Observer-based state feedback. • Experiments are designed to • Disturbance observer based CNN based on transfer verify the effectiveness of the fuzzy control method is learning. proposed method. presented to achieve disturbance The control structure of AMB system rejection.

MP3-3(5) 18:00-18:15 MP3-3(6) 18:15-18:30 Decomposition of Complex System and Multi- Application of Maximum Power Point Tracking level Control Method based on Control Control in Pendulum Wave Energy Converter Allocation Chunjie Wang1, Xuece Li1，Peng Chen1, Lin Cui2, 1Tianjin Complex System Control Theory and Application. 2 National Ocean Technology Center. Jia Chao,Li Ziyu Tianjin, China Tianjin University of Technology Control Engineering • This paper analyzes the dynamic Tianjin, China • Aiming at the control problems of model of pendulum wave power Complex redundant complex redundant systems, a system generation system under the effect method of decomposing complex of regular sea waves. The first Control Second I-th subsystem nonlinear systems into subsystems subsystem distribution subsystem • Design the optimal perturbation step and designing discrete controllers is Decoupling size perturb-and-observe method. First sub- Second sub- I-th sub- proposed. controller controller controller • It overcomes the shortcomings of • The coupling problem is solved local optimum and the power through control allocation, and the oscillation like traditional perturb- control of the complex redundant and-observe method. The system Pendulum wave energy Overall control strategy system is realized simply and can be quickly to track nearby the generation system model. effectively. maximum power point. 15

IEEE ICMA 2021 Conference Digest

MP1-4: Control Theory and Application (IV)

Session Chairs: Liwei Shi, Beijing Institute of Technology Cheng Yang, Beijing Institute of Technology Online Conference Room 4, UTC+9(Tokyo Time): 13:30-15:00, Monday, 9 August 2021

MP1-4(1) 13:30-13:45 MP1-4(2) 13:45-14:00 SINS Stationary Base Initial Alignment Based on Simulation Design of Intelligent Vehicle Backtracking Scheme Transverse Control System Xuan Xiao, Yongyan Zhang Mingqiu Guo, Yunde Shi*, and Dan Xia School of Automation, Beijing Institute of Technology Department of Mechanical Engineering, Southeast University Beijing, China Nanjing, China • Backtracking scheme improves the alignment accuracy by • Vehicle three-degree-of-freedom processing the data of inertial dynamics model. measurement unit forward and • MPC algorithm with new constraints backward repeatedly. equations. • The model of reverse navigation • Joint simulation of CarSim and and filtering is derived. Simulink. • An optimal backtracking • Noble accuracy of high-speed number criterion is designed to trajectory tracking control. Intelligent Vehicle judge whether filtering process is converged. Error of Yaw Angle (Vehicular)

MP1-4(3) 14:00-14:15 MP1-4(4) 14:15-14:30 High Voltage Gain DC-DC Converter with Bridge- Application of SOC estimation based on Double-Voltage Cell Based on Coupled Inductor unscented Kalman filter in photovoltaic-mains Chunjie Wang, Hao Yu, and Lin Cui Tianjin Complex System Control Theory and Application complementary power management system Key Laboratory, Tianjin University of Technology Tianjin, China Lei Shao, Chenyang Yan, and Huapeng Zeng • A high voltage gain DC-DC converter School of Electrical and Electronic Engineering, Tianjin University of Technology, China N N Do4 S1 S 2 C based on coupled inductors is proposed. * o2

Do3 C • Thevenin model was used to build a • The converter adopts interleaved parallel o1 N P2 Do2 first-order RC model of the battery. structure, and uses the bridge-double- * S N 2 voltage cell as the output, which can P1 • The state of charge of the battery was Do1 achieve high voltage gain with low duty estimated by using untraceable Kalman S Cc V 1 ratio. in filter (UKF) algorithm. • Analyzing the operating principle of the • Compared with the traditional converter. The circuit topology of the Extended Kalman Filter algorithm, the proposed converter estimated value of SOC obtained by SOC comparison curve • Building a prototype for verifying the UKF is closer to the real value of SOC. correctness of the theroy.

MP1-4(5) 14:30-14:45 MP1-4(6) 14:45-15:00 Research on Improved p-q Reactive Current Model Reference Scheduling and Robust Detection Method Based on Synchronous Resilient H-infinity Control Co-design with Time- Rotary Angle delay Shunli Zhao Xuesong Zhou, Weibao Zhong, Youjie Ma, Kairui Guo 1. Tianjin Key Laboratory for Control Theory & Applications in Complicated System, Tianjin Tianjin University of Technology University of Technology, Tianjin, China 391, Binshui Xidao, Xiqing District, Tianjin, 300384, China 2. School of Electrical and Electronic Engineering, Tianjin University of Technology, Tianjin,  ua d ua 1 X • A voltage detection method LPF sin = China dt 100 Z Y based on the synchronous rotary LPF cos = • A co-design of the robust resilient Actuator Plant Sensor Z angle is proposed. sin cos H-infinity control and MRS x()k u a ud u xk() ... xk() • the voltage detection method is u LPF d scheme is studied in the paper. 1 n b abc− dq dq−− abc  u uq uq introduced into the traditional p- c LPF • The effect of the weighted matrix Network δ()k u u q detection method to form an   on the performance of the MRS

improved p-q reactive current ia p p ia1 i LPF i scheme is studied.  ()k b abc− pq q q pq− abc a2 detection method based on the i ia3 c LPF • The effectivity is demonstrated synchronous rotary angle. u()k xˆ()k Schematic diagram by a numerical example. Controller Scheduling scheme in NCS 16

IEEE ICMA 2021 Conference Digest

MP2-4: Control Theory and Application (V)

Session Chairs: Liwei Shi, Beijing Institute of Technology Wei Zhou, Beijing Institute of Technology Online Conference Room 4, UTC+9(Tokyo Time): 15:15-16:45, Monday, 9 August 2021

MP2-4(1) 15:15-15:30 MP2-4(2) 15:30-15:45

Code Summarizer Fault Grouping Bias Strategy of Thrust Shreya Mehta, Sneha Patil, Nikita Shirguppi and Dr. Mrs. Vahida Attar Department of Computer and IT, Allocation for Dynamic Positioning Ship College of Engineering, Pune, India Fuguang Ding, Junfeng Zhang* and Yuanhui Wang* College of Intelligent Science and Engineering, Harbin Engineering University • Source Code Summarization implies Harbin, China generating summary in natural • Solve the problem of thrust language from a given code snippet. allocation when the thruster fails • We propose an approach that uses during the dynamic positioning. UAST (Universal Abstract Syntax • Propose a grouping bias thrust Tree) of the source code to generate allocation algorithm for the thruster tokens and then use the Transformer in failure mode. model with self-attention • Propose a minimum bias algorithm, mechanism that uses Encoder- which could avoid the thrusters to Decoder. Dynamic Positioning Ship enter the thrust forbidden zone due • BLEU, METEOR and ROUGE-L Transformer Model to bias operation. are the three metrics we have used for evaluation of our model.

MP2-4(3) 15:45-16:00 MP2-4(4) 16:00-16:15

Robust adaptive back-stepping control for Research on Control Strategy of Bidirectional nonlinear systems with unknown backlash-like DC-DC Converter Based on Enhanced ADRC hysteresis Xuesong Zhou, Bingjie Xie, Youjie Ma, and Qian Liu School of Electrical and Electronic Engineering, Tianjin University of Technology Yanyan Chen, Yue Wang, Qi Dong, Guiming Zhu, Jingjing Feng, and Feiyun Xiao Tianjin, China

Jiangsu JARI Information Technology Co., Ltd, Liangyungang, Jiangsu, China • Objective: To solve the problem of Distributed Generation Bidirectional Battery • A robust adaptive control method stable control of DC bus voltage in the DC-DC independent photovoltaic energy based on backstepping approach is DC-DC Energy Storage （MPPT） presented. storage system. System • The control algorithm can ensure the • Method: An enhanced LADRC is DC Bus global stability of the nonlinear proposed, which introduces differential closed loop. signal of total disturbance into the DC-AC DC-DC DC Load • This method may be used in the traditional LESO . AC Load DC Load Load mechanical actuators, piezo-electric • Conclusion: The disturbance estimation Independent Photovoltaic materials, and tendon sheath and disturbance rejection ability of the The tracking performance and the Power Generation System mechanism system. control input for example 1 system are improved.

MP2-4(5) 16:15-16:30 MP2-4(6) 16:30-16:45

Reduced-order ADRC of DC-Link Voltage of Application of Model Compensated Active Photovoltaic Grid-connected Inverter Disturbance Rejection Control in Electric Xuesong Zhou, Qian Liu, Youjie Ma and Bingjie Xie Vehicle Charging School of Electrical and Electronic Engineering, Tianjin University of Technology Youjie Ma, Min Yan, Xuesong Zhou, and Jie Yin Tianjin, China School of Electrical and Electronic Engineering,Tianjin University of Technology Tianjin,China • Problem: The control performance Ps Pdc Pg • The rectifier of the charging system of the DC bus voltage is constrained Front stage Rear stage is ig r u0 u y by uncertain factors such as is taken as the research object. kp 1/b0 G(s) i dc L e temperature, irradiance and the grid. a • A Model Compensated Active b0 + + i eb kd C a z1 - - ib ec 1/s  u udc Disturbance Rejection Control (MC- 1 • Solve: A Reduced-order active pv ic disturbance rejection control ADRC) is applied to DC link. f0 z2 (ADRC) controller is presented in • MC-ADRC uses some known 1/s 2 Topology structure of two-stage z3 this paper to achieve strong disturbances for feedforward 1/s 3 PV grid-connected system tolerance to various disturbances. compensation, which reduces the observation pressure of Linear MC-LADRC control Extended State Observer(LESO). block diagram

IEEE ICMA 2021 Conference Digest

MP3-4: Neuro, Fuzzy, and Intelligent Control

Session Chairs: Jin Guo, Beijing Institute of Technology Chuqiao Lyu, Beijing Institute of Technology Online Conference Room 4, UTC+9(Tokyo Time): 17:00-18:30, Monday, 9 August 2021

MP3-4(1) 17:00-17:15 MP3-4(2) 17:15-17:30 A System to Identify Walking Pattern Using Learning the Spatial Perception and Obstacle Machine Learning for a Load Reduction Avoidance with the Monocular Vision on a Exoskeleton Robot Quadrotor Hanqing Zhao, Hidetaka Nambo Jiajun Ou1, Xiao Guo2, Wenjie Lou3 and Ming Zhu3 Artificial Intelligence Laboratory, 1School of Aeronautic Science and Engineering, Beihang University, Beijing, China Graduate School of Natural Science and Technology 2Research Institute for Frontier Science, Beihang University, Beijing, China Electrical Engineering and Computer Science, 3Institute of Unmanned System, Beihang University, Beijing, China Kanazawa University, Kanazawa, Japan • An unsupervised contrastive • Gait cycle state recognition and learning method to extract the time-series gait prediction based on spatial representation. the distribution of plantar pressure. • No scene images with prior • Iot method to collect plantar information is required. pressure distribution data • Select optimal obstacle avoidance • CNN+LSTM+Ensemble Learning Gait cycle action from multiple features in for Multi-Model Processing. plantar pressure distribution sequence rather than the latest The Trajectory of Successful one only. Obstacle Avoidance

MP3-4(3) 17:30-17:45 MP3-4(4) 17:45-18:00

Fault Diagnosis Method of Reciprocating The Vector Control Scheme for Amphibious Spherical Robots Based on Reinforcement Learning Compressor Based on Domain Adaptation under He Yin, Shuxiang Guo, Liwei Shi, Mugen Zhou, Xihuan Hou, Zan Li, Debin Xia Multi-working Conditions School of Life Science,Beijing Institute of Technology Lijun Zhang1, Lixiang Duan2, Xiaocui Hong2, and XinYun Zhang2 Beijing, China 1College of Mechanical and Transportation Engineering, China University of Petroleum(Beijing) 2College of Safety and Ocean Engineering, China University of Petroleum(Beijing), • The proposed two-layer network Beijing, China framework uses the adaptive ability of reinforcement learning to realize the • A fault diagnosis method of reciprocating compressor based on domain adaptation is control of the amphibious spherical proposed in this paper. robot. • The combination of CEEMDAN and WT can • The proposed two-layer reinforcement be effective in reducing the noise-induced learning framework for amphibious interference. spherical robot can better process the • The ResNet50-MK-MMD method is used for upper and lower control commands Two-layer reinforcement learning fault diagnosis under multi-working condition, and reduce the difficulty of training. framework for amphibious spherical robots with the average accuracy reaching above 97%. The reciprocating compressor

MP3-4(5) 18:00-18:15 MP3-4(6) 18:15-18:30

Fuzzy-Improved Linear Active Disturbance Perturbation Fuzzy Neural Fractional-order Sliding Rejection Control of Solid State Transformer Mode Control of Micro Gyroscope Juntao Fei, Jie Zhuo, Fang Chen Inverter Stage College of IoT Engineering, Hohai University, China Youjie Ma, Tong Zhang, Xuesong Zhou Y • A fractional-order sliding mode  Output Layer School of Electrical and Electronic Engineering, Tianjin University of Technology W W W control method based on a double 1 2 m Tianjin,China   1 2 m recurrent perturbation fuzzy neural −1 −1 z ··· z Recurrent Layer

network is proposed to achieve r r2 r · 1 · m pre pre· pre • The first order differential term 1 2    m Wro1 1 2 m W adaptive estimation of unknown   ···  ro2 Rule Layer of the given signal is introduced system model.    11 12 1m  21  22  2m into LSEF for compensation • The sine-cosine perturbation ··· ··· Membership Function Layer

membership function is used to for  elimination. 1 2 uncertainty of rules in neural Input Layer x x • Fuzzy control combined with network to increase the accuracy. 1 2 improved linear state error • Simulations show the proposed feedback control rate. method can accurately estimate the Structure of DRPFNN The Solid State Transformer unknown nonlinear term and reduce • MATLAB & Simulink was used the output tracking error. for modeling.

IEEE ICMA 2021 Conference Digest

MP1-5: Organized session: Biomimetic Underwater Robots

Session Chairs: Keisuke Morishima, Osaka University Liwei Shi, Beijing Institute of Technology Online Conference Room 5, UTC+9(Tokyo Time): 13:30-15:00, Monday, 9 August 2021

MP1-5(1) 13:30-13:45 MP1-5(2) 13:45-14:00 D* Lite-Based Navigation Algorithm for Multiple Motion Stability Evaluation of the Improved Spherical Underwater Robots Collaboration Spherical Underwater Robot with Awa Tendeng, Shuxiang Guo, Ruochen An, and Chunying Li Intelligent Mechanical System Engineering, Kagawa University Hybrid Propulsion Devices Takamatsu, Japan Chunying Li1, Shuxiang Guo1, 2*, Jian Guo2, Ruochen An1, and Tendeng Awa1 1Kagawa University, Hayashi-cho, Takamatsu, 761-0396, Japan • This paper explores the problem of 2Tianjin University of Technology, Tianjin, 300384, China safe navigation in non-deterministic • A hybrid propulsion device is designed underwater environment for multiple for the spherical underwater robot to robot’s collaboration. enhance motion stability and operation • To ensure an optimal path to the goal, D* Lite evaluation with a single robot efficiency. we implemented the D* Lite • The motion experiments in the pool are algorithm. carried out, including forward-motion and • The collaborative navigation is rotation-motion experiments.

simulated with Gazebo software. (m) Coordinates Y • The experimental result showed that the improved hybrid propulsion devices we X coordinates (m) designed had better performance. Hybrid modes of the HPSUR Multiple robots navigation

MP1-5(3) 14:00-14:15 MP1-5(4) 14:15-14:30 Dynamic Modeling and Optimization of Robotic Numerical Exploration on Pitching Motion of Fish Based on Passive Flexible Mechanism Robotic Dolphin Realized by Pectoral Fin Qianqian Zou, Ben Lu, Yuzhuo Fu, Xiaocun Liao, Zhuoliang Zhang, Chao Zhou Cao, Zhihan Li, Xufeng Zhou and Dan Xia* Institute of Automation, Chinese Academy of Sciences School of Mechanical Engineering, Southeast University Beijing, China Nanjing, China • The elastic tail is composed of a passive flexible joint and a rigid caudal fin. • The pitching ability of robotic • The pseudo-rigid-body model dolphin is relative to the flapping (PRBM) is ﬁrstly used to analyze method of symmetrical fin. the kinematics of the ﬂexible joint. • By changing the water-striking • A dynamic model of the robotic fish angle over a fin movement cycle， is established by the Lagrangian The robotic fish with flexible the dolphin robot can obtain great dynamic method. passive mechanisms propulsion and lift force. • Passive flexible joints can be The Trajectories and rotation utilized to optimize performance by adjusting the stiffness.

MP1-5(5) 14:30-14:45 MP1-5(6) 14:45-15:00 Trajectory Tracking Control of an Amphibious Study on Hydrodynamic Characteristics and Spherical Robot Using MPC Approach Dynamics Model of Underwater Spherical Robot Meng Liu1 , Shuxiang Guo1,2∗ , Liwei Shi1∗ ,Xihuan Hou1,2 , He Yin1, Ao Li1 ,Zan Li1 ,Debin Ao Li1, Shuxiang Guo1,2*, Liwei Shi1, Xihuan Hou1, Zan Li1, Debin Xia1 Xia1 Mugen Zhou1 1 Key Laboratory of Convergence Medical Engineering System and Healthcare Technology, the Key Laboratory of Convergence Medical Engineering System and Healthcare Technology, Ministry of Industry and Information Technology, Beijing Institute of Technology, Beijing, China Beijing Institute of Technology 2 Faculty of Engineering, Kagawa University, Takamatsu, Japan Beijing, China • Use MPC to achieve straight and • The establishment of dynamic model square trajectory tracking in Gazebo on underwater spherical robot. platform. • The experimental verification and • Design an adaptive parameter analysis of the dynamic model. adjustment method. • The hydrodynamic characteristics of • Ensure the continuity of the control the spherical robot under different output of the water jet motor, disturbance flow. The Amphibious Spherical Robot The Underwater Spherical Robot

IEEE ICMA 2021 Conference Digest

MP2-5: Organized session: New Mobile Mechanism and Control

Session Chairs: Shoichi Maeyama, Kagawa University Toshinobu Takei, Hirosaki University Online Conference Room 5, UTC+9(Tokyo Time): 15:15-16:45, Monday, 9 August 2021

MP2-5(1) 15:15-15:30 MP2-5(2) 15:30-15:45

Power-assisted Three-wheel Carrier Cart With Position Control for Half-Drone Wheeled Inverted Pendulum Robot Stair-climbing Ability Rintaro Hoji1, Shoichi Maeyama1, Takuro Kono1, Toshinobu Takei2 and Shin’ichi Yuta3 1Intelligent of Mechanical System Engineering, Kagawa University, Takamatsu, Japan Isaku Nagai*, Daisuke Kijihana**, and Keigo Watanabe* 2Graduate School of Science and Engineering, Hirosaki University, Hirosaki, Japan *Division of Industrial Innovation Sciences, Okayama University, Japan 3SIT Research Laboratories, Shibaura Institute of Technology, Minato-ku Japan **Altech Corporation, Japan • Development of a half-drone wheeled inverted pendulum robot that integrates • We propose a power-assisted carrier wheels and a drone cart using Tri-star wheels to reduce Inverted pendulum Grounding mode mode the burden of carrying luggage. • Driving in various environments with two operation mode (Inverted pendulum • The user can move the cart with a mode and Grounding mode) constant force regardless of load weight and terrain conditions. • Proposal of the integrated position control system on the inverted • Without a step sensor, automatic pendulum mode for the first prototype switching between planar movement robot. and stair climbing is achieved. Power-assisted carrier cart • Verification of the system by simulation The first prototype half-drone and experiment using the real robot wheeled inverted pendulum robot

MP2-5(3) 15:45-16:00 MP2-5(4) 16:00-16:15 Application of maximum hands-off Study on Propulsion of a distributed control to a quadrotor group Snake Robot with Torque Propagation Kimiko Motonaka1, Takuya Watanabe1, Yuhwan Kwon2, Masaaki Nagahara3, and Seiji Miyoshi1 Based on Modified Curvature Derivatives 1Kansai University, Osaka, Japan 2Nara Institute of Science and Technology, Nara, Japan Yongdong Wang, Tetsushi Kamegawa and Akio Gofuku 3The University of Kitakyushu, Fukuoka, Japan Graduate School of Interdisciplinary Science and Engineering in Health Systems Okayama University, Okayama, Japan

• We apply the maximum hands-off • The idea of torque propagation converts the distributed control to the multi-agent angular space to the torque space by system constructed by the proportional gain based on the local curvature quadrotors. derivatives(i.e., joint angle) of the robot. • In simulation and actual experiment, • The idea of using torque propagation for snake four quadrotors at different altitudes robot propulsion is modified and validated. finally converged to the same • We conduct several sets of experiments and altitude using the state of the compare the proportional gain of the new algorithm. neighboring quadrotor. Illustration of the snake robot's propulsion by Environment of an actual experiment modified curvature derivatives in simulation

MP2-5(5) 16:15-16:30 MP2-5(6) 16:30-16:45 Prototype of wheeled stilts-type personal Development of Weeding Turtle Robot micro-mobility for Paddy Fields Issa Nakamura, Tatsuya Kato, and Hirofumi Ohtsuka Toshinobu Takei*, Masahiro Takeshita, and Akira Torige Electronics and Information Systems Engineering Course, *Faculty of Science and Engineering, Hirosaki University,Hirosaki, Aomori, Japan National Institute Technology, Kumamoto Collage, Kumamoto, Japan • By separating the left and right, it is • Recently, weeding robots have been possible to move even in steps and to attracting attention due to the straddle the dent, not only forward and growing demand for organic backward in flat space. farming. • Inverted pendulum control is applied to • 3 types of prototypes and fin parts each of the left and right sides. were designed. • The length, width, and depth of the • Perform weeding operation using mobility are 1480 mm, 200 mm, and only reciprocating motion of the 110 mm, respectively, and the weight of four legs and waist. mobility is 3.5 kg for both left and right. Wheeled stilts-type • The performance was verified in a • Low height-cylindrical wheel with personal micro-mobility Weeding Turtle Robot ． pseudo paddy field. inside motor 20

IEEE ICMA 2021 Conference Digest

MP3-5: Organized session: Biomimetic Measurement and Control in Robotics

Session Chairs: Keigo Watanabe, Okayama University Fusaomi Nagata, Sanyo-Onoda City University Online Conference Room 5, UTC+9(Tokyo Time): 17:00-18:30, Monday, 9 August 2021

MP3-5(1) 17:00-17:15 MP3-4(2) 17:15-17:30 A Range-finding System Using Multiple Lasers Production of a Small-sized Tandem Rotor for an Underwater Robot with Pectoral-fin Aircraft with Two Tiltable Coaxial Rotors Propulsion Mechanisms and Improving Its and Its Experiments Accuracy by a Gimbal Mechanism Hideaki Komura, Keigo Watanabe, and Isaku Nagai Takashi Utsumi, Keigo Watanabe, and Isaku Nagai Graduate School of Natural Science and Technology, Okayama University, Okayama, Japan Graduate School of Natural Science and Technology,Okayama University Okayama, Japan • Reduce the size of the aircraft to • Show a manta robot developed in our make it suitable for confined spaces. laboratory to be used this time. • Change a servo motor to a multi- • Explain a range-finding system using revolutional type to remove the multiple lasers based on a ray tracing constraint of rotational angle. method, and a gimbal mechanism. • Verify whether the manufactured • Conduct some experiments in the air tandem rotor aircraft can move and the water to demonstrate the Gimbal mechanism and horizontally in any direction on the usefulness of the proposed method. a camera module ground as designed. Small-sized Tandem Rotor Aircraft

MP3-5(3) 17:30-17:45 MP3-5(4) 17:45-18:00 Experimental Verification on Visual Feedback Control and Transfer Learning-Based CNN the Rotational and Sway Control of a Suspended for a Pick and Place Robot on a Sliding Rail Fusaomi Nagata1), Kohei Miki1), Keigo Watanabe2) and M.K. Habib3) Horizontally Movable Multi-Rotor 1)Sanyo-Onoda City University, 2)Okayama University, 3)American University in Cairo Kazuya Miyamoto, Shinsuke Kanda, Keigo Watanabe, and Isaku Nagai • CNN&SVM design tool is developed. Natural Science and Technology, Okayama University, Okayama, Japan • A transfer learning-based CNN is designed using the tool to estimate the • Study on a new control method for orientations of workpieces. swing and rotation in suspending • A pick and place robot on a sliding and transporting a load. rail is introduced while implementing • Control a wire tip directly using a a visual feedback controller and the horizontally movable multi-rotor. transfer learning-based CNN to cope • Demonstrate rotational and sway with wide range working area. control using a prototype when • The visual feedback controller enables suspended with no loads. the robot to perform pick and place Horizontally Movable Multi-Rotor Pick and place robot with a CNN task with no calibration. fixed on a sliding rail for wide range working space.

MP3-5(5) 18:00-18:15 MP3-5(6) 18:15-18:30 A Probabilistic models fusion based contact Design of a Novel Cable-Driven 3-DOF Series- detection for quadruped robot Parallel Wrist Module for Humanoid Arms Xuesheng Li, Yage Shen*, Cong Luo, Qiwei Xu Xinlei Chen, Chen Li Zhihao Liang, Bin Wang, Yaowei Song, Tao Zhang, Chaoqun Xiang and Yisheng Guan* School of Aeronautics and Astronautics Delu Dynamics Technology Co. Ltd. Guangdong University of Technology（GDUT） University of Electronic Science and Tech of China Guangzhou, Guangdong, China. Chengdu, Sichuan, China Chengdu, Sichuan, China • This paper proposes a probabilistic • The modular design of the wrist improves the models fusion based contact humanoid arm’s versatility and extensibility. detection for quadruped robot. • The wrist module size and DOF configuration • The proposed algorithm has the are consistent with humans. property of strong reliability and • The cable-driven method is used to ensure high accuracy, and does not require wrist lightweight and flexibility. complex sensors. • The design of the cable-driven wrist module • With the proposed contact detection Unitree A1 robot traverses stairs allows for decoupling of the 3-DOF movement. algorithm, Unitree A1 can traverse • Wrist forward and inverse kinematics of the unstructured terrain, like stairs and 3-DOF series-parallel structure are analyzed. slope. 3-DOF Wrist Module

IEEE ICMA 2021 Conference Digest

MP1-6: Signal and Image Processing (IV)

Session Chairs: Haiyuan Wu, Wakayama University Junchao Zhu, Tianjin University of Technology Online Conference Room 6,, UTC+9(Tokyo Time): 13:30-15:00, Monday, 9 August 2021

MP1-6(1) 13:30-13:45 MP1-6(2) 13:45-14:00 Identification of Binge Drinkers via Convolutional Neural Simulation Research of One-Dimensional Network and Support Vector Machine Guangfei Li, Sihui Du, Jiaxi Niu, Zhao Zhang, Tianxin Gao, Wuyi Wang, Magnetic Particle Imaging Chiang-Shan R. Li and Xiaoying Tang Xiaojun Chen, Xiao Han, Xiaolin Wang, and Xiaoying Tang Department of Biomedical engineering, Beijing Institute of Technology，Beijng, China School of Life Science, Beijing Institute of Technology NO. 5,Zhongguancun South Street, Haidian District, Beijing 100081, China

•Deep learning would address how well • Research of one-dimensional MPI the neural and psychosocial markers simulation is conducted. distinguish binge and non-binge drinkers. • The reconstruction is realized by •We built SVM, 2D-/3D-CNN models of the conjugate gradient method gray matter volumes to distinguish binge and non-binge drinkers. based on the system matrix. •Rule-breaking behavior and cerebellum • Three factors on reconstruction GMVs showed greatest difference performance are also discussed: between the two groups. the particle size, the sample

concentration and the code spacing Fig. 1 The basic principle of MPI: a) the drive field b) on the reconstruction accuracy. the magnetization curve c) the magnetization over time d) the spectrum of voltage signal. Workflow of 2D-CNN construction

MP1-6(3) 14:00-14:15 MP1-6(4) 14:15-14:30 Three-dimensional Ambient Perception Method A Novel Target Detector of Marine Radar Based Based on Binocular Stereo Visual and Structural on HOG Feature Zhizhong Lu, Yue Shi Light Intelligent Systems Science and Engineering, Harbin Engineering University Qi Zeng,Junchao Zhu,Baofeng Zhang, Huifeng Cao,Chang Jia Harbin, Heilongjiang China Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems, Tianjin University of Technology • Read the original radar file. Tianjin, China • An improved gray gravity center • Extract radar image feature extraction algorithm is used to information. extract the center of grating fringe. • Use machine learning methods to • The left and right images are classify. matched by a sweeping grating • Separate the target from image to matching method based on polar obtain the target binary image. The Original Radar Image constraint.

• 3D reconstruction of the scanned Binocular stereo vision experimental object platform

MP1-6(5) 14:30-14:45 MP1-6(6) 14:45-15:00

Directional Navigability Analysis of Local Gravity Research on Vehice Motion Control Based on Map Based on Selective Ensemble Learning Electromyography Signal Collected and Fangming Li1, Dan Zhao1, and Fanming Liu2 1. The 29th research institute of CETC, Chengdu, China 2. College of Automation, Harbin Engineering University. Processed by MCU and Computer Platform Harbin, China Baofeng Gao* , Binkai Yang , Kaijun Guo, Tao Chen , Mingliang Zheng • Aiming at the problems that human factors have great Beijing Institute of Technology, Beijing, China influence on navigability feature extraction and threshold setting of directional navigability analysis of local gravity maps, a directional navigability analysis method based on Log- Gabor filter banks and selective ensemble learning is proposed. • EMG is a kind of bioelectric signal, • First, directional navigability feature maps in five scales and which is distinguishable. eight typical directions are established by using the Log-Gabor filter banks. Second, a selective ensemble learning based on • We designed the app-assisted binary gravitation field algorithm and optimized voting scheme is used as directional navigability analysis model. Finally, control and display on the Android simulation experiments are conducted to verify the proposed method. side of the mobile phone through the • Results show that the proposed method can effectively avoid open source platform. the blindness when artificially extracting directional navigability features and evaluation threshold setting, and the • We finally control the operation of directional navigability analysis for local gravity map can be achieved automatically. Compared with random forest and the vehice through EMG. Test chart of signal selective ensemble learning based on BPSO, the accuracy of Directional navigability analysis model based on acquisition experiment the proposed method is highest. selective ensemble learning 22

IEEE ICMA 2021 Conference Digest

MP2-6: Vision System, Robotic Vision

Session Chairs: Jian Guo, Tianjin University of Technology Xiaoliang Jin, Kagawa University Online Conference Room 6,, UTC+9(Tokyo Time): 15:15-16:45, Monday, 9 August 2021

MP2-6(1) 15:15-15:30 MP2-6(2) 15:30-15:45 Towards a Micron-level Line Structured Light Instance Segmentation of Low-texture Industrial Yue Guo, Yanmei Li, Haitao Song, Wenhao He, and Kui Yuan Parts Based on Deep Learning Institute of Automation, Chinese Academy of Sciences, Beijing, China School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China Yue Zhang, Zelin Shi, Chungang Zhuang* Beijing College of Finance and Commerce, Beijing, China Mechanical Engineering，Shanghai Jiao Tong University Shanghai, China • We propose a method for generating datasets for industrial parts in a physical simulation environment. • We propose a neural network based

Thickness Measurement Framework on point cloud feature fusion for instance segmentation of industrial • A semantic centroid method to localize central points of line parts. segments. • We conducted an instance • Several detailed factors that may ruin measurement performances are segmentation experiment on the data The Instance Segmentation Results analyzed. set, and the effect was very good. • General conclusions that are also beneficial for other high-precision applications based on line structured lights.

MP2-6(3) 15:45-16:00 MP2-6(4) 16:00-16:15 An Improved QPSO Algorithm Based on EXIF for YOLO Target Detection Algorithm with Camera Self-calibration Deformable Convolution Kernel Hui Wang, Shuai Zhang, Lijun Yu and Ce Shi Pengxiao Bao1,2 , Feng Gao1,2 , Liwei Shi1,2* , Shuxiang Guo1,2,3* College of Intelligent Systems Science and Engineering, Harbin Engineering University, Harbin, • Binocular vision technology is an China important branch of computer vision • This paper proposes a YOLO target technology. detection algorithm with deformable • We propose a new method of camera convolution kernel. self-calibration by improving an • The network can adaptively change existing QPSO algorithm with the the receptive field of feature points EXIF information of digital camera according to the shape of the target, photos. thereby extracting features more • The improved algorithm has an effectively. Improved QPSO algorithm average accuracy of above 94% and • The algorithm can effectively Comparison of target detection the prediction accuracy of the improve the accuracy of target results camera’s focal length is above 95%. detection while guaranteeing detection speed.

MP2-6(5) 16:15-16:30 MP2-6(6) 16:30-16:45 Research on CAMshift Algorithm Based on Improved MB-LBP Feature Extraction Algorithm Feature Matching and Prediction Mechanism Based on Reduced-dimensional HOG Hui Wang, Quchang Zhang, Lijun Yu and Zhiqiang Wang Lijun Yu, Qing Li, Hui Wang and Ce Shi College of Intelligent Systems Science and Engineering, Harbin Engineering University, College of Intelligent Systems Science and Engineering, Harbin Engineering University, Harbin 150001, China Harbin 150001, China • This paper proposes an improved • This paper proposes an improved MB- CAMshift Algorithm Based on Feature LBP feature extraction algorithm based Matching and Prediction Mechanism. on dimensionality reduction HOG. • The CAMShift algorithm is optimized • Through feature extraction of face by ORB feature matching algorithm CAMshift Algorithm image and hand image, it is verified Image feature extraction and Kalman filter to improve the that the algorithm in this paper is tracking effect when the target is superior to the traditional algorithm. disturbed by complex background and • Compared with MB-LBP and HOG occlusion. feature extraction algorithms, the algorithm in this paper can extract • The improved algorithm can accurately Image 3D mesh map track the moving targets in the video effective features more completely. stream in real time. Improved Algorithm 23

IEEE ICMA 2021 Conference Digest

MP3-6: Sensor Design, and Novel Sensing Systems

Session Chairs: Jian Guo, Tianjin University of Technology Xiaoliang Jin, Kagawa University Online Conference Room 6,, UTC+9(Tokyo Time): 17:00-18:30, Monday, 9 August 2021

MP3-6(1) 17:00-17:15 MP3-6(2) 17:15-17:30 Extraction of Pedestrian Position and Attribute Motion Measurement and Segmentation Information Based on the Integration of LiDAR Toward Automated Sewing Operations and Smartphone Sensors Yaqiang Mo, Yuto Nakagawa, Kotaro Nagahama and Kimitoshi Yamazaki Shinshu University Zhengshu Zhou, Saya Kitamura, Yousuke Watanabe, Shunya Yamada, and Hiroaki Takada Nagano, Japan Institute of Innovation for Future Society, Nagoya University Nagoya, Aichi, Japan • We propose a measurement system that acquires the operation skills of a • LiDAR & GPS matching to obtain human operator in a sewing task. the coordinates and probability of pedestrians’ id and location • We propose a force sensor to measure the force on the fingertip • Human feature filters to exclude without losing the sense of feeling. non-pedestrian clusters • we established automated • IMU and GPS fusion to analyze the classification and segmentation speed and course of the pedestrians The Proposed Method to Extract methods for the motions acquired to improve matching accuracy Pedestrian Position and Attribute from a simulated sewing operation. Process of simulated Information sewing operation

MP3-6(3) 17:30-17:45 MP3-6(4) 17:45-18:00

A Multi-sensor Intelligent Surface Garbage An Energy-efficient and Redundancy-reduced Cleaning Robot Protocol of WSN under Non-Uniform Deployment Yunliang Wang,Yiwen Zhao,Yanjuan Wu,Sai Zhang, and Jian Wang Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems, Tianjin Yuxiu Lin, Xiaomei Xie*, Mingzhu Wei, Tonghui Zeng, Xin Chen, Xiangfei Wu, Omar Mechali School of Aeronautics and Astronautics, University of Electronic Science & Technology of China University of Technology Chengdu, Sichuan Province, China Tianjin, China • Divide the WSN under non-uniform • The Cleaning Robot has multiple deployment into several regions by sensors. DPC algorithm. • The robot is able to approach surface • Sleep redundant nodes by coverage garbage automatically based on rate and the number of neighboring information collected by multiple nodes to reduce redundancy and save sensors. energy. • The robot can accurately identify • Sort the weight values of nodes in the The Number of Dead nodes garbage and can run stably and flexibly. same region in descending order and • The control system realizes the The Cleaning Robot then select a suitable number of intelligent control of unmanned vehicle. cluster nodes.

MP3-6(5) 18:00-18:15 MP3-6(6) 18:15-18:30 Information Manifold and Ricci Curvature Research on the State Acquisition Method Mo Tao1,2, Shaoping Wang1*, Hong Chen2 , Zhi Liu2 and Yi Lei2 of Anti-seismic Support and Hanger 1.School of Automation Science and Electrical Engineering, Beihang University, Beijing , China Yingxin Wang, Junchao Zhu, Qingliang Zhu,Yongcheng Zhu 2.Wuhan Second Ship Design and Research Institute, Wuhan , Hubei Province, China Engineering Research Center of Optoelectronic Devices & Communication Technology, Tianjin University of Technology,Tianjin, China • This paper discusses affine immersion of • Data monitoring system based on multivariable Gaussian statistical embedded and sensor technology manifolds for multi-sensor networks. design. • The Ricci curvature of information space • The sensing equipment installed on is calculated, the degree of curvature of the seismic support and hanger the manifold to represent the changing senses the change signal, which can trend of information. collect real-time data on the force • The potential function that can represent state, deformation and sway law of the manifold, the high-dimensional the seismic support and hanger, and manifold of the sensor network upload the data to the host computer. Data collection systems information is embedded into the • Improve the safety performance of Euclidean space. seismic supports and hangers. 24

Tuesday August 10, 2021

Morning Sessions

TA1-1 Intelligent Mechatronics and Application (I) TA1-2 Manipulator Control and Manipulation (I) TA1-3 Mobile Robot System (I) TA1-4 Biomimetic Systems TA1-5 Industrial, Manufacturing Process and Automation TA1-6 Robot Navigation and Control Algorithm TA2-1 Intelligent Mechatronics and Application (II) TA2-2 Manipulator Control and Manipulation (II) TA2-3 Mobile Robot System (II) TA2-4 Mobile Robot System (V) TA2-5 Elements, Structures, and Mechanisms TA2-6 Organized session: Soft Robotics

Tuesday August 10, 2021

Afternoon Sessions

TP1-1 Medical, Biomedical and Rehabilitation Systems (I) TP1-2 Organized session: Medical Robots for Minimal invasive Surgery (I) TP1-3 Mobile Robot System (III) TP1-4 Human-System Interaction and Interface TP1-5 Intelligent Biomedical Instrument Technology TP2-1 Medical, Biomedical and Rehabilitation Systems (II) TP2-2 Organized session: Medical Robots for Minimal invasive Surgery (II) TP2-3 Mobile Robot System (IV) TP2-4 Medical, Biomedical and Rehabilitation Systems

IEEE ICMA 2021 Conference Digest TA1-1: Intelligent Mechatronics and Application (I)

Session Chairs: Takashi kei Saito, Akita Prefectural University Peng Shi, Kagawa University Online Conference Room 1, UTC+9(Tokyo Time): 9:30-11:00, Tuesday, 10 August 2021

TA1-1(1) 9:30-9:45 TA1-1(2) 9:45-10:00 Deep Q-learning for Control: Technique and Implementation Considerations on a Mechanism and Design of Tableware Tidying-up Physical System: Active Automotive Rear Spoiler Case, on ICMA 2021 Velazquez Espitia Victor Miguel, Gonzalez Gonzalez Jose Angel, Mata Juarez Omar, Ponce Pedro and Molina Arturo Robot for Self-Service Restaurant

Intelligent of Mechanical System Engineering, Kagawa University Deheng Zhu, Hiroaki Seki, Tokuo Tsuji, Tatsuhiro Hiramitsu Takamatsu, Japan Division of Mechanical Science and Engineering, Kanazawa University Kanazawa, Japan • Active automotive rear spoiler with two DOF that increases stability • Automated tableware tidying-up when taking turns. robot is developed. • Step by step Deep Reinforcement • Sort tableware on conveyor belt. Learning implementation guide for • The robot arm is parallel type, agent’s successful training. air driven, simple mechanism • A more effective, dynamic, control and the system consists of of mechanical systems, able to multiple arms. adapt to changing scenarios and Active Automotive Rear Spoiler • Leftover food and garbage is unforetold variables. also processed. Tableware Tidying-up Robot • Uses the advantages of AI in the • Mechanism is made and tested. control of mechanical systems.

TA1-1(3) 10:00-10:15 TA1-1(4) 10:15-10:30 Intraoperative Kinematic Analysis of Autonomous Injection molding of cell membrane perforator Cornea Suturing Surgical Robot for Keratoplasty with photochemical perforation function and Xiaojing Feng, Xiaodong Zhang, Xiaojun Shi Li Li School of Mechanical Engineering, Department of Ophthalmology, surface modification Xi’an Jiaotong University First Affiliated Hospital of Xi’an Jiaotong University Kohei Kobayashi, Kenji Imamura, Kiriro Suzuki, Koki Takemasa, and Takashi Kei Saito Shaanxi, China Shaanxi, China Department of Intelligent Mechatronics, Akita Prefectural University • Intraoperative procedures Yurihonjo, Japan

including needle insertion and 11mm Upper Layer: • Our photochemical cell membrane Iron powder the surgeon’s knot are analyzed. perforation method exhibits high mixed with TPE Middle Layer: • Quantitive kinematic model of efficiency, but requires a cell NaCl solution 5mm needle insertion and knot tying membrane perforator, which LowerLayer: TPE is built according to requires 4 h for production. Bottom Layer: Hp-OcOH intraoperative procedures. • We present a process using an solution infiltrated • Cornea suturing robot is injection molding machine for into TPE designed and the suturing manufacturing new perforators for Thermoplastic Elastomer Based procedure is reproduced based Keroatoplastic robot mass production, while preserving Cell Membrane Perforator on this robot. the perforating function. For Injection Molding

TA1-1(5) 10:30-10:45 TA1-1(6) 10:45-11:00 Classification and detection method of Blood A Novel Remote Control Method Oriented to lancet based on VGG16 network Underwater Manipulators Feng Zhao1, Baofeng Zhang1，Zhili Zhang2, Xinghui Zhang2, Chunyu Wei3 Zhiheng Zheng*, Jirong Xie, Tao Su, Xinguang Li and Yu Ni 1Tianjin University of Technology. 2Tianjin Sino-German University of Applied Sciences. China Ship Scientific Research Center (CSSRC) 3Tianjin Yuhang Runming Technology Development Co, Ltd. Wuxi, Jiangsu, China Teijin, China • Intuitive operation of multi-DOF • The Canny detection algorithm is used to multi-target detection on blood lancet underwater manipulators. images. • Position and orientation mapping • The parameters of the pre-trained vgg16 scheme of the end effector. model are fine tuned, and the trained • Color ball detection of the control rod. network is used for blood lancet classification detection. • Multi-vision positioning system. • The experiment compared GoogLeNet • Lab component analysis under and ResNet two algorithms, the results different illumination conditions. Control of multi-DOF Underwater Manipulators show that the VGG16 network has better performance than others. The Experimental Process

IEEE ICMA 2021 Conference Digest TA2-1: Intelligent Mechatronics and Application (II)

Session Chairs: Yan Zhao, Beijing Institute of Technology Peng Shi, Kagawa University Online Conference Room 1, UTC+9(Tokyo Time): 11:15-12:15, Tuesday, 10 August 2021

TA2-1(1) 11:15-11:30 TA2-1(2) 11:30-11:45 Pallet Detection and Estimation for Development of Interface for Turning Motion Fork Insertion with RGB-D Camera Assist of Power Assist System using 3D sensor Ryosuke Iinuma1, Yusuke Hori1, Hiroyuki Onoyama1, Yukihiro Kubo1, Takanori Fukao2,1 Hironao Yamada, Takahiro Ikeda, Satoshi Ueki, Katsutoshi Ootsubo 1Ritsumeikan University, Shiga, Japan, Intelligent of Mechanical System Engineering, Gifu University, Gifu, Japan 2University of Tokyo, Tokyo, Japan. • A new interface for turning • We developed an autonomous forklift motion assist of power assist system for the fork insertion in system using a 3D sensor was outdoor environments where the developed and its effectiveness ground is inclined. was verified. • Pallet is detected and its position and • Devised a method that utilizes orientation are estimated by the result Autonomous forklift the correlation between the of semantic segmentation and point direction of the upper body of cloud. the operator performing the • Experimental results showed the turning motion and the system achieved the fork insertion on movement of the operation unit. Power Assist System the situation assuming the real task. Fork insertion

TA2-1(3) 11:45-12:00 TA2-1(4) 12:00-12:15 Design and Simulation of Lift Feedback Fin Stabilizer A Game-based fault-tolerant path planning Control System Based on Hardware-in-the-Loop System algorithm for space manipulator Lihua Liang, Jianfeng Li, Zhiwen Le and Songtao Zhang College of Intelligent Systems Science and Engineering, Harbin Engineering University Shuo Wang, James Zhiqing Wen, Decheng Zhou and Yujin Wu Harbin, Heilongjiang, China Engineering Research Center for Intelligent Robotics, Jihua Laboratory Foshan, China

• Proposed a lift feedback fin stabilizer • A game-based RRT* algorithm is proposed to find a reasonable path in both Cartesian control system based on measuring space and joint space. the pressure difference between the • In this algorithm, the rasterized grids are fin driving cylinders. used to improve the convergence rate, while a game-based steer mechanism is designed • A hardware-in-the-loop simulation to balance the objectives with different system for ship rolling is built. dimensions. • Simulation results demonstrate that the • Results provide a better simulation proposed algorithm can search a path Path planned by game-based environment for the design of fin satisfies the requirements from two spaces. stabilizer control system. RRT* algorithm Hardware-in-the-Loop System

IEEE ICMA 2021 Conference Digest

TP1-1: Medical, Biomedical and Rehabilitation Systems (I)

Session Chairs: Jin Guo, Beijing Institute of Technology Dongdong Bu, Beijing Institute of Technology Online Conference Room 1, UTC+9(Tokyo Time): 13:30-15:00, Tuesday, 10 August 2021

TP1-1(1) 13:30-13:45 TP1-1(2) 13:45-14:00 Development of a Knee Extensors Training Prototype of the Record and Evaluation System System with Active Resistance Component of the Locomotion Experience Zixi Gu, Ko Matsuhiro, Jiei Yanagi, Kazuhiko Mizukami, Ryuya Watanabe, Hiroaki Eto, Using the Electric Mobility Support Device Sarah Cosentino, Atsuo Takanishi Rikuya Tsujinaka, Toshihiko Yasuda, Yasutaka Nishioka and Mitsuhiro Yamano Faculty of Science and Engineering, Waseda University Department of Mechanical System Engineering , University of Shiga prefecture Shinjuku City, 1690072, Tokyo, Japan Hikone, Shiga, Japan • To better integrate knee extensor • Locomotion experience is very important in child development. We training in daily life we are are developing the electric mobility support device for physically developing a training device which handicapped children. can apply additional mechanical • We propose a system recording the locomotion experience using load on the knee joint during daily the locomotion device and evaluating the normal walking. development of the children, constructed by • In this study, we improved the the omnidirectional camera and the micro- training device with active computer with SD card. resistance component to adjust the Prototype of Training device training load in real-time. • We also propose two methods indicating the improvement of the operation ability. Electric Mobility Support Device

TP1-1(3) 14:00-14:15 TP1-1(4) 14:15-14:30 A New Anthropomorphic Thumb Configuration ESPI: Dynamic Safety Evaluation Index for With Passive Finger Torsion Human-Exoskeleton System Ang Ke, Jian Huang, Jiping He Fashu Xu, Xinyue Gu, Jing Qiu, Rui Huang*, Hong Cheng School of Artificial Intelligence and Automation Huazhong University of Science and Machine Intelligence Institute, School of Automation Engineering Technology, Wuhan, China University of Electronic Science and Technology of China，Chengdu, China In this paper, the Enhanced Stability Pyramid Index (ESPI) is proposed to evaluate the system's safety, • Proposed a new thumb structure with especially in a dynamic situation. To incorporate ‘311’ joint configuration, in which the dynamic information of the system, this index employs TM joint has three joints. eXtrapolated Center of Mass (XCoM) instead of the center of mass (CoM). Meanwhile, Time-to-Contact • The couple finger torsion improves the (TTC), the urgency of safety, is utilized as an automatic direction of the fingertips on the weight assignment factor instead of the traditional abdomen during gripping. manual one. The system's current safety state can also be • Optimal design of the coupled TM represented by scalar and vector to provide more The Enhanced information for the exoskeleton and pilot. Finally, the joint based on thumb dexterity. Stability Pyramid indicator's effectiveness is verified by the human- Index (ESPI). exoskeleton walking simulation controlled by the linear The Prosthetic Hand inverted pendulum in Gazebo.

TP1-1(5) 14:30-14:45 TP1-1(6) 14:45-15:00 Study on Tracking Control of Vascular Design of a Variable Stiffness Series Mechanism Qiang Fu, Xun Li, Jian Guo, Shuxiang Guo, Zhuocong Cai and Jiajun Fu Interventional Surgical Robot based on Tianjin Key Laboratory for Control Theory & Application in Complicated Systems and Intelligent Robot Laboratory, Tianjin University of Technology Autocoupling PID Tianjin, China Shuxiang Guo1,2, Zefa Sun1 and Jian Guo1* 1Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems • Series actuator with variable and Intelligent Robot Laboratory, Tianjin University of Technology,Tianjin, China stiffness. 2 Department of Intelligent Mechanical Systems Engineering Faculty of Engineering, • Estimation of interaction force Kagawa University, Kagawa, Japan without force sensor. •In order to solve the problem of • Simpler stiffness characteristic master-slave control of vascular design. interventional surgical robot, the traditional PID controller was analysed, and the self-coupling PID controller was designed. Autocoupling PID control block diagram The Variable Stiffness Series Mechanism

IEEE ICMA 2021 Conference Digest

TP2-1: Medical, Biomedical and Rehabilitation Systems (II)

Session Chairs: Jin Guo, Beijing Institute of Technology Dongdong Bu, Beijing Institute of Technology Online Conference Room 1, UTC+9(Tokyo Time): 15:30-17:00, Tuesday, 10 August 2021

TP2-1(1) 15:30-15:45 TP2-1(2) 15:45-16:00

Design and Characterization of a Cable-Driven Development of a walking-trajectory Series Elastic Actuator Based Torque measurement system Nina Tajima, Koichiro Kato, Daigo Kurokawa, Kanako Amano, and Yuka Kato Transmission for Back-Support Exoskeleton and Nobuto Matsuhira Hongpeng Liao, Hugo Hung-tin Chan, Fei Gao, Xuan Zhao, and Wei-Hsin Liao Department of Engineering Science and Mechanics, Division of Mathematical Sciences, The Chinese University of Hong Kong, Hong Kong, China Shibaura Institute of Technology Tokyo Woman’s Christian University Koto, Japan Suginami, Japan • A cable-driven series elastic actuator • Aiming to improve the (SEA) based torque transmission measurement accuracy of walking- system for back-support exoskeleton. trajectory of multiple people using • Compliance assistive torque LRF. LRF transmission and large assistive torque • Use Gaussian process regression to transmission capability are achieved. predict and correct for missing data. • Experimentally characterize the • Measurement of trajectories of The results of each experiment designed torque transmission system multiple persons within a radius of measured by the conventional and show the capability of providing 4m for up to 10 persons. system and the predictive system timely and sufficient assistive torque. Kinetics of Transmission

TP2-1(3) 16:00-16:15 TP2-1(4) 16:15-16:30 A New Collision Detection Algorithm for Study on Contact Force Prediction for the Vascular Interventional Surgery Vascular Interventional Surgical Robot based on Simulation Training System Parameter Identification Jian Guo1,2 ,Zhentao Wang 1 ,Shuxiang Guo1* Shuxiang Guo1,2, Xianguo Liao 1 and Jian Guo1* 1Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems 1Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems and Intelligent Robot Laboratory, Tianjin University of Technology,Tianjin, China and Intelligent Robot Laboratory, Tianjin University of Technology,Tianjin, China 2Intelligent Mechanical Systems Engineering Department Faculty of Engineering, 2Department of Intelligent Mechanical Systems Engineering Faculty of Engineering, Kagawa University, Kagawa, Japan Kagawa University, Kagawa, Japan • In this paper,A new soft tissue •In this paper, the contact force is hybrid collision detection modeled in the master manipulator side, algorithm is proposed. Through which is used to predict the contact the two-stage detection, a lot of force when the slave manipulator side redundant calculation in the contacts the real tissue. Autoregressive original collision detection least squares method is used for algorithm is eliminated, and the parameter identification. This method response time of collision Collision method simulation can perform force prediction well. detection is greatly shortened. diagram Force model prediction system

TP2-1(5) 16:30-16:45 TP2-1(6) 16:45-17:00

Design and Evaluation of a Novel Magnetoactive Design of a Novel Drug-Delivery Capsule Robot Biopsy Capsule Endoscope Robot 1,2 1 1 1 Shuxiang Guo , Yaqi Hu , Jian Guo , Qiang Fu Jian Guo1,2 ,Xinyi Liu 1 ,Shuxiang Guo1* ,Qiang Fu1* 1Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems and 1 Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems and Intelligent Robot Laboratory, Intelligent Robot Laboratory, Tianjin University of Technology,Tianjin, China Tianjin University of Technology, Tianjin, China 2 2 Department of Intelligent Mechanical Systems Engineering, Faculty of Engineering, Intelligent Mechanical Systems Engineering Department Faculty of Engineering, Kagawa Kagawa University, Kagawa, Japan University, Kagawa, Japan • In this paper, a novel biopsy capsule • A novel micro in-pipe robot with a endoscope robot controlled by an propeller is proposed in this paper, it is a external magnetic field is designed. robot that can release drug. According to the working principle of • Fluid simulation model is established. the CAM structure, the biopsy module • The feasibility of the capsule robot is is proposed and manufactured. At the verified through experiments. same time, the control system of the Schematic diagram of the robot is introduced, including the Biopsy robot based on CAM micro-robot driving system of the capsule robot and structure the biopsy process of the biopsy module.

IEEE ICMA 2021 Conference Digest TA1-2: Manipulator Control and Manipulation (I)

Session Chairs: Tatsuo Arai, Akita Prefectural University He Yin, Beijing Institute of Technology Online Conference Room 2, UTC+9(Tokyo Time): 9:30-11:00, Tuesday, 10 August 2021

TA1-2(1) 9:30-9:45 TA1-2(2) 9:45-10:00

A Six-degree-of-freedom Face Tracking Method for Kinematics Redundancy Identification of Non-contact Physiological Detection Robot Arbitrary Serial Robot Haoran Zhang, Bingzheng Fan, Xiaoman Zhang, Hang Zhan, and Xiaojian Li Lingxiao Li, Guojiang Li,Lu Li, Yanan Wang, Baolin Feng Key Laboratory of Process Optimization and Intelligent Decision-Making (Ministry of Education) Institute of intelligent machines, Hefei Institute of Physical Science School of Management, Hefei University of Technology Hefei, China Hefei, Anhui Province, China • The proposed method enables the • Derivation of Error model between robot to actively cooperate with the end-effector and DH parameters patient to perform physiological detection. • Redundant parameter analysis of • A six-degree-of-freedom position- serial robot with different structure. based visual servoing method is • Homogeneous transformation matrix proposed to complete the real-time between robot system and world face tracking task. system • A decoupling P-type controller is designed to make the movement • Self calibration based on plane smoother. constraint. Non-redundant calibration The Real-time Face Tracking

TA1-2 (3) 10:00-10:15 TA1-2(4) 10:15-10:30 Control Method for Robotic Manipulation of Real-Time Pixel-Wise Grasp Detection Based Heavy Industrial Cables on RGB-D Feature Dense Fusion Fangli Mou, Dan Wu Department of Mechanical Engineering, Tsinghua University, Yongxiang Wu, Yili Fu*, and Shuguo Wang Beijing, China State Key Laboratory of Robotics and System, Harbin Institute of Technology Harbin, China • The first attempt to consider and g g z z • An end-to-end grasp detector y y solve the payload effect caused by Fn−1 F4 F3 F2 F1 x x {}Ttool {}Ttool

Ft based on RGB-D fusion for heavy DLOs. (a) Example of a manipulating (b) The MSD model of cable. cable. predicting grasp confidence • A general and effective control g g and pose of each pixel.

Fm−1 F ' F ' F ' F 3 2 1 z 1 z method is presented for robot ' F n−1 F y 2 y • A efficient sampling and x x {}Ttool {}T manipulating heavy cables with high tool (c) The model transformation. (d) The cable model with feature points. matching strategy based on UIS precision. Cable Model algorithm and the index image. • A computationally-efficient method • Our method runs at a real-time for simulating the payload effect speed of 8 ms per image and according to the extracted cable achieves an 97% grasp success features. rate on household object. Control Scheme

TA1-2 (5) 10:30-10:45 TA1-2 (6) 10:45-11:00 Magnetically Driven Rolling Motion for Real-time Gesture Recognition Based on Magnetic Cylindrical Microrobots Improved Artificial Neural Network and sEMG Hongzhe Liao, Xiaoming Liu, Dan Liu, Yu Ning, Qiang Huang, Tatsuo Arai Beijing Advanced Innovation Center for Intelligent Robots and Systems, and School of Signals Mechatronical Engineering,Beijing Institute of Technology, China Wenzhe Zhang, Liguo Shuai and Haoxuan Kan • A magnetic microrobot system is set School of Mechanical Engineering, Southeast University, Nanjing, China up with five electromagnets on • Optimize the original neural network different axes to drive the microrobot. model for gesture recognition. • The microrobot is fabricated by a • Use time-domain features and method based on heat-curing using a convolution results as input features. glass capillary tube as the mold. • Add a dropout layer to the neural • The microrobot can be rotated by an network model. external rotating magnetic field to roll • Use the cross-validation method of on two-dimensional plane. majority voting and consecutive • The microrobot can be controlled to voting to determine the final Real-time Gesture Recognition Model move along a planned trajectory with The Magnetic Microrobot System recognized gestures. high position accuracy.

IEEE ICMA 2021 Conference Digest TA2-2: Manipulator Control and Manipulation (II)

Session Chairs: Habib Maki K, The American University in Cairo He Yin, Beijing Institute of Technology Online Conference Room 2, UTC+9(Tokyo Time): 11:15-12:15, Tuesday, 10 August 2021

TA2-2(1) 11:15-11:30 TA2-2(2) 11:30-11:45 Stabilization of Double Inverted Pendulum (DIP) on Perovskite Photovoltaic Tracking Systems for Efficient a Cart using Optimal Adaptive Sliding Mode Control Irrigation in the Agricultural Field: Morocco Case (OASMC) Study Maki K. Habib and Samuel A. Ayankoso u Yousra Bensouda, Yasmine Boudihaj, Nada Sebti, Hajar Hassina and Yassine Salih-Alj The American University in Cairo, Cairo, Egypt. [email protected] and [email protected] School of Science and Engineering, Al Akhawayn University in Ifrane Ifrane, Morocco • The potential applications of an inverted pendulum is the primary source of motivation for researchers to initiate innovative and new • Optimized efficiency of the solar B panel cells using 2T-PVK/Si development. A DIP system is a nonlinear, X tandems inherently unstable, and underactuated system. It has a single input and multiple outputs. The A • Optimized irrigation mathematical model of this system is normally • Minimized to no pollution formulated either using Newtonian or Euler- The Perovskite Atomic Structure Lagrange dynamic approaches. • Little maintenance • This article aims to analyze the stability of a DIP supported by the development of OASMC and the performance is compared to LQR The DIP on a cart controller.

TA2-2(3) 11:45-12:00 TA2-2(4) 12:00-12:15 Data-Driven Modeling: Concept, Techniques, Challenges and a Case Study 1Maki K. Habib, 1Samuel A. Ayankoso and 2Fusaomi Nagata 1The American University in Cairo, Cairo, Egypt. [email protected] and [email protected] 2Graduate School of Science & Engineering, Sanyo-Onoda City University, Japan • There is a surge in the use of data-driven models in different application domains. Such models are developed using experimental input/output data measured • The proposed method infers an end-to- from real-world systems. Data-driven based end policy for imitation after one video modeling is described through a system demonstration is given. Specifically, a identification process that involves acquiring input-output data, selecting a model class, target recognition module is adopted in estimating model parameters, and then the model architectur. validating the estimated model. • This paper presents the data-driven • The real-world experiments on a modeling as a concept and as a technique. UR10e robot arm are conducted with Also, it presents the challenges and a criteria new scenarios or with new objects, to consider when developing a data-driven Modeling approaches for Fig. The robot (right) imitates to approach a model. Besides the paper introduces a case illustrating the effectiveness of our engineering systems method. container which is shown in the human study demonstration (left).

IEEE ICMA 2021 Conference Digest TP1-2: Organized session: Medical Robots for Minimal invasive Surgery (I)

Session Chairs: Keisuke Morishima, Osaka University Cheng Yang, Beijing Institute of Technology Online Conference Room 2, UTC+9(Tokyo Time): 13:30-15:00, Tuesday, 10 August 2021

TP1-2(1) 13:30-13:45 TP1-2(2) 13:45-14:00 Trajectory Planning of an Underactuated Cable- Resistance Recognition of Moving Guidewire in Vascular Driven Planar Device for the Trunk Interventional Operation Chuqiao Lyu1, Shuxiang Guo1,2*, Youchun Ma1, Yue Wang1, Chenguang Yang1 Lailu Li, Shuoyu Wang, and Guang Yang 1 Key Laboratory of Convergence Biomedical Engineering System and Healthcare Technology, The Ministry of Industry and Information Technology, School of Life Science, Beijing Institute of Technology, School of Systems Engineering, Kochi Unviersity of Technology No.5, Zhongguancun South Street, Haidian District, Beijing 100081, China Kami, Japan 2 Faculty of Engineering, Kagawa University, 2217-20 Hayashi-cho, Takamatsu, Kagawa 760-8521, Japan

• An underactuated cable-driven planar • We try to build a resistance recognition device for the trunk (CAPT) is algorithm in a non-vascular environment. Cable I proposed. Cable II • We divide the moving guidewire into resistance state and non-resistance. • CAPT provides certain trunk motion • We get the Resnet-18 pre-training model restraints dring gait rehabilitation based on extensive X-ray datasets from activities. C2L as the feature network of Faster RCNN.

• System kinematic and dynamic models Omni gait trainer • We train the whole network, and the are established. training results show that the resistance recognition model can give a good • Trajectory planning study based on the precision, which is 95.93%. radau pseudospectral method is Cable-driven planar device • Our method provides a good idea for proposed. for trunk (CAPT) posture restraint combining expert experience and AI （c）. Faster RCNN algorithm in vascular interventional surgery. Resistance Recognition of Moving Guidewire

TP1-2(3) 14:00-14:15 TP1-2(4) 14:15-14:30 A teleoperation control method for vascular Evaluation of a Reinforcement Learning interventional surgery robot Algorithm for Vascular Intervention Surgery Hang Yuan, Nan Xiao*, Mengqi Cheng, Kaidi Wang Fanxu Meng, Shuxiang Guo, Wei Zhou, and Zhengyang Chen Key Laboratory of Convergence Biomedical Engineering System and Healthcare Technology, Key Laboratory of Convergence Biomedical Engineering System and Healthcare Technology, The Ministry of Industry and Information Technology, School of Life Science, Beijing Institute of Technology Beijing Institute of Technology, Haidian District, Beijing 100081, China Beijng, China • A model based on Unity was • Prosed a teleoperation provided to implement the control method. reinforcement learning environment. • The 4G network and • Asynchronous advantage actor-critic data transfer server are method was used to train the virtual used to realize master- model. slave communication. • The model was successfully trained • Carried out three The teleoperating system models in virtual environment and could experiments for system lead to more practical applications. verification. Aortic arch and catheter in Unity

TP1-2(5) 14:30-14:45 TP1-2(6) 14:45-15:00 Mechanism Design, Kinematic and Evaluation of a Clamping Mechanism for Hydrodynamic Simulation of a wave-driven Vascular Interventional Surgery Robotic System amphibious bionic robot Chaochao Shi, Shuxiang Guo Intelligent of Mechanical System Engineering, Kagawa University Zhongyin Zhang, Liwei Shi, Shuxiang Guo, He Yin, Ao Li , Pengxiao Bao, Meng Liu Takamatsu, Japan Key Laboratory of Biomimetic Robots and Systems, Ministry of Education, School of Life • It simulates the twisting of human Science, fingers for the catheter during the Beijing Institute of Technology, Haidian District, Beijing 100081, China actual vascular intervention. • Ray structure design includes • Its main structure uses two mobile that bindiny mechanism, the platforms. general structure of a fish fin, angle adjusting mechanism. • The bottom mobile platform simulates the human wrist and • We added connections and realizes the forward and backward drives to the original model to of the catheter. make Kinematic simulation. • The top platform’s screw adopts • Fin hydrodynamic simulation The wave-driven bionic robot positive and negative threads. The Slave manipulator about pressure and velocity.

IEEE ICMA 2021 Conference Digest TP2-2: Organized session: Medical Robots for Minimal invasive Surgery (II)

Session Chairs: Shuxiang Guo, Akita Prefectural University Cheng Yang, Beijing Institute of Technology Online Conference Room 2, UTC+9(Tokyo Time): 15:30-17:00, Tuesday, 10 August 2021

TP2-2(1) 15:30-15:45 TP1-2(2) 15:45-16:00 A Novel Tremor Suppression Method for Preliminary Method for Reducing Contact Force between Endovascular Interventional Robotic Systems Catheter Tip and Vessel Wall in Endovascular Surgery Xiaoliang Jin1, Shuxiang Guo1,2,3*, Jian Guo3*, Peng Shi1, and Xinming Li1 Xinming Li, Shuxiang Guo*, Peng Shi and Xiaoliang Jin Graduate School of Engineering, Kagawa University 1Graduate School of Engineering, Kagawa University, Takamatsu 761-0396, Kagawa, Japan. Takamatsu, Japan 2Beijing Institute of Technology, Beijing, China. 3Tianjin University of Technology, Tianjin, China. • A novel method based on active • In some special environments, the blood restraint and passive modification vessels are more curved, and it is difficult for tremor suppression in robot- for the catheter to pass through smoothly even with the guidance of the guidewire. assisted system. • In this paper, a preliminary concept was • Comparative experiments and presented to reduce the contact force on simulation results indicate that the the catheter tip and improve the safety of The developed robot-assisted system proposed method is capable of the system. suppressing burst tremor and • The experiments in “Vitro” were carried regular tremor under the operation out to evaluate the proposed method, and the results were indicated that the electro of surgeon. Result for hybrid tremors magnetic force can effectively deflect the tip of the catheter. The schematic diagram of the deflection of catheter tip

TP2-2(3) 16:00-16:15 TP2-2(4) 16:15-16:30 Centerline Extraction Method for Virtual Force analysis of catheter tip influenced by Vascular Model in Virtual Reality Interventional multiple factors in interventional surgery robot Training Systems Chenguang Yang 1, Shuxiang Guo 1,2*, Chuqiao Lyu 1, Youchun Ma1, Yue Wang 1 School of Life Science, Beijing Institute of Technology Peng Shi1, Shuxiang Guo1,2*, Xiaoliang Jin1 and Xinming Li1 Beijing, China 1Department of Intelligent Mechanical Systems Engineering, Faculty of Engineering and Design, • We designed and implemented three Kagawa University, Takamatsu, Japan 2Key Laboratory of Convergence Medical Engineering System and Healthcare Technology, The Ministry blood vessel phantoms in vitro based of Industry and Information Technology, School of life science, on real human body. Beijing Institute of Technology,Beijing, China • We chose three risk factors • The centerline of vasculature is according to master-slave robot considered as a generalized rotational operation. symmetry and the center point of each • We used the master-slave robot cylindrical shape is located at system previously studied in the centerline. laboratory verified the specific • A pre-processing strategy is proposed effects of the risk factors. The force measurement device to merge duplicate vertices to improve Centerline Extraction the operation efficiency. • We conducted experiment and made conclusions in the end.

TP2-2(5) 16:30-16:45 TP2-2(6) 16:45-17:00 An Image Information-based Classification Method for Study on the Automatic Surgical Method of the Vascular Interventional Surgery Operating Skills Vascular Interventional Surgical Robot Based on Yue Wang1, Jin Guo*, Shuxiang Guo1,2, Chuqiao Lyu1, Youchun Ma1, Chenguang Yang1, Zeyu Li1 Deep Learning Key Laboratory of Convergence Biomedical Engineering System and Healthcare Technology, The Ministry of Industry and Information Technology, School of Life Science, Jian Guo1, Suxiang Feng1 and Shuxiang Guo1,2 Beijing Institute of Technology, Beijing, China 1 Tianjin University of Technology ，Tianjin, China Takamatsu, Japan 2 Graduate School of Engineering, Kagawa University, Takamatsu, Japan • Images of different skill groups • Applying the AlexNet deep operating guidewires were collected. learning net to complete automatic surgery judgment mechanism . • We got the distal end coordinates of guidewire in each image. • Error analysis is carried out through experiments, and the • We generated a trajectory from the experimental results show that the distal end coordinates of an operation. automatic surgical judgment • VGG network was used to classify the mechanism based on deep generated tracks and distinguish the learning is capable of completing The structure of Alexnet operations of different skill groups. The Accuracy of Network Classification automatic judgment.

IEEE ICMA 2021 Conference Digest TA1-3: Mobile Robot System (I)

Session Chairs: Wangli He, East China University of Science and Technology Chunying Li, Kagawa University Online Conference Room 3, UTC+9(Tokyo Time): 9:30-11:00, Tuesday, 10 August 2021

TA1-3(1) 9:30-9:45 TA1-3(2) 9:45-10:00 Swipe Gesture Robotic System Based on Hexapod Robot with Ground Reaction Force Image Processing and Servo Compensation Sensor on Rough Terrain Koichi Tokairin, Daisuke Koyama, Meifen Cao, Koumei Yamashita, Takaki Kiyozumi and Yusuke Tada Kotaro Hashikura, Md Abdus Samad Kamal and Kou Yamada Tokyo Metropolitan College of Industrial Technology, Tokyo, Japan Mechanical Science and Technology, Gunma University, Kiryu, Japan • A hexapod robot and a 3-axis ground reaction force sensor have • For resolving COVID-19 been designed and developed in this research. problem or last one mile problem. • The developed ground reaction force sensor detects the contact • We are developing a robotic between foot and ground as a thrust load, and the contact between system such that the mobile robot foot and wall as a radial load. is controlled with swipe gesture. • A posture control method of • The GUI system generates an walking on rough terrain is appropriate robot reference by proposed for the hexapod robot processing the detected human Proposed gesture robotic system using the force sensor developed motion. in this research. • The proposed robotic system enabled the mobile robot to trace • Experimental verification results a human more efficiently. are reported in this paper. Walking Experiments on Random Step Field

TA1-3(3) 10:00-10:15 TA1-3(4) 10:15-10:30

Research and Analysis of Comprehensive Mobile Robot Path Planning Based on Improved Optimization Method for Energy Consumption Ant Colony Fusion Dynamic Window Approach Lei Shao, Qi Li, Chao Li, Wentao Sun and Trajectory Error of the Leg Structure Based Tianjin Key Laboratory for Control Theory & Application in Complicated Systems, College of on Virtual Model Control Electrical and Electronic Engineering, Tianjin University of Technology, China Geqi Lin1, Wenchuan Jia1, Shugen Ma1,2, Jianjun Yuan1, Yi Sun1 •An adaptive distance induction factor 1. Shanghai University, Shanghai, China is designed and combined with MMAS 2. Ritsumeikan University, Shiga, Japan. to improve the pheromone update rule. • A VMC based comprehensive optimization •Improve the probability transition rule method for quadruped gait is proposed. by constructing corner suppression • Several fitness functions are designed factor. according to different sets of motion •DWA is fused to track the global path energy consumption and tracking error. points generated by the Ant Colony. • Co-simulation and basic physical experiments are performed. Path planning of fusion algorithm comprehensive optimization results

TA1-3(5) 10:30-10:45 TA1-3(6) 10:45-11:00 Interaction-Aware Crowd Navigation via Control of A Lower Limb Exoskeleton Robot by Augmented Relational Graph Learning Qing Xu1, Wangli He1, and Naoyuki Kubota2 Upper Limb sEMG Signal 1 Key Laboratory of Smart Manufacturing in Energy Chemical Process, Ministry of Education, Shuxiang Guo，Yibin Ding and Jian Guo East China University of Science and Technology Tianjin Key Laboratory for Control Theory, Tianjin University of Technology Shanghai, China Tianjin,China 2 The Graduate School of System Design Tokyo Metropolitan University • Analyze generating principle of Tokyo, Japan EMG signal. • We propose CEM-RGL, an augmented relational graph • The EMG signals of upper limbs based reinforcement learning, for crowd navigation; were collected and filtered.. • We incorporate cross entropy method • Extracted four time domain into a relational graph learning characteristics of EMG signal. framework to get sufficient samples • Designed BP neural network in continuous state-action space; classifier and classified two kinds of • We introduces graph attention network and reward shaping to extract arm movements. efficient representation and accelerate the training convergence. The Structure of Robot

IEEE ICMA 2021 Conference Digest TA2-3: Mobile Robot System (II)

Session Chairs: Wangli He, East China University of Science and Technology Chunying Li, Kagawa University Online Conference Room 3, UTC+9(Tokyo Time): 11:15-12:15, Tuesday, 10 August 2021

TA2-3(1) 11:15-11:30 TA2-3(2) 11:30-11:45 LIDAR Graph SLAM based Autonomous Vehicle Maps Local Path Planning using Distance-Type Fuzzy using XY and Yaw Dead-Reckoning Measurements Reasoning Method in Unstructured and Dynamic Mohammad Aldibaja, Reo Yanase, Takahiro Furuya, Akitaka Oko and Naoki Suganuma Construction Sites Advanced Mobility Research Institute, Kanazawa University Guang Yang1, Shuoyu Wang1, Hajime Okamura1, Yasuhiro Ueda2, Toshiaki Yasui2, Tetsuya Yamada2, Kanazawa, Japan Yuki Miyazawa2, Satomi Yoshida2, Yuta Inada2, Shingo Ino3, Kazuo Okuhata3, Yoshinobu Mizobuchi3 • A Unique Graph Slam (GS) 1 Department of Intelligent Mechanical Systems Engineering, Kochi University of Technology, Kami, Japan framework is proposed to generate 2Maeda Corporation, Tokyo, Japan precise maps based on Dead- 3Satt Systems, Kochi, Japan Reckoning (DR) Measurements. • A local path planning method that utilizes the • The relative yaw angle errors are distance-type fuzzy reasoning method is compensated in the image domain proposed. by Fourier Mellin Transform. • Both non-sparse and sparse input fuzzy sets • XY relative positions are accuratly representing sensor readings can be properly processed. recovered by Phase Correlation. • And the reasoned results can be trusted to • The cost-function is designed to follow the rules in the rule bases. globally optimized the compensated DR Map with XY-Yaw Deviations • The presented approach has been evaluated in positions by (GS-Yaw ➔ GS-XY). and GS Accurate Map real construction sites. Material Transportation Robot

TA2-3(3) 11:45-12:00 TA2-3(4) 12:00-12:15 Study on Positioning for The Spherical Multi-robot Multi-path Intelligent Planning Based Amphibious Robot Based on Visual-Inertia on Universal Conflict Resolution and Free Localization Crossing Emergence Jian Guo1 ,Xiangyu Chen1, Shuxiang Guo1,2, Jigang Xu3 Zeyu Zhou, Mingyang Li, Jingxi Zhang, Xiongwei Wu and Wei Tang 1Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems and Control Engineering, Northwestern Polytechnical University Biomedical Robot Laboratory, Tianjin University of Technology, Tianjin, China Takamatsu, Japan 2Intelligent Mechanical Systems Engineering Department Faculty of Engineering, Kagawa University, Kagawa, Japan • An algorithm of UCR-FCE is 3Qinghai Haidong,810700,China proposed to solve multi-robot path conflict. • Using the visual positioning information estimate the zero • UCR-FCE can solve all the conflict deviation of the IMU scenarios with sufficient path nodes and no path deadlock. • Two localization algorithms are fused by untraceable Kalman filter • A lightweight multi-robot to form a more robust localization information interaction mechanism algorithm The trajectory of the robot running the is built to achieve real-time path UCR-FCE algorithm communication between robots. improved algorithm

IEEE ICMA 2021 Conference Digest TP1-3: Mobile Robot System (III)

Session Chairs: Qiang Fu, Tianjin University of Technology Xinming Li, Kagawa University Online Conference Room 3, UTC+9(Tokyo Time): 13:30-15:00, Tuesday, 10 August 2021

TP1-3(1) 13:30-13:45 TP1-3(2) 13:45-14:00 A kind of Sea-fog Removal Algorithm Fusing Design of Automatic Picking Robot Based on 2- Image Enhancement and End-to-end Network DOF Stabilized Platform Zhongli Ma, Lili Wu*, Linshuai Zhang,Jiadi Li, Yuehan Zeng Chao Zhang , Bing Zhang , Hao Zhang School of Control Engineering,Chengdu University of Information Technology Innovation Center Of Robotics And Intelligent Equipment, Dongguan University Of Technology Chengdu, China, Dongguan, China • A kind of single sea fog image defogging Banana picking robot includes method is proposed.which fuses the image three parts: enhancement with end-to-end network. • The robot moving chassis which • Firstly, the image enhancement based on can adapt to the complex ground Multi Scale Retinex;Then, the end-to-end environment in the field; network was used to extract the transmi- The principle of proposed • A mechanical arm with large span ttance, and then the transmittance was fusion defogging algorithm operation ability in the vertical optimized by the guided filter. direction and the horizontal • The experimental results show that the direction; proposed single-image sea fog removal • A grip-shear integrated operation algorithm has obvious defogging effects handle at the end of the manipulator Banana picking robot and Original picture Our result and good image restoration ability. arm. the picking operation

TP1-3(3) 14:00-14:15 TP1-3(4) 14:15-14:30 Research on underwater target detection method Unit Dual Quaternion Based Set-Point Control based on improved MSRCP and YOLOv3 with Input Constraints Tongxu Guo, Yanhui Wei, Hong Shao, and Boye Ma Institute of Intelligent Science and Engineering, Harbin Engineering University Wenxin Hu, Yajing Wang and Xiangke Wang Harbin, China College of Intelligence Science and Technology, National University of Defense Technology Changsha, China • Image processing and target detection algorithms for • This paper presents a first attempt to underwater creatures. solve the set-point control problem for the unit dual quaternion (UDQ) • Improved underwater MSRCR based kinematic system subject to image enhancement technology input constraints. based on adaptive algorithm. • It strictly proves that the UDQ- • Comprehensive types and huge logarithm based proportional control number of data sets. law can still make this system stable • Fast recognition algorithm based by using Lyapunov method. on YOLOv3. Underwater target detection Set-Point Control Scheme

TP1-3(5) 14:30-14:45 TP1-3(6) 14:45-15:00 Finite-Time Attitude Control of Uncertain A Variable Artificial Potential Field Method for Quadrotor Aircraft via Continuous Terminal Gait Generation of Quadruped Robot Feng Zhuo, Wenchuan Jia, Shugen Ma, Jianjun Yuan, Yi Sun Sliding-Mode-Based Active Anti-Disturbance School of Mechatronic Engineering and Automation Shanghai Key Laboratory of Intelligent Manufacturing and Robotics Approach Shanghai University, Shanghai, China Omar Mechali, Jamshed Iqbal, Abdesselam Mechali, Xiaomei Xie, Limei Xu School of Aeronautics and Astronautics, University of Electronic Science and Technology of • Proposing a variable artificial China, Chengdu, China potential field (VAPF) to generate the gait of a quadruped • This paper addresses the problem of robot. robust attitude control for a quadrotor. • Constructing the point potential field-based model in different • A new continuous terminal sliding stages. mode-based active anti-disturbance control (CTSMBAADC) is proposed. • Designing an attitude controller to generate the actual gait with • Experiments are conducted to the VAPF. the target and application point corroborate the theoretical findings. Fig. 1 Experiment setup for the attitude control of the quadrotor.

IEEE ICMA 2021 Conference Digest TP2-3: Mobile Robot System (IV)

Session Chairs: Qiang Fu, Tianjin University of Technology Xinming Li, Kagawa University Online Conference Room 3, UTC+9(Tokyo Time): 15:30-17:00, Tuesday, 10 August 2021

TP2-3(1) 15:30-15:45 TP2-3(2) 15:45-16:00 Development of Prototype Electric-driven Active Cross-domain Self-localization Using 2-DoF Monopod Robot for Hopping Motion Pole-like Landmarks Asahi Anzai, Toshihide Xuechao Chen, Kenji Hashimoto, Tanaka Kanji Doi, Kazuki Hashida, Lianqiang Han, Department of Mechanical Engineering, University of Fukui Graduate School of School of Mechatronical Engineering Informatics / Humanoid Fukui, Japan Science and Engineering, Engineering, Beijing Robotics Institute (HRI), • The novel task of active cross- Meiji University, Institute of Technology, Meiji University / Waseda University, Kanagawa, Japan Beijing, China Kanagawa, Japan domain self-localization is addressed • We developed a one-legged hopping robot • A novel holistic pole landmark MH-1. detector is developed and employed • MH-1 is an electrically driven 2-DoF • Pole detector and self-localization monopod with pitch axes at the hip and knee system are implemented joints. • Experiments using public NCLT • MH-1 has a height of 510mm and a mass of dataset are conducted 6.9kg. System architecture • MH-1 realized hopping within the constraint that the robot can only move vertically. MH-1

TP2-3(3) 16:00-16:15 TP2-3(4) 16:15-16:30 Underwater Motion Characteristics Evaluation of Natural Residual Reinforcement Learning for a Bio-inspired Father-son Robot Bicycle Robot Control on ICMA 2021 Ruochen An, Shuxiang Guo, Chunying Li, and Tendeng Awa Xianjin Zhu1, Xudong Zheng3, Qiyuan Zhang1, Zhang chen2, Yu Liu1, Bin Liang2 Graduate School of Engineering, Kagawa University 1Harbin Institute of Technology, 2Tsinghua University, 3Tsinghua Shenzhen International Takamatsu, Japan • A father-son robot system for the • This work focuses on balance underwater sample acquisition was control and path tracking of bicycle proposed. robot by using the proposed NRRL algorithm. • Control circuit of the father-son robot system. • NRRL decomposes the main tasks into subtasks and introduces • The underwater experiments of the qualitative prior knowledge based father robot of the father-son robot on the RRL algorithm. system with the disturbance of the wind. • Simulation results show better performance and better sample • The fluid simulation of the drive Schematic of the father-son efficiency of the proposed NRRL The bicycle robot device of the father-son robot robot system algorithm. system.

TP2-3(5) 16:30-16:45 TP2-3(6) 16:45-17:00 Study on a Novel Network Nod Monitoring Amphibian robot small data set aquatic animal System based on the Spherical Multi-robot classification based on transfer learning Shuxiang Guo1,2 ,Ran Wang1 ,Jian Guo1* ,Jigang Xu3* Shuxiang Guo1,2 ,Shaolong Wang1 ,Jian Guo1* ,Jigang Xu3* 1Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems and 1Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems and Intelligent Robot Laboratory, Tianjin University of Technology,Tianjin, China Intelligent Robot Laboratory, Tianjin University of Technology,Tianjin, China 2Intelligent Mechanical Systems Engineering Department Faculty of Engineering, Kagawa 2Intelligent Mechanical Systems Engineering Department Faculty of Engineering, Kagawa University, Kagawa, Japan University, Kagawa, Japan 3Qinghai Haidong,810700,China 3Qinghai Haidong,810700,China • The spherical multi-robot system is • The paper proposes a small data set mainly used for water quality aquatic animal classification model monitoring in large-scale aquaculture. based on convolutional neural • The experimental results show that it network and transfer learning in a is feasible to construct the whole spherical robot. node monitoring system based on • The model optimizes the accuracy of Mesh. With this system, the multi- aquatic animal target recognition, and robot system can accomplish more the test accuracy reaches 93.11%, Network nod monitoring complex tasks. with good stability and high precision. Examples of aquatic animal system. images 36

IEEE ICMA 2021 Conference Digest

TA1-4: Biomimetic Systems

Session Chairs: Yan Zhao, Beijing Institute of Technology Ao Li, Beijing Institute of Technology Online Conference Room 4, UTC+9(Tokyo Time): 9:30-11:00, Tuesday, 10 August 2021

TA1-4(1) 9:30-9:45 TA1-4(2) 9:45-10:00 An Elastic Biomimetic Fish Tail and Its A Combined Indoor Self-positioning Method for Undulation Fitting Method of Body Wave Robotic Fish Based on Multi-sensor Fusion Xiaocun Liao, Yuzhuo Fu, Ben Lu, Qianqian Zou, Zhuoliang Zhang and Chao Zhou Yuzhuo Fu, Ben Lu, Xiaocun Liao, Qianqian Zou, Zhuoliang Zhang and Chao Zhou Institute of Automation, Chinese Academy of Sciences Institute of Automation, Chinese Academy of Sciences Beijing, China Beijing, China • A novel wire-driven fish tail • A combined indoor self-positioning with two joints, using spring- method for robotic fishin this paper steel-sheet to mimic fish spine. is introduced. • The kinematics model for • An ostracion-like robotic fish is used predicting the attitude of fish as the experimental object to achieve tail. centimeter-level positioning with an • A based-vision biomimetic average positioning error of 4.492 undulation fitting method for A Novel Wire-driven Fish Tail cm in a short-distance range and solving the control coupling decimeter-level positioning with an problem. error of 2.049 dm in a medium- Positioning results of robotic fish distance range.

TA1-4(3) 10:00-10:15 TA1-4(4) 10:15-10:30 Design of a Novel Quadruped Robot Based on Influence of the Pappus Shape of Bionic Tensegrity Structures dandelion aircraft on Drag Characteristics Ruiqi Gao, Yixiang Liu, Qing Bi, Bin Yang, and Yibin Li Min Chen, Wen Zhu, Yongchen Liu, Xiaomei Xie, Lei Guo and Ao Jiao 1. School of Control Science and Engineering, Shandong University University of Electronic Science and Technology of China 2. Shandong Institute of Advanced Technology, Chinese Academy of Sciences Chengdu, Sichuan Province, China 3. Engineering Research Center of Intelligent Unmanned Systems, Ministry of Education • According to the morphometric results, 4. Tianjin Institute of Aerospace Mechanical and Electrical Equipment the geometric and simulation models • This paper introduces tensegrity structure of bionic dandelion aircraft was into the spine and limb structure of constructed. quadruped robot . • The cylindrical and cuboid pappus • The tensegrity structures based quadruped shape of bionic dandelion aircraft are robot has high flexibility, adaptability, and concerned mainly in simulation. safety. • This paper proves that the porosity of bionic dandelion aircraft is not the • This study lays a foundation for the Prototype of the absolute factor to determine the drag research of quadruped robot based on tensegrity structures coefficient, and the pappus shape is Bionic dandelion aircraft with tensegrity structure. based quadruped robot also an important factor. different shapes of pappi

TA1-4(5) 10:30-10:45 TA1-4(6) 10:45-11:00 Structural synthesis of nested hyper-redundant Structure design and mathematical modeling of mechanisms inspired by the mackerel fishes bionic butterfly flapping wing aircraft Weiwei Chen and Hongzhou Jiang School of Mechatronics Engineering, Harbin Institute of Technology Yue Zhu, Hengjing Huang,Yanlong Wang,Zhenfeng Han and Jun Zhong Harbin, China Mechanical&Electrical Engineering, Hohai University, Changzhou , Jiangsu Province, China • Using the code-based approach, this article comes up with a • This paper studies a bionic butterfly simplified characteristic code for different kind of nested hyper- flapping wing robot. The design of the redundant mechanisms inspired by the macherel fishes. robot is based on a mechanism composed of a crank slider and a guide RM rod. Without the need for program POT AOT control, the speed and torque can be changed.The kinematics and dynamics N N+1 N+2 N+3 N+4 N+5 N+6 BB modeling of the flapping-wing mechanism was carried out, and the feasibility of the mechanism was verified by MATLAB simulation. Drawing of the nested The bionic butterfly flapping hyper-redundant mechanisms wing aircraft

IEEE ICMA 2021 Conference Digest

TA2-4: Mobile Robot System (V)

Session Chairs: Mitsuharu Matsumoto, The University of Electro-Communications Ao Li, Beijing Institute of Technology Online Conference Room 4, UTC+9(Tokyo Time): 11:15-12:15, Tuesday, 10 August 2021

TA2-4(1) 11:15-11:30 TA2-4(2) 11:30-11:45

Garbot - Semantic Segmentation for Material Investigation of driving principle of Recycling and 3D Reconstruction Utilizing non-electrically driven robots using sound Robotics waves as a power source Shinichi Otake and Mitsuharu Matsumoto Siva Ariram, Tuulia Pennanen, Antti Tikanmäki, and Juha Röning Department of Informatics, The University of Electro-Communications, Tokyo, Japan Biomimetics and Intelligent Systems Group, University of Oulu, Finland • This paper describe non-electrically • Semantic segmentation directly from driven robot, assuming the use of the images of landfills can be robots in situations where electricity utilized in the earth movers to cannot be used. segregate the garbage autonomously. • We investigated the feasibility of a • The project also contributes to robot that operates using sound reconstruct the segmented images to waves as a drive source. build a 3D map and this exploits the use of earth moving vehicles to • Through several experiments, we navigate autonomously by localizing confirmed that the developed robot the segmented objects. could be controlled arbitrarily even Typical movement Garbot - Semantic Segmentation if the installation location changes. of two small cups on styrofoam

TA2-4(3) 11:45-12:00 TA2-4(4) 12:00-12:15 Design and Dynamics Simulation of a A Path Planning Method for the Spherical Triphibious Robot in Webots Environment Amphibious Robot Based on Improved A* Xiaoyi Gu, Anwei Zhang, Li Yuan, Yuanhao Xia Northeastern University of China，Liaoning，China Algorithm Jian Guo1,2 ,Xiaojie Huo1 ,Shuxiang Guo1* ,Jigang Xu3* • Introduce the purpose of the design 1Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems and of a triphibious robot. Intelligent Robot Laboratory, Tianjin University of Technology,Tianjin, China 2 • Describe the advantages of the Intelligent Mechanical Systems Engineering Department Faculty of Engineering, Kagawa University, Kagawa, Japan mechanical structure and three kinds 3Qinghai Haidong,810700,China of working methods according to different environments. • This paper presents an improved A* algorithm for path planning of the • Verify the feasibility of the model in The triphibious robot Webots Environment. spherical amphibious robots. • The multiple target points path planning problem of spherical amphibious robot is solved. Multiple Target Points Path Planning

IEEE ICMA 2021 Conference Digest

TP1-4: Human-System Interaction and Interface

Session Chairs: Yi Liu, Beijing Institute of Technology Ziyi Yang, Kagawa University Online Conference Room 4, UTC+9(Tokyo Time): 13:30-15:00, Tuesday, 10 August 2021

TP1-4(1) 13:30-13:45 TP1-4(2) 13:45-14:00

Preliminary Evaluation of a Performance-based A Behavior Study on the Effect of Baby Schema Stiffness Control for Upper Limb Elbow Joints on Face Discrimination for Robot Appearance Rehabilitation Design Ziyi Yang, Shuxiang Guo and Yi Liu Lichang Yao, Qi Dai, Yiyang Yu, Yuki Nishioka, Qiong Wu, Mengni Zhou, Jiajia Yang, Satoshi Department of Intelligent Mechanical Systems Engineering, Graduate School of Engineering, Kagawa University Takamatsu，Japan Takahashi, Yoshimichi Ejima, Jinglong Wu The Graduate School of Interdisciplinary Science and Engineering in Health Systems Okayama University • The purpose of this study was to discuss whether human will have behavioral special reactions when observing the infants’ faces. • Human would react quickly to infants, even infants not with high baby schema parameters.

Fig. 1. Dynamic modeling of the PVSED • This conclusion takes a new idea of how human react to • The elbow joint output stiffness would be regulated according the patient’s training performance which infants, and it was hope to support will be determined using the real-time position in improving robot appearance tracking error. design Fig. 2 Interaction force evaluation results in Low-stiffness condition Reaction time to the adults and the infants' face under three conditions.

TP1-4(3) 14:00-14:15 TP1-4(4) 14:15-14:30 A Basic Psychophysics Study of Visual Masking A Basic Psychophysics Study of Sound Effect on Kanji Recognition for Image Reliability Effects on Audiovisual Integration for Recognition Technology Developing New Virtual Reality Device Qi Dai, Lichang Yao, Ikue Hattori, Qiong Wu, Jiajia Yang, Satoshi Takahashi, Yoshimichi Ejima and Jinglong Wu * Hongtao Yu, Mengni Zhou, Jiajia Yang, Satoshi Takahashi, Yoshimichi Ejima, Jinglong Wu* Cognitive Neuroscience Laboratory., Okayama University Cognitive Neuroscience Lab, Okayama University, Japan Qiong Wu Okayama, Japan Department of Psychology, Suzhou University of Science and Technology, China • Present study measured the sensitivity Qi Li index of Kanji detection and Brain informatics Laboratory, Changchun University of Science and Technology, China discrimination in the masking paradigm • Living picture category have a using the Signal Detection Theory. significant advantage than non- • The results show that the longer the living picture category under presentation time in visual processing, semantic unreliable the higher the sensitivity index. multisensory condition, • Study the human visual processing, indicated a robust semantic hope to support the improvement of representation for living objects. image recognition technology.

TP1-4(5) 14:30-14:45 TP1-4(6) 14:45-15:00

Design of the Lower Limb Rehabilitation Design of front feed PID Control System for the Training System Based on Virtual Reality Limb Rehabilitation Robot Based on BP Neural Jian Guo1* and Kai Zhao1 ，ShuxiangGuo1,2 1Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems and Network Intelligent Robot Laboratory, Tianjin University of Technology,Tianjin, China Jian Guo1 ,Fudong Huo1 ,Shuxiang Guo1,2* 2Intelligent Mechanical Systems Engineering Department Faculty of Engineering, Kagawa 1Tianjin Key Laboratory for Control Theory & Applications In Complicated systems and University, Kagawa, Japan Intelligent Robot Laboratory Tianjin University of Technology,Tianjin,China 2Intelligent Mechanical System Engineering Department Faculty of Engineering,Kagawa A virtual reality rehabilitation University,Kagawa, Japan training system based on the lower limb rehabilitation robot • A front feed PID control system of our team was proposed, based on neural network is proposed. which organically combined • Neural network front feed PID can visual feedback, vibration effectively improve the response feedback and sound feedback speed and tracking effect of the with the lower limb system. rehabilitation robot. Schematic diagram of rehabilitation robot Right hip tracking simulation

IEEE ICMA 2021 Conference Digest

TP2-4: Medical, Biomedical and Rehabilitation Systems

Session Chairs: Yi Liu, Beijing Institute of Technology Ziyi Yang, Kagawa University Online Conference Room 4, UTC+9(Tokyo Time): 15:30-17:00, Tuesday, 10 August 2021

TP2-4(1) 15:30-15:45 TP2-4(2) 15:45-16:00 Design A Novel of Path Planning Method for The Proposal of Classification Table for Transfer Vascular Interventional Surgery Robot based on Assistance Based on Condition of Care-Receiver DWA Model Chiharu Ishii and Ryo Sugiyama Department of Mechanical Engineering, Hosei University Jian Guo1, Han Zhao1 and Shuxiang Guo1,2 Tokyo, Japan 1 Tianjin University of Technology ，Tianjin, China 2 Graduate School of Engineering, Kagawa University, Takamatsu, Japan • A hands-free activation device to • The surgical navigation system supply compressed air to the proposed in this article is designed artificial muscle attached to the based on the DWA(Dynamic assist suit “Sustainable” was window algorithm) model. developed. • The experimental analysis and • For transfer work in nursing care, verification of surgical operations a classification table, in which compared with or without Working flow chart of path caregivers can choose a suitable navigation system was planning system assistance method corresponding demonstrated. to the condition of the care- Transfer Work receiver, was proposed.

TP2-4(3) 16:00-16:15 TP2-4(4) 16:15-16:30 Development of a Novel Intraoperative A Training System for Vascular Interventional Information Monitor System for the Vascular Surgeons based on Local Path Planning Interventional Surgery Robotic System Jian Guo1, Yue Sun1 and Shuxiang Guo1,2 1 1，2*, 1 1 1 Tianjin University of Technology ，Tianjin, China Wei Zhou , Shuxiang Guo Zhengyang Chen , Fanxu Meng 2 Graduate School of Engineering, Kagawa University, Takamatsu, Japan 1The Ministry of Industry and Information Technology, Beijing Institute of Technology 2 Faculty of Engineering, Kagawa University Japan • This paper presents a system of • Display the collision force catheter local path planning based information of the slave on improved RRT algorithm. manipulator. • This paper compares the local • Display the displacement path and planning time of the information with the errors of traditional RRT algorithm with the master manipulator and that of the improved RRT slave manipulator. algorithm. • Display the video image of the slave side. The developed information Flow chart of dynamic local path planning • Warning and monitor system monitor system.

TP2-4(5) 16:30-16:45 TP2-4(6) 16:45-17:00

Prediction of Physiological Tremor Base on Performance Evaluation of the Vascular Model Deeplearning for Vascular Interventional Based on the Nonlinear Viscoelastic Tensor- Surgery Robot Mass Method Zhengyang Chen, Shuxiang Guo, Wei Zhou, Fanxu Meng Liuqing Zhang, Shuxiang Guo, Cheng Yang Key Laboratory of Convergence Biomedical Engineering System and Healthcare Technology, Beijing Institute of Technology The Ministry of Industry and Information Technology, School of Life Science, Beijing Institute of Technology, Beijing, China • Proposes a multi-step signal prediction method based on • A vascular tissue model based LSTM with moving windows. on nonlinear viscoelastic tensor mass method is proposed. • Predict the tremor signal in real- time which reduce the influence • The algorithm can realize real- of the phase delay. time visual feedback and tactile feedback. • Improves the accuracy of tremor prediction compared with the • The method can be used to other prediction method. The MW Multi-step LSTM simulate the nonlinear behavior of blood vessels. Virtual Reality Simulation Platform

IEEE ICMA 2021 Conference Digest

TA1-5: Industrial, Manufacturing Process and Automation

Session Chairs: Aiguo Ming, The University of Electro-Communications Liang Zheng, Changchun University of Science and Technology Online Conference Room 5, UTC+9(Tokyo Time): 9:30-11:00, Tuesday, 10 August 2021

TA1-5(1) 9:30-9:45 TA1-5(2) 9:45-10:00 Error Analysis of Vertical Beam Inclination of A CMOS Soil Yeast Count to Frequency Delta-Robot Three-Dimensional Printers Converter for Sensing Yeast Count Applications Yuezong Wang, Youfan Peng, Jiqiang Chen, Yangyang Lu Faculty of Materials and Manufacturing, Beijing University of Technology Cheng-Ta Chiang, Chung-Yu Hsu, and Yao-Chieh Huang Beijing, China Department of Electrical Engineering, National Chia Yi University, Chia Yi, Taiwan • Simulating of parallel motion mechanism • The proposed chip can be easily • Deriving non-ideal utilized to sense soil yeast in real parallel printing model time. • Analyzing the influence • The adjustment sensitivity circuits of error source on 3D could adjust the sensitivities under printing precision different kinds of soil. Setup and model Block diagram of the proposed chip

TA1-5(3) 10:00-10:15 TA1-5(4) 10:15-10:30 Research on Nonlinear Ultrasonic Testing Economic Dispatch of Microgrid Based on Technology of Microcracks in Metallic Materials Adaptive Mutation Particle Swarm Optimization Qinxue Pan, Shuangyang Li, Lang Xu, Yunmiao Zhang, Meile Chang, Xiaoyu Xu, Sa Li, Wei Li School of Mechanical Engineering, Beijing Institute of Technology Ji Li, Wenlong Nie, Xiaoning Xu, Lei Shao, Wentao Sun Beijing, China College of Electrical and Electronic Engineering, Tianjin University of Technology • Nonlinear ultrasonic can be used Tianjin, China to detect microcracks in material. • Overcome the disadvantages of Simulation and experiment prove Particle Swarm Optimization for its feasibility. low solving accuracy. • A nonlinear ultrasonic testing • The inertia weight is decreased by system is built based on RITEC an adaptive normal distribution and RAM 5000 and other modules. the movement strategy of the particle position is updated with the • There is an approximate positive iterations. correlation linear relationship Comparison convergence Nonlinear Ultrasonic • Solves the simulation of the between the relative nonlinear of the process in Wind- Measurement System operation cost model of wind-solar coefficient and the ultrasonic Sunny propagation distance. storage under four typical weather conditions.

TA1-5(5) 10:30-10:45 TA1-5(6) 10:45-11:00 Design of Remote Control Inverter Based on Spacecraft Equipment Health Condition MQTT Communication Protocol Monitoring Based on Augmented Dickey-Fuller Kunpeng Yang, Baofeng Zhang, Jinwei Zhang, Junchao Zhu Tianjin Key Laboratory for Control Theory & Application in Complicated Systems Tianjin test and Gaussian Mixture Model University of Technology, Tianjin Jinsheng Environmental Technology Fusheng Zhang1, Shumei Zhang2*, Wentao Wu2, Chao Tan2, Yang Zhao1 Co., Ltd. ,Tianjin ,China 1Institute of Spacecraft System Engineering, China Academy of Space Technology, Beijing, China 2School of Electrical and Information Engineering, Tianjin University, Tianjin, China • This article proposes the MQTT • A health condition monitoring method communication protocol on the ADF-GMM (Augmented Dickey- (a) ESP8266 hardware module. Fuller-Gaussian mixture model) is • The PC terminal starts the MQTT proposed for spacecraft equipment . cloud server and subscribes messages • ADF test is conducted to select the key to the MQTT cloud server as an features that contain performance (b) degradation information. MQTT communication client. Hardware physical map • The upper computer adopts a QT • A mixed Gaussian model is established design to achieve the purpose of real- to monitor the health condition and Fig. 4. Bearing health monitoring: time control of the inverters of evaluate the degradation of the system (a) BID in logarithmic scale; using the BID index. different slave stations. (b) Reliability based on BID.

41 IEEE ICMA 2021 Conference Digest

TA2-5: Elements, Structures, and Mechanisms

Session Chairs: Xianqiang Bao, Beijing Institute of Technology Lingling Zheng, Kagawa University Online Conference Room 5, UTC+9(Tokyo Time): 11:15-12:15, Tuesday, 10 August 2021

TA2-5(1) 11:15-11:30 TA2-5(2) 11:30-11:45

Heel Brake for Personal Mobility Vehicle Evaluation of Resonant Tank Topologies for High- Shizuki Kawauchi, Isaku Nagai and Keigo Watanabe Graduate School of Natural Science and Technology, Okayama University Voltage Power Supply in Ion Propulsion Application Okayama, Japan Minghai Dong, Hui Li, Shan Yin, Zhenni Zeng, Jinshu Lin, Xi Liu, Yingzhe Wu • By using the restoring force School of Aeronautics & Astronautics, University of Electronic Science and Technology of China Chengdu, China of the spring, unintentional braking does not occur. • the resonant tank topology is evaluated. • Braking with the heel • A selecting criteria for the resonant tank is prevents the pilot from falling proposed. due to the backward leaning Shaft • The gain valley could be more easily posture and shifting of the founded in the 4-element resonant tank center of gravity. Hall sensor compared with others, and it’s more suitable for the high-voltage application. Spring Block diagram of HV resonant converter • The amount of deceleration differs depending on the input Overview of heel brake angle.

TA2-5(3) 11:45-12:00 TA2-5(4) 12:00-12:15 A Study of Ship Supportability Evaluation Using A Bio-Inspired Quadruped Robot with Foldable AHP and DS Evidence Theory Torso Capable of Omnidirectional Motion Ling Xiong, Peng Shang, Jiale Tian, Jian Zhou Zhe Wang, Yixiang Liu, Qing Bi, Bin Yang, and Rui Song 92942 People’s Liberation Army; Xi’an Jiaotong University; 1. School of Control Science and Engineering, Shandong University, China China 2. Engineering Research Center of Intelligent Unmanned Systems, Ministry of Education 3.Shandong Institute of Advanced Technology, Chinese Academy of Sciences, China • A new model is proposed to evaluate 4. Tianjin Institute of Aerospace Mechanical and Electrical Equipment, China ship supportability using AHP and DS • A novel variable structure evidence theory. omnidirectional mobile bio-inspired • AHP accounts both qualitative and quadruped robot is proposed. quantitative information to create the evaluation index system. • A fully symmetrical mechanism with • DS evidence theory is used to reduce variable configuration is designed to error and uncertainty in the expert realize a foldable torso. scores. • The quadruped robot with a foldable • Example is given to show the torso and omnidirectional mobility is effectiveness and accuracy of this more adaptable to a complex The state of quadruped robot method. Flow Chart of Evaluation environment.

42 IEEE ICMA 2021 Conference Digest

TP1-5: Intelligent Biomedical Instrument Technology

Session Chairs: Xianqiang Bao, Beijing Institute of Technology Lingling Zheng, Kagawa University Online Conference Room 5, UTC+9(Tokyo Time): 13:30-15:00, Tuesday, 10 August 2021

TP1-5(1) 13:30-13:45 TP1-5(2) 13:45-14:00

Design of Temperature Measurement Motor Fault Diagnosis System Based on Single Identification Instrument based on OpenMV and Chip Microcomputer and Artificial Intelligence Mengfu Hao, Xin Wang, Peiyu Li, Fan He, Liu Yang, Meng Xiao, Xueqi Bian and Zixuan Zhang MLX90614 Tianjin Key Laboratory of Control Theory & Applications in Complicated Systems,Tianjin Xin Wang, Yuanshuang Yang,Zixuan Zhang,Yuechang Luo and Zepeng Li University of Technology Automation, TianJin University Of Technology Tianjin 300384, China Tianjin, China (1) A set of motor data acquisition (1) Temperature processing based on device is designed to collect various MLX90614 and Kalman filter. data information of the motor; (2) Temperature compensation based (2) A deep learning algorithm is on polynomial fitting. proposed to analyze the data and (3) Face recognition based on local reduce the time investment of facial features and OpenMV maintenance personnel; (3) Using the computer to carry out Motor fault simulation on-line real-time diagnosis to get the platform fault information

TP1-5(3) 14:00-14:15 TP1-5(4) 14:15-14:30 A Novel Target Tracking System for the A Novel Surgeon Training System for the Amphibious Robot based on Improved Camshift Vascular Interventional Surgery based on Algorithm Augmented Reality Shuxiang Guo1,2 ,Handong Cheng1 ,Jian Guo1* ,Jigang Xu3* Jian Guo1,2 ,Shichen Jia1 ,Shuxiang Guo1* 1Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems and 1Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems and Biomedical Robot Laboratory, Tianjin University of Technology, Tianjin, China Intelligent Robot Laboratory, Tianjin University of Technology,Tianjin, China 2Intelligent Mechanical Systems Engineering Department Faculty of Engineering, 2Intelligent Mechanical Systems Engineering Department Faculty of Engineering, Kagawa Kagawa University, Kagawa, Japan University, Kagawa, Japan 3Qinghai Haidong,810700,China • This paper proposed a method to • We presents a target tracking system apply augmented reality technology based on Camshift algorithm. to the training system, which could • We improve the accuracy and robustness reduce the workload of modeling of traditional Camshift algorithm in and give doctors a more realistic and target tracking. familiar visual experience. • We improved Camshift algorithm Training system based combined with Kalman filter was fused. The target tracking experiment on augmented reality

43 IEEE ICMA 2021 Conference Digest

TA1-6: Robot Navigation and Control Algorithm

Session Chairs: Daisuke Kurabayashi, Tokyo Institute of Technology Ruochen An, Kagawa University Online Conference Room 6, UTC+9(Tokyo Time): 9:30-11:00, Tuesday, 10 August 2021

TA1-6(1) 9:30-9:45 TA1-6 (2) 9:45-10:00

Searching Algorithm for Autonomous Performance Evaluation of the Assembly Mechanism for Distributed Agents based on Chemotactic Multimodule Capsule Robots Docking Behavior of C. elegans Lingling Zheng and Shuxiang Guo Intelligent of Mechanical System Engineering, Kagawa University *1 *2 *1 Takuto Yanagisawa , Yuya Hattori , and Daisuke Kurabayashi Takamatsu, Japan *1 Tokyo Institute of Technology, Tokyo, Japan *2 National Institute of Technology, Kure College, Hiroshima, Japan • Proposed a concept of multiply • A searching algorithm for a multimodule capsule robots and a peak environment was proposed. novel assembly method. • The behavior model was modified so • By applying a force on the that an agent visits a global peak claws, the deformation of the with high probability, without being claws under this force is (a) Rod mechanism. (b) Claw mechanism. captured by local peaks. calculated. Assembly mechanism. Fig: The red cross indicates • By Gaussian mixture model, a the global peak of the five- • In the simulation, the claw mechanism with six claws was analyzed, potential map to locate a global peak well potential. and according to the simulation results, the maximum displacement was obtained. of the claw tip was 1 mm under the load with a force of 39.42 mN.

TA1-6 (3) 10:00-10:15 TA1-6 (4) 10:15-10:30 Distributed Leader-Follower Formation Control Consensus Based Distributive Task Allocation of Quadrotors Swarm Subjected to Disturbances for Multi-AUV in Searching and Detecting Di Wu, Liukun Zhang, Lichao Hao Omar Mechali, Jamshed Iqbal, Jingxiang Wang, Xiaomei Xie, Limei Xu IDepartment of Intelligent Science and Engineering School of Aeronautics and Astronautics, University of Electronic Science and Technology of Harbin Engineering University China, Chengdu, China Harbin, Heilongjiang Province, China • Marine environment exploration • This paper presents a novel robust based on multi-AUV leader-follower formation control for a team of quadrotors subject to • Task allocation strategies in targets lumped disturbances. searching and detecting. • A robust fixed-time position control • A mathematical model that depict the law is synthesized to achieve the territory competition process among desired geometrical pattern and to predators in natural district is constructed. track the reference trajectory. Fig. 1 Desired formation pattern and communication topology graph among the networked quadrotor system. • ROS/Gazebo simulation is • The definition of Pareto consensus is conducted to validate the designed put forward to describe the consensus Task area curve of each AUV and flight control algorithm. state of detecting task allocation. Path planning after negotiation

TA1-6 (5) 10:30-10:45 TA1-6 (6) 10:45-11:00 Water Surface Targets Recognition and Tracking A Path Planning Algorithm based on Artificial Potential Field Method and Ant Colony Algorithm Based on Improved YOLO and KCF Algorithms Zheping Yan, Liyuan Jiang and Di Wu Zhongli Ma, Yaohan Zeng, Lili Wu, Linshuai Zhang, Jiadi Li, Huixin Li College of Automation,Harbin Engineering University School of Control Engineering, Chengdu University of Information Technology Harbin, China Chengdu, China • Establishment and preprocessing of • Assign a set of path point and calculate the target data set. evaluation function of each point. • Select the next target path point and • Target recognition algorithm based trajectory point. on improved YOLO v3. • Determine whether the target is found and • Target recognition algorithm based whether the number of iterations exceeds on improved KCF. the maximum. • Test and analysis of water surface • Compare the distance between the UUV and last selected target path point, Repeat multi-target recognition and tracking the above steps. Path Planning Result algorithm based on improved YOLO and KCF. Overall Test Results

44 IEEE ICMA 2021 Conference Digest

TA2-6: Organized session: Soft Robotics

Session Chairs: Qiang Fu, Tianjin University of Technology Ruochen An, Kagawa University Online Conference Room 6, UTC+9(Tokyo Time): 11:15-12:15, Tuesday, 10 August 2021

TA2-6(1) 11:15-11:30 TA2-6(2) 11:30-11:45 Development of Pneumatic Driven Robot System A Robot Finger with Many Joints which can Entry and Retire from the Gas Pipe Driven by One Motor Using Shape Memory Gel Satoshi Kodama*, Koki Shindo*, Shotaro Senzaki*, Toko Nakamura*, Yoka Konishi* and Tendon-Driven Mechanism Yasuhiko Ohnuki**, Minoru Konno**, Hiroyuki Isii* and Atsuo Takanishi* Mitsuhiro Yamano*1, Naoya Hanabata*2, Akira Okamoto*2, Toshihiko Yasuda*1, Yasutaka * Faculty of Science and Engineering, Waseda University Tokyo, Japan Nishioka*1 ** Tokyo Gas Co., Ltd. Pipeline Network Company Kanagawa, Japan *1School of Engineering/ *2Graduate School of Engineering, The University of Shiga Prefecture, Japan • WATER-5 is a pneumatically driven robot that moves inside gas pipes and acquires states of pipe walls with its camera. MD Nahin Islam Shiblee, Kazunari Yoshida, Hidemitsu Furukawa, Riichiro Tadakuma Graduate School of Sci. and Eng. Yamagata University, Japan • The robot can enter 3-type curved sections and select branches. • Light and low-cost robot finger • The robot can retire from the gas pipes with a winder. using shape memory gel and tendon- driven mechanism is proposed. 540 Unit : mm • Many joints are driven by one motor. • The finger is applicable to both bending of selected joint and y Rear Elastic Flexible Front Active Bending adaptive grasping by its many joints. z x Gripper Actuator Gripper Module The Robot Finger WATER-5 Driven by One Motor

TA2-6(3) 11:45-12:00 TA2-6(4) 12:00-12:15 Development of a Wrist Sphygmomanometer with an Active Soft Mechanism Prototype of the Assist System for Early Stage of *Nobuaki YOSHIMURA, Yasutaka NISHIOKA, Toshihiko YASUDA, Mitsuhiro YAMANO, Standing Up with a Pneumatic Actuator Minoru TANIGUCHI, Chisato UESAKA, Tuyoshi HAMAGUCHI and Masao SHIMIZU *Department of Mechanical Systems Engineering, University of Shiga prefecture Ryuta Takemon, Toshihiko Yasuda, Yasutaka Nishioka and Mitsuhiro Yamano Hikone, Shiga, Japan Mechatronics Lab., University of Shiga prefecture, Hikone, Japan • More measurements during the day are • Though some elderly or handicapped persons needed to help manage one's physical need help to stand up, the early stage of standing- condition. up movement needs the big physical burden for • The purpose is to reduce the number of both the care-receiver and the care-giver. measurement procedures in a compact • We proposed the standing-up assist system with size by performing the measurement only the pneumatic actuator operated by care-receiver by placing the wrist from above. himself using the computer control system. • We designed three types of triangular • We designed new structure of the actuator for prism type soft actuators with active The wrist shortening of standing-up time. mechanisms. sphygmomanometer • We confirm the usefulness of the system • My own pulsation was measured using The actuator and system the device I made. mounted on a wheelchair by experiments.

45 Index of Session Chairs He, Wangli TA2-3 -A- Hirata, Hideyuki MP1-2 An, Ruochen MP2-2 An, Ruochen TA1-6 -J- An, Ruochen TA2-6 Jia, Chao MP3-3 Arai, Tatsuo TA1-2 Jin, Xiaoliang MP2-6 Jin, Xiaoliang MP3-6 -B-

Bao, Xianqiang TA2-5 -K- Bao, Xianqiang TP1-5 Kosuge, Kazuhiro MP2-2 Bu, Dongdong TP1-1 Koyanagi, Ken'ichi MP1-1 Bu, Dongdong TP2-1 Koyanagi, Ken'ichi MP2-1 Kurabayashi, Daisuke TA1-6 -F-

Fu, Qiang TP1-3 -L- Fu, Qiang TP2-3 Li, Ao TA1-4 Fu, Qiang TA2-6 Li, Ao TA2-4 Fukuda, Toshio MP1-1 Li, Chunying TA1-3 Li, Chunying TA2-3 -G- Li, Xinming TP1-3 Guo, Jian MP2-6 Li, Xinming TP2-3 Guo, Jian MP3-6 Liu, Yi TP1-4 Guo, Jin MP3-4 Liu, Yi TP2-4 Guo, Jin TP1-1 Lyu, Chuqiao MP3-4 Guo, Jin TP2-1 Guo, Shuxiang TP2-2 -M- Maeyama, Shoichi MP2-5 -H- Maki K., Habib TA2-2 He, Wangli TA1-3 Matsumoto, Mitsuharu TA2-4

Ming, Aiguo TA1-5 Morishima, Keisuke MP1-5 -X- Morishima, Keisuke TP1-2 Xie, Biyun MP2-1 -N- -Y-

Nagata, Fusaomi MP3-5 Yang, Cheng MP1-4 Yang, Cheng TP1-2 -P- Yang, Cheng TP2-2 Pan, Qinxue MA1-P Yang, Ziyi MA1-P Pan, Qinxue MP3-2 Yang, Ziyi TP1-4 Yang, Ziyi TP2-4 -S- Yin, He TA1-2 Saito, kei Takashi TA1-1 Yin, He TA2-2 Sawada, Hideyuki MP1-3 Sawada, Hideyuki MP2-3 -Z- Shi, Liwei MP1-4 Zhao, Yan TA2-1 Shi, Liwei MP2-4 Zhao, Yan TA1-4 Shi, Liwei MP1-5 Zheng, Liang MP1-3 Shi, Peng MP3-1 Zheng, Liang TA1-5 Shi, Peng TA1-1 Zheng, Lingling TA2-5 Shi, Peng TA2-1 Zheng, Lingling TP1-5 Song, Yu MP2-3 Zhou, Wei MP1-2 Song, Yu MP3-3 Zhou, Wei MP3-2 Zhou, Wei MP2-4 -T- Zhu, Junchao MP1-6

Takei, Toshinobu MP2-5 -W-

Watanabe, Keigo MP3-5 Wu, Haiyuan MP3-1 Wu, Haiyuan MP1-6

Index of Authors

Burk, Daniel MP2-3 -A-

Aldibaja, Mohammad TA2-3 -C- Amano, Kanako TP2-1 Cai, Liming MA1-P An, Ruochen MP1-5 Cai, Zhuocong TP1-1 An, Ruochen MP1-5 Calderón Ch., J. Alan MA1-P An, Ruochen TP2-3 Cao, Hui MA1-P An, Xingrun MA1-P Cao, Huifeng MP1-6 Anzai, Asahi TP2-3 Cao, Jiabo MP1-5 Arai, Tatsuo TA1-2 Cao, Meifen TA1-3 Ariram, Siva MP1-1 Chan, Hugo Hung-tin MP1-1 Ariram, Siva TA2-4 Chan, Hugo Hung-tin TP2-1 Ashley, Joshua MP1-1 Chang, Meile TA1-5 Attar, Vahida MP2-4 Chen, Chenfei MA1-P Ayankoso, Samuel A. TA2-2 Chen, Fang MP3-4 Ayankoso, Samuel A. TA2-2 Chen, Hong MP3-6 Chen, Jianjun MA1-P -B- Chen, Jiqiang TA1-5 Bai, Zhengkun MP2-3 Chen, Longmiao MP1-2 Bai, Zhengkun MP2-3 Chen, Min MP2-1 Bao, Pengxiao MP2-6 Chen, Min TA1-4 Bao, Pengxiao TP1-2 Chen, Peng MP3-3 Barriga, Benjamín MA1-P Chen, Tao MP1-6 Bensouda, Yousra TA2-2 Chen, Weiwei TA1-4 Bi, Qing TA1-4 Chen, Xiangyu TA2-3 Bi, Qing TA2-5 Chen, Xiaojun MP1-6 Bian, Xueqi TP1-5 Chen, Xin MP3-6 Bjørnøy, Erik MA1-P Chen, Xinlei MA1-P Boudihaj, Yasmine TA2-2 Chen, Xinlei MP3-5

Chen, Xuechao TP2-3 Dong, Minghai TA2-5 Chen, Yanyan MP2-4 Dong, Qi MP2-4 Chen, Yuxiang MP3-1 Dou, Jianping MP1-2 Chen, Zhang TP2-3 Du, Lifeng MP3-3 Chen, Zhengyang TP1-2 Du, Sihui MP1-6 Chen, Zhengyang TP2-4 Du, Yimeng MP2-1 Chen, Zhengyang TP2-4 Duan, Fuhai MA1-P Chen, Zhixin MP3-3 Duan, Lixiang MP3-4 Cheng, Handong TP1-5 Cheng, Hong TP1-1 -E- Cheng, Mengqi MP2-1 Ejima, Yoshimichi TP1-4 Cheng, Mengqi TP1-2 Ejima, Yoshimichi TP1-4 Chiang, Cheng-Ta TA1-5 Ejima, Yoshimichi TP1-4 Chu, Yuyi MA1-P Emaru, Takanori MP2-1 Cong, Xiaodan MA1-P Eto, Hiroaki TP1-1 Cortez, Ricardo MP3-3 Cosentino, Sarah TP1-1 -F- Cui, Lin MP3-3 Fan, Bingzheng TA1-2 Cui, Lin MP1-4 Fang, Yuefeng MP3-3 Cui, Mingshuai MA1-P Fei, Juntao MP3-4 Feng, Baolin TA1-2 -D- Feng, Jingjing MP2-4 Dai, Qi TP1-4 Feng, Junlong MP3-1 Dai, Qi TP1-4 Feng, Sheng MA1-P Dai, Xiaobo MA1-P Feng, Sheng MA1-P Dai, Xiaobo MA1-P Feng, Suxiang TP2-2 Dai, Yuehong MP1-1 Feng, Xiaojing TA1-1 Ding, Fuguang MP2-4 Franke, Jörg MP1-2 Ding, Yibin TA1-3 Fu, Jiajun TP1-1 Doi, Toshihide TP2-3 Fu, Mingyu MA1-P Dong, Chao MP3-2 Fu, Qiang TP1-1

Fu, Qiang TP2-1 Graichen, Knut MP2-3 Fu, Qiang TP2-1 Gu, Xiaoyi TA2-4 Fu, Yili TA1-2 Gu, Xinyue TP1-1 Fu, Yuzhuo MP1-5 Gu, Zixi TP1-1 Fu, Yuzhuo TA1-4 Guan, Lianwu MA1-P Fu, Yuzhuo TA1-4 Guan, Yisheng MP2-2 Fukao, Takanori TA2-1 Guan, Yisheng MP3-5 Furukawa, Hidemitsu TA2-6 Guan, Zhiwei MA1-P Furuya, Takahiro TA2-3 Guo, Hongpeng MA1-P Guo, Jian MP1-2 -G- Guo, Jian MP1-5 Gao, Baofeng MP1-6 Guo, Jian TP1-1 Gao, Fei MP1-1 Guo, Jian TP1-1 Gao, Fei TP2-1 Guo, Jian TP2-1 Gao, Feng MP2-6 Guo, Jian TP2-1 Gao, Huimin MA1-P Guo, Jian TP2-1 Gao, Qiang MA1-P Guo, Jian TP2-1 Gao, Qiang MA1-P Guo, Jian TP2-2 Gao, Qiang MA1-P Guo, Jian TP2-2 Gao, Qiang MA1-P Guo, Jian TA1-3 Gao, Ruiqi TA1-4 Guo, Jian TA2-3 Gao, Tianxin MP1-6 Guo, Jian TP2-3 Gao, Yanbin MA1-P Guo, Jian TP2-3 Gao, Yunfei MA1-P Guo, Jian TA2-4 Gao, Zhiqiang MP2-3 Guo, Jian TP1-4 Garrido, Ruben MP3-3 Guo, Jian TP1-4 Ge, Yao MP1-2 Guo, Jian TP2-4 Geng, Haipeng MP3-3 Guo, Jian TP2-4 Gofuku, Akio MP2-5 Guo, Jian TP1-5 Gonzalez Gonzalez, Jose Angel TA1-1 Guo, Jian TP1-5 Gou, Huabei MP2-3 Guo, Jin TP2-2

Guo, Kaijun MP1-6 Habib, Maki K. MP3-5 Guo, Kairui MP1-3 Habib, Maki K. TA2-2 Guo, Kairui MP1-4 Habib, Maki K. TA2-2 Guo, Lei TA1-4 Hamaguchi, Tuyoshi TA2-6 Guo, Mingqiu MP1-4 Han, Lianqiang TP2-3 Guo, Shuxiang MP3-1 Han, Siqi MP3-2 Guo, Shuxiang MP1-2 Han, Xiao MP1-6 Guo, Shuxiang MP1-2 Han, Zhenfeng TA1-4 Guo, Shuxiang MP3-4 Hanabata, Naoya TA2-6 Guo, Shuxiang MP1-5 Hao, Lichao TA1-6 Guo, Shuxiang MP1-5 Hao, Mengfu TP1-5 Guo, Shuxiang MP1-5 Hao, Zengchao MP3-1 Guo, Shuxiang MP1-5 Hashida, Kazuki TP2-3 Guo, Shuxiang MP2-6 Hashikura, Kotaro TA1-3 Guo, Shuxiang TP1-1 Hashimoto, Kenji TP2-3 Guo, Shuxiang TP1-1 Hassina, Hajar TA2-2 Guo, Shuxiang TP2-1 Hattori, Ikue TP1-4 Guo, Shuxiang TP2-1 Hattori, Yuya TA1-6 Guo, Shuxiang TP2-1 He, Fan TP1-5 Guo, Shuxiang TP2-1 He, Jiping TP1-1 Guo, Shuxiang TP1-2 He, Wangli TA1-3 Guo, Shuxiang TP1-2 He, Wenhao MP2-6 Guo, Shuxiang TP1-2 Hiramitsu, Tatsuhiro TA1-1 Guo, Shuxiang TP1-2 Hofmann, Christian MP1-2 Guo, Shuxiang TP2-2 Hoji, Rintaro MP2-5 Guo, Tongxu TP1-3 Holseker, Erlend MA1-P Guo, Xiao MP2-3 Hong, Xiaocui MP3-4 Guo, Xiao MP3-4 Hori, Shigeki MP2-3 Guo, Yue MP2-6 Hori, Yusuke TA2-1 Hou, Xihuan MP1-2 -H- Hou, Xihuan MP3-4

Hou, Xihuan MP1-5 Jia, Chao MP3-3 Hou, Xihuan MP1-5 Jia, Shichen TP1-5 Hsu, Chung-Yu TA1-5 Jia, Wenchuan TA1-3 Hu, Wenxin TP1-3 Jia, Wenchuan TP1-3 Hu, Yaqi TP2-1 Jiang, Hongzhou TA1-4 Huang, Hengjing TA1-4 Jiang, Liyuan TA1-6 Huang, Jian TP1-1 Jiao, Ao TA1-4 Huang, Lin MP2-1 Jin, Xiaoliang TP2-2 Huang, Nieyong MP2-2 Jin, Xiaoliang TP2-2 Huang, Qiang TA1-2 Jin, Xiaoliang TP2-2 Huang, Rui TP1-1 Huang, Xuchao MP2-2 -K- Huang, Yao-Chieh TA1-5 Kamal , Md Abdus Samad TA1-3 Huo, Fudong TP1-4 Kamegawa, Tetsushi MP2-5 Huo, Xiaojie TA2-4 Kan, Haoxuan TA1-2 Kanda, Shinsuke MP3-5 -I- Kang, Tong MP1-1 Igo, Naoki MA1-P Kang, Yi MA1-P Iinuma, Ryosuke TA2-1 Kapila, Vikram MP1-2 Ikeda, Takahiro TA2-1 Kato, Koichiro TP2-1 Imamura, Kenji TA1-1 Kato, Tatsuya MP2-5 Inada, Yuta TA2-3 Kato, Yuka TP2-1 Ino, Shingo TA2-3 Kawashima, Kenji MP2-3 Iqbal, Jamshed TP1-3 Kawauchi, Shizuki TA2-5 Iqbal, Jamshed TA1-6 Ke, Ang TP1-1 Ishii, Chiharu TP2-4 Kei Saito, Takashi TA1-1 Isii, Hiroyuki TA2-6 Khodabandeh, Arvin MA1-P Kijihana, Daisuke MP2-5 -J- Kimura, Noriyuki MA1-P Ji, Yuehui MA1-P Kitamura, Saya MP3-6 Jia, Chang MP1-6 Kiyozumi, Takaki TA1-3

Kobayashi, Kazuma MA1-P Li, Chao TA1-3 Kobayashi, Kohei TA1-1 Li, Chen MA1-P Kobayashi, Yukinori MP2-1 Li, Chen MP3-5 Kodama, Satoshi TA2-6 Li, Chiang-Shan R. MP1-6 Kogan, Vladyslav MP1-2 Li, Chuan MP3-2 Komura, Hideaki MP3-5 Li, Chunying MP1-5 Konishi, Yoka TA2-6 Li, Chunying MP1-5 Konno, Minoru TA2-6 Li, Chunying TP2-3 Kono, Takuro MP2-5 Li, Fangming MP1-6 Koyama, Daisuke TA1-3 Li, Guangfei MP1-6 Koyanagi, Ken'ichi MP1-1 Li, Guojiang TA1-2 Kubo, Yukihiro TA2-1 Li, Guoyuan MA1-P Kubota, Naoyuki TA1-3 Li, Hui TA2-5 Kurabayashi, Daisuke TA1-6 Li, Huixin TA1-6 Kurokawa, Daigo TP2-1 Li, Ji MP3-2 Kwon, Yuhwan MP2-5 Li, Ji TA1-3 Li, Ji TA1-5 -L- Li, Jiadi MA1-P Le, Zhiwen TA2-1 Li, Jiadi TP1-3 Lei, Yi MP3-6 Li, Jiadi TA1-6 Lengua, Juan Carlos MA1-P Li, Jianfeng TA2-1 Li, Ang MP2-3 Li, Juan MA1-P Li, Ang MP2-3 Li, Junfang MA1-P Li, Ao MP1-2 Li, Junjie MA1-P Li, Ao MP1-5 Li, Lailu TP1-2 Li, Ao MP1-5 Li, Lei MP2-1 Li, Ao TP1-2 Li, Li TA1-1 Li, Bin MA1-P Li, Lingxiao TA1-2 Li, Bin MA1-P Li, Lu TA1-2 Li, Bingjue MA1-P Li, Mingyang TA2-3 Li, Chao MP3-2 Li, Peiyu TP1-5

Li, Qi MA1-P Li, Zeyu TP2-2 Li, Qi TP1-4 Li, Zhihan MP1-5 Li, Qin MP2-1 Li, Zichen MP2-3 Li, Qing MP2-6 Li, Zichen MP2-3 Li, Sa MP3-2 Li, Ziyu MP3-3 Li, Sa TA1-5 Liang, Bin TP2-3 Li, Shuangyang MP3-2 Liang, Lihua TA2-1 Li, Shuangyang TA1-5 Liang, Zhihao MP3-5 Li, Wei MP3-2 Liao, Hongpeng MP1-1 Li, Wei TA2-2 Liao, Hongpeng TP2-1 Li, Wei TA1-5 Liao, Hongzhe TA1-2 Li, Xiaojian TA1-2 Liao, Wei-Hsin MP1-1 Li, Xinguang TA1-1 Liao, Wei-Hsin TP2-1 Li, Xinming TP2-2 Liao, Xianguo TP2-1 Li, Xinming TP2-2 Liao, Xiaocun MP1-5 Li, Xinming TP2-2 Liao, Xiaocun TA1-4 Li, Xuece MP3-3 Liao, Xiaocun TA1-4 Li, Xuesheng MA1-P Lieret, Markus MP1-2 Li, Xuesheng MP3-5 Lin, Geqi TA1-3 Li, Xun TP1-1 Lin, Jinshu TA2-5 Li, Yanmei MP2-6 Lin, Mengxiang MP3-3 Li, Yao MP3-1 Lin, Yuhang MP1-2 Li, Yibin TA1-4 Lin, Yuxiu MP3-6 Li, Yingna MP3-2 Liu, Bin MA1-P Li, Yuchuan MP3-2 Liu, Chunping MA1-P Li, Zan MP1-2 Liu, Cong MA1-P Li, Zan MP3-4 Liu, Dan TA1-2 Li, Zan MP1-5 Liu, Dongdong MP1-2 Li, Zan MP1-5 Liu, Fanming MP1-6 Li, Zejie MP2-1 Liu, Hongli MP3-1 Li, Zepeng TP1-5 Liu, Jiafeng MP2-3

Liu, Junjie MA1-P Lu, Zhizhong MP1-6 Liu, Junjie MA1-P Luo, Cong MP3-5 Liu, Junjie MA1-P Luo, Rui MP3-2 Liu, Junjie MA1-P Luo, Yuechang TP1-5 Liu, Lei MA1-P Lyu, Chuqiao TP1-2 Liu, Meng MP1-2 Lyu, Chuqiao TP2-2 Liu, Meng MP1-5 Lyu, Chuqiao TP2-2 Liu, Meng TP1-2 Lyu, Dingchong MP3-3 Liu, Qian MP2-4 Liu, Qian MP2-4 -M- Liu, Qiang MP2-2 Ma, Boye TP1-3 Liu, Xi TA2-5 Ma, Huichen MP2-2 Liu, Xiaoming TA1-2 Ma, Shugen TA1-3 Liu, Xinyi TP2-1 Ma, Shugen TP1-3 Liu, Yali MP3-2 Ma, Xiang MP3-2 Liu, Yana MP2-1 Ma, Youchun TP1-2 Liu, Yanhong MP3-1 Ma, Youchun TP2-2 Liu, Yi TP1-4 Ma, Youchun TP2-2 Liu, Yixiang TA1-4 Ma, Youjie MA1-P Liu, Yixiang TA2-5 Ma, Youjie MP1-3 Liu, Yongchen TA1-4 Ma, Youjie MP1-3 Liu, Yu TP2-3 Ma, Youjie MP1-3 Liu, Yunqing MA1-P Ma, Youjie MP1-3 Liu, Zhanshuo MP3-1 Ma, Youjie MP1-3 Liu, Zhi MP3-6 Ma, Youjie MP2-3 Lou, Wenjie MP3-4 Ma, Youjie MP1-4 Lozano, John MA1-P Ma, Youjie MP2-4 Lu, Ben MP1-5 Ma, Youjie MP2-4 Lu, Ben TA1-4 Ma, Youjie MP2-4 Lu, Ben TA1-4 Ma, Youjie MP3-4 Lu, Yangyang TA1-5 Ma, Zhongli MA1-P

Ma, Zhongli TP1-3 Motonaka, Kimiko MP2-5 Ma, Zhongli TA1-6 Mou, Fangli TA1-2 Maeyama, Shoichi MP2-5 Mao, Zemin MA1-P -N- Mao, Zemin MA1-P Nagahama, Kotaro MP3-6 Mata Juarez, Omar TA1-1 Nagahara, Masaaki MP2-5 Matsuhira, Nobuto TP2-1 Nagai, Isaku MP2-5 Matsuhiro, Ko TP1-1 Nagai, Isaku MP3-5 Matsumoto, Mitsuharu TA2-4 Nagai, Isaku MP3-5 Mechali, Abdesselam TP1-3 Nagai, Isaku MP3-5 Mechali, Omar MP3-6 Nagai, Isaku TA2-5 Mechali, Omar TP1-3 Nagata, Fusaomi MP3-5 Mechali, Omar TA1-6 Nagata, Fusaomi TA2-2 Mehta, Shreya Rajendra MP2-4 Nakagawa, Yuto MP3-6 Meng, Fanxu TP1-2 Nakamura, Issa MP2-5 Meng, Fanxu TP2-4 Nakamura, Toko TA2-6 Meng, Fanxu TP2-4 Nambo, Hidetaka MP3-4 Miki, Kohei MP3-5 Ni, Yu TA1-1 Mills, James K. MP1-1 Nie, Heng MA1-P Mills, James K. MP2-2 Nie, Wenlong TA1-5 Min, Bo MP1-3 Ning, Yu TA1-2 Minamoto, Masahiko MP2-3 Nishioka, Yasutaka TP1-1 Mitsui, Satoshi MA1-P Nishioka, Yasutaka TA2-6 Miyamoto, Kazuya MP3-5 Nishioka, Yasutaka TA2-6 Miyazaki, Tetsuro MP2-3 Nishioka, Yasutaka TA2-6 Miyazawa, Yuki TA2-3 Nishioka, Yuki TP1-4 Miyoshi, Seiji MP2-5 Niu, Jiaxi MP1-6 Mizobuchi, Yoshinobu TA2-3 Niu, Yong MP3-1 Mizukami, Kazuhiko TP1-1 Noda, Kentaro MP1-1 Mo, Yaqiang MP3-6 Nussbaum, Doron MA1-P Molina, Arturo TA1-1

Qiu, Xiaoqing MA1-P -O-

Ohnuki, Yasuhiko TA2-6 -R- Ohtsuka, Hirofumi MP2-5 Ravankar, Ankit A. MP2-1 Okamoto, Akira TA2-6 Röning, Juha MP1-1 Okamura, Hajime TA2-3 Röning, Juha TA2-4 Oko, Akitaka TA2-3 Ruan, Jinliang MA1-P Okuhata, Kazuo TA2-3 Onoyama, Hiroyuki TA2-1 -S- Ootsubo, Katsutoshi TA2-1 Sabater, José Maria MP2-2 Oshima, Toru MP1-1 Salih-Alj, Yassine TA2-2 Otake, Shinichi TA2-4 Satake, Toshifumi MA1-P Ou, Jiajun MP3-4 Sebti, Nada TA2-2 Seki, Hiroaki TA1-1 -P- Senzaki, Shotaro TA2-6 Pan, Junqing MP1-3 Shan, Chenghao MP2-1 Pan, Junqing MP1-3 Shang, Peng MP2-2 Pan, Qinxue MP3-2 Shang, Peng TA2-5 Pan, Qinxue TA1-5 Shao, Hong TP1-3 Patil, Sneha Sunil MP2-4 Shao, Lei MP3-1 Peng, Youfan MP2-1 Shao, Lei MP3-2 Peng, Youfan TA1-5 Shao, Lei MP1-4 Peng, Yueyan TA2-2 Shao, Lei TA1-3 Pennanen, Tuulia MP1-1 Shao, Lei TA1-5 Pennanen, Tuulia TA2-4 Shao, Xin MP3-2 Ponce, Pedro TA1-1 Shao, Xiuqiang MP1-2 Shen, Yage MP3-5 -Q- Shi, Ce MP2-6 Qin, Jing MP1-1 Shi, Ce MP2-6 Qiu, Chenguang MA1-P Shi, Chaochao TP1-2 Qiu, Jing TP1-1 Shi, Liwei MP3-4

Shi, Liwei MP1-5 Shuxiang, Guo TP2-4 Shi, Liwei MP1-5 Shuxiang, Guo TP2-4 Shi, Liwei MP2-6 Shuxiang, Guo TP2-4 Shi, Liwei TP1-2 Shuxiang, Guo TP2-4 Shi, Peng TP2-2 Shuxiang, Guo TP1-5 Shi, Peng TP2-2 Shuxiang, Guo TP1-5 Shi, Peng TP2-2 Shuxiang, Guo TA1-6 Shi, Xiaojun TA1-1 Si, Xiaxi MP1-1 Shi, Yue MP1-6 Sneltvedt, Isak Gamnes MA1-P Shi, Yunde MP1-4 Solano, Gonzalo MA1-P Shi, Zelin MP2-6 Song, Haitao MP2-6 Shiblee, MD Nahin Islam TA2-6 Song, Rui TA2-5 Shimizu, Masao TA2-6 Song, Shouan MP3-1 Shindo, Koki TA2-6 Song, Yaowei MP3-5 Shirguppi, Nikita Shailesh MP2-4 Song, Yu MA1-P Shuai, Liguo TA1-2 Song, Yu MA1-P Shuxiang, Guo TP2-2 Song, Yu MA1-P Shuxiang, Guo TP2-2 Su, Diankang MP1-3 Shuxiang, Guo TP2-2 Su, Diankang MP1-3 Shuxiang, Guo TP2-2 Su, Tao TA1-1 Shuxiang, Guo TP2-2 Suganuma, Naoki TA2-3 Shuxiang, Guo TA1-3 Sugiyama, Ryo TP2-4 Shuxiang, Guo TA2-3 Sun, Liran MP2-3 Shuxiang, Guo TP2-3 Sun, Wentao MP3-1 Shuxiang, Guo TP2-3 Sun, Wentao TA1-3 Shuxiang, Guo TP2-3 Sun, Wentao TA1-5 Shuxiang, Guo TA2-4 Sun, Yanhua MP2-2 Shuxiang, Guo TP1-4 Sun, Yi TA1-3 Shuxiang, Guo TP1-4 Sun, Yi TP1-3 Shuxiang, Guo TP1-4 Sun, Yue TP2-4 Shuxiang, Guo TP2-4 Sun, Zefa TP1-1

Suzuki, Kiriro TA1-1 Tendeng, Awa TP2-3 Tian, Jiale MA1-P -T- Tian, Jiale TA2-5 Tada, Yusuke TA1-3 Tian, Yiran MA1-P Tadakuma, Riichiro TA2-6 Tikanmäki, Antti MP1-1 Tafur, Julio MA1-P Tikanmäki, Antti TA2-4 Tajima, Nina TP2-1 Tokairin, Koichi TA1-3 Takada, Hiroaki MP3-6 Tomura, Toyoaki MA1-P Takahashi, Satoshi TP1-4 Torige, Akira MP2-5 Takahashi, Satoshi TP1-4 Tsuji, Tokuo TA1-1 Takahashi, Satoshi TP1-4 Tsujinaka, Rikuya TP1-1 Takanishi, Atsuo TP1-1 Tsukagoshi, Takuya MP1-1 Takanishi, Atsuo TA2-6 Takata, Daisuke MP1-1 -U- Takei, Toshinobu MP2-5 Ueda, Kazuma MA1-P Takei, Toshinobu MP2-5 Ueda, Yasuhiro TA2-3 Takemasa, Koki TA1-1 Ueki, Satoshi TA2-1 Takemon, Ryuta TA2-6 Uesaka, Chisato TA2-6 Takeshita, Masahiro MP2-5 Umetani, Yamato MP2-3 Tamamoto, Takumi MP1-1 Utsumi, Takashi MP3-5 Tan, Chao TA1-5 Tan, Jianhui MP1-2 -V- Tan, Junbo MP1-3 Vachmanus, Sirawich MP2-1 Tanaka, Kanji TP2-3 Velazquez Espitia, Victor Miguel TA1-1 Tang, Wei TA2-3 Vivas, Andres MP2-2 Tang, Xiaoying MP1-6 Völz, Andreas MP2-3 Tang, Xiaoying MP1-6 Taniguchi, Minoru TA2-6 -W- Tao, Mo MP3-6 Wang, Bin MP3-5 Tendeng, Awa MP1-5 Wang, Bo MP3-3 Tendeng, Awa MP1-5 Wang, Changliang MP1-3

Wang, Chunguang MA1-P Wang, Xiaoming MA1-P Wang, Chunjie MP3-3 Wang, Xin TP1-5 Wang, Chunjie MP1-4 Wang, Xin TP1-5 Wang, Hui MP2-6 Wang, Xueqian MP1-3 Wang, Hui MP2-6 Wang, Yajing TP1-3 Wang, Hui MP2-6 Wang, Yan MA1-P Wang, Jian MA1-P Wang, Yanan TA1-2 Wang, Jian MA1-P Wang, Yanlong TA1-4 Wang, Jian MP3-6 Wang, Yingxin MP3-6 Wang, Jiaqi MA1-P Wang, Yongdong MP2-5 Wang, Jingxiang TA1-6 Wang, Yuanhui MP2-4 Wang, Junyao MP1-1 Wang, Yuchao MA1-P Wang, Jutao MA1-P Wang, Yue MP2-4 Wang, Kaidi MP2-1 Wang, Yue TP1-2 Wang, Kaidi TP1-2 Wang, Yue TP2-2 Wang, Manyi MP3-3 Wang, Yue TP2-2 Wang, Qiao MP3-2 Wang, Yuezong MP2-1 Wang, Qiao MP3-2 Wang, Yuezong TA1-5 Wang, Qiusu MA1-P Wang, Yunliang MA1-P Wang, Ran TP2-3 Wang, Yunliang MA1-P Wang, Shaolong TP2-3 Wang, Yunliang MP2-3 Wang, Shaoping MP3-6 Wang, Yunliang MP2-3 Wang, Shuguo TA1-2 Wang, Yunliang MP3-6 Wang, Shuo TA2-1 Wang, Zhe TA2-5 Wang, Shuoyu TP1-2 Wang, Zhentao TP2-1 Wang, Shuoyu TA2-3 Wang, Zhiqiang MP2-6 Wang, Tao MP3-2 Wang, Zhisen MA1-P Wang, Wuyi MP1-6 Watanabe, Keigo MP2-5 Wang, Xiangke TP1-3 Watanabe, Keigo MP3-5 Wang, Xiaofei MA1-P Watanabe, Keigo MP3-5 Wang, Xiaolin MP1-6 Watanabe, Keigo MP3-5

Watanabe, Keigo MP3-5 Wu, Lili TP1-3 Watanabe, Keigo TA2-5 Wu, Lili TA1-6 Watanabe, Takuya MP2-5 Wu, Litao MA1-P Watanabe, Ryuya TP1-1 Wu, Qiong TP1-4 Watanabe, Yousuke MP3-6 Wu, Qiong TP1-4 Wei, Chunyu TA1-1 Wu, Qiong TP1-4 Wei, Congcong MP1-3 Wu, Wentao TA1-5 Wei, Jiaxin MP3-3 Wu, Xiangfei MP3-6 Wei, Liuxuan MP3-3 Wu, Xiongwei TA2-3 Wei, Meng MP3-2 Wu, Yanjuan MA1-P Wei, Meng MP3-2 Wu, Yanjuan MA1-P Wei, Mingzhu MP3-6 Wu, Yanjuan MP2-3 Wei, Yanhui TP1-3 Wu, Yanjuan MP2-3 Wei, Zikang MA1-P Wu, Yanjuan MP3-6 Wen, Guoqiang MA1-P Wu, Yingzhe TA2-5 Wen, Hongyu MP1-3 Wu, Yongxiang TA1-2 Wen, James Zhiqing TA2-1 Wu, Yujin TA2-1 Wen, James Zhiqing TA2-2 Wu, Chenguang MP3-2 -X- Wu, Dan TA1-2 Xia, Dan MP1-4 Wu, Di TA1-6 Xia, Dan MP1-5 Wu, Di TA1-6 Xia, Debin MP1-2 Wu, Haiyuan MP3-1 Xia, Debin MP3-4 Wu, Haiyuan MP3-1 Xia, Debin MP1-5 Wu, Jiabin MA1-P Xia, Debin MP1-5 Wu, Jiabin MA1-P Xia, Yuanhao TA2-4 Wu, Jin MA1-P Xiang, Chaoqun MP3-5 Wu, Jinglong TP1-4 Xiao, Feiyun MP2-4 Wu, Jinglong TP1-4 Xiao, Meng TP1-5 Wu, Jinglong TP1-4 Xiao, Nan MP2-1 Wu, Lili MA1-P Xiao, Nan TP1-2

Xiao, Xuan MP1-4 Xu, Yujie MA1-P Xie, Bingjie MP2-4 Xie, Bingjie MP2-4 -Y- Xie, Biyun MP1-1 Yamada, Hironao TA2-1 Xie, Jirong TA1-1 Yamada, Kou TA1-3 Xie, Xiaomei MP2-1 Yamada, Shunya MP3-6 Xie, Xiaomei MP3-6 Yamada, Tetsuya TA2-3 Xie, Xiaomei TP1-3 Yamaguchi, Shota MA1-P Xie, Xiaomei TA1-4 Yamano, Mitsuhiro TP1-1 Xie, Xiaomei TA1-6 Yamano, Mitsuhiro TA2-6 Xing, Huiming MP3-1 Yamano, Mitsuhiro TA2-6 Xiong, Ling TA2-5 Yamano, Mitsuhiro TA2-6 Xu, Fashu TP1-1 Yamashita, Koumei TA1-3 Xu, Jigang TA2-3 Yamazaki, Kimitoshi MP3-6 Xu, Jigang TP2-3 Yan, Chenyang MP1-4 Xu, Jigang TP2-3 Yan, Min MP2-4 Xu, Jigang TA2-4 Yan, Zheping TA1-6 Xu, Jigang TP1-5 Yanagi, Jiei TP1-1 Xu, Lang MP3-2 Yanagisawa, Takuto TA1-6 Xu, Lang TA1-5 Yanase, Reo TA2-3 Xu, Limei TP1-3 Yang, Bin TA1-4 Xu, Limei TA1-6 Yang, Bin TA2-5 Xu, Qing TA1-3 Yang, Binkai MP1-6 Xu, Qiwei MA1-P Yang, Cheng TP2-4 Xu, Qiwei MP3-5 Yang, Chenguang TP1-2 Xu, Xiaoning MP3-1 Yang, Chenguang TP2-2 Xu, Xiaoning TA1-5 Yang, Chenguang TP2-2 Xu, Xiaoyu TA1-5 Yang, Guang TP1-2 Xu, Xichen MA1-P Yang, Guang TA2-3 Xu, Xinxin MP3-2 Yang, Hongbo MA1-P Xu, Xinxin MP3-2 Yang, Jiajia TP1-4

Yang, Jiajia TP1-4 Yin, Jie MA1-P Yang, Jiajia TP1-4 Yin, Jie MP1-3 Yang, Jun MP1-3 Yin, Jie MP2-4 Yang, Kunpeng TA1-5 Yin, Shan TA2-5 Yang, Lei MP3-1 Yin, Zhenshuo MP2-2 Yang, Liu TP1-5 Yoshida, Kazunari TA2-6 Yang, Qiongnan MA1-P Yoshida, Satomi TA2-3 Yang, Weichang MP2-2 Yoshimura, Nobuaki TA2-6 Yang, Xuxi MA1-P Yu, Hao MP1-4 Yang, Xuyun TA2-2 Yu, Hongtao TP1-4 Yang, Yuanshuang TP1-5 Yu, Lie MA1-P Yang, Zhen MA1-P Yu, Lie MA1-P Yang, Zhi MP2-2 Yu, Lijun MP2-6 Yang, Ziyi TP1-4 Yu, Lijun MP2-6 Yao, Bo Wen MP1-1 Yu, Lijun MP2-6 Yao, Jiayi MP3-3 Yu, Yiyang TP1-4 Yao, Lichang TP1-4 Yuan, Hang MP2-1 Yao, Lichang TP1-4 Yuan, Hang TP1-2 Yasuda, Toshihiko TP1-1 Yuan, Jiace MP2-3 Yasuda, Toshihiko TA2-6 Yuan, Jianjun TA1-3 Yasuda, Toshihiko TA2-6 Yuan, Jianjun TP1-3 Yasuda, Toshihiko TA2-6 Yuan, Kui MP2-6 Yasui, Toshiaki TA2-3 Yuan, Li TA2-4 Ye, Min MP3-2 Yuan, Ruikun MA1-P Ye, Min MP3-2 Yue, Caicheng MP1-2 Ye, Tingfeng MA1-P Yuta, Shin'ichi MP2-5 Ye, Xiufen MP3-1 Yin, He MP1-2 -Z- Yin, He MP3-4 Zeng, Huapeng MP1-4 Yin, He MP1-5 Zeng, Jianhui MA1-P Yin, He TP1-2 Zeng, Qi MP1-6

Zeng, Tonghui MP3-6 Zhang, Linshuai TA1-6 Zeng, Yaohan MA1-P Zhang, Liukun TA1-6 Zeng, Yaohan TA1-6 Zhang, Liuqing TP2-4 Zeng, Yuehan TP1-3 Zhang, Michael MA1-P Zeng, Zhenni TA2-5 Zhang, Qiyuan TP2-3 Zhan, Hang TA1-2 Zhang, Quchang MP2-6 Zhan, Yu MA1-P Zhang, Sai MA1-P Zhan, Yu MA1-P Zhang, Sai MA1-P Zhang, Anwei TA2-4 Zhang, Sai MP3-6 Zhang, Baofeng MA1-P Zhang, Sen MA1-P Zhang, Baofeng MP1-6 Zhang, Shihui MA1-P Zhang, Baofeng TA1-1 Zhang, Shuai MP2-6 Zhang, Baofeng TA1-5 Zhang, Shumei TA1-5 Zhang, Bing TP1-3 Zhang, Songtao TA2-1 Zhang, Chao TP1-3 Zhang, Tao MP3-5 Zhang, Chongyu MA1-P Zhang, Tong MP3-4 Zhang, Fusheng TA1-5 Zhang, Wenzhe TA1-2 Zhang, Hao TP1-3 Zhang, Winston MA1-P Zhang, Haobo MA1-P Zhang, Xiao MP2-1 Zhang, Haoran TA1-2 Zhang, Xiaodong TA1-1 Zhang, Houxiang MA1-P Zhang, Xiaoman TA1-2 Zhang, Jingxi TA2-3 Zhang, Xinghui TA1-1 Zhang, Jinwei TA1-5 Zhang, Xinyun MP3-4 Zhang, Jiwei MP3-1 Zhang, Yongcheng MP3-6 Zhang, Jiwei MP3-1 Zhang, Yongyan MP1-4 Zhang, Junfeng MP2-4 Zhang, Yue MP2-6 Zhang, Justin MA1-P Zhang, Yunmiao MP3-2 Zhang, Juzhong MA1-P Zhang, Yunmiao TA1-5 Zhang, Lijun MP3-4 Zhang, Zhao MP1-6 Zhang, Linshuai MA1-P Zhang, Zhili MA1-P Zhang, Linshuai TP1-3 Zhang, Zhili TA1-1

Zhang, Zhongyin TP1-2 Zhong, Weibao MP1-4 Zhang, Zhuoliang MP1-5 Zhou, Chao MP1-5 Zhang, Zhuoliang TA1-4 Zhou, Chao TA1-4 Zhang, Zhuoliang TA1-4 Zhou, Chao TA1-4 Zhang, Zixuan MA1-P Zhou, Decheng TA2-1 Zhang, Zixuan TP1-5 Zhou, Decheng TA2-2 Zhang, Zixuan TP1-5 Zhou, Jian MA1-P Zhao, Chaodan MP1-2 Zhou, Jian MA1-P Zhao, Dan MP1-6 Zhou, Jian MP2-2 Zhao, Feng TA1-1 Zhou, Jian TA2-5 Zhao, Han TP2-4 Zhou, Junjie MP2-2 Zhao, Hanqing MP3-4 Zhou, Mengni TP1-4 Zhao, Jianping MP3-1 Zhou, Mengni TP1-4 Zhao, Kai TP1-4 Zhou, Mugen MP1-2 Zhao, Shunli MP1-4 Zhou, Mugen MP3-4 Zhao, Xia MP1-2 Zhou, Mugen MP1-5 Zhao, Xuan MP1-1 Zhou, Shengmin MA1-P Zhao, Xuan TP2-1 Zhou, Songxin MP2-2 Zhao, Yang TA1-5 Zhou, Wei TP1-2 Zhao, Yiwen MA1-P Zhou, Wei TP2-4 Zhao, Yiwen MA1-P Zhou, Wei TP2-4 Zhao, Yiwen MP3-6 Zhou, Weidong MP2-1 Zhao, Yu MA1-P Zhou, Xuesong MA1-P Zheng, Lingling TA1-6 Zhou, Xuesong MP1-3 Zheng, Mingliang MP1-6 Zhou, Xuesong MP1-3 Zheng, Wei MP3-3 Zhou, Xuesong MP1-3 Zheng, Xudong TP2-3 Zhou, Xuesong MP1-3 Zheng, Zhiheng TA1-1 Zhou, Xuesong MP1-3 Zhong, Jun TA1-4 Zhou, Xuesong MP2-3 Zhong, Weibao MP1-3 Zhou, Xuesong MP1-4 Zhong, Weibao MP1-3 Zhou, Xuesong MP2-4

Zhou, Xuesong MP2-4 Zhou, Xuesong MP2-4 Zhou, Xuesong MP3-4 Zhou, Xufeng MP1-5 Zhou, Yinuo MP2-2 Zhou, Zeyu TA2-3 Zhou, Zhengshu MP3-6 Zhou, Zhuo MP2-2 Zhu, Deheng TA1-1 Zhu, Guiming MP2-4 Zhu, Haiqiang MA1-P Zhu, Jiangang MA1-P Zhu, Jiangang MA1-P Zhu, Junchao MP1-6 Zhu, Junchao MP3-6 Zhu, Junchao TA1-5 Zhu, Ming MP2-3 Zhu, Ming MP3-4 Zhu, Qingliang MP3-6 Zhu, Sanying MP2-2 Zhu, Wen TA1-4 Zhu, Xianjin TP2-3 Zhu, Yaqiao MA1-P Zhu, Yue TA1-4 Zhuang, Chungang MP2-6 Zhuo, Feng TP1-3 Zhuo, Jie MP3-4 Zou, Qianqian MP1-5 Zou, Qianqian TA1-4 Zou, Qianqian TA1-4

Memo

Advisory Council Honorary Chairs: T. J. Tarn, Washington University, USA Toshio Fukuda, Meijo University, Japan Advisory Council Chairs: IEEE ICMA 2022 Xiaoyun Qin, Guangxi Univ. of Science and Tech., China Tianyou Chai, Northeastern University, China 2022 IEEE International Conference on Hegao Cai, Harbin Institute of Technology, China Jie Chen, Tongji University, China Mechatronics and Automation A.A. Goldenberg, University of Toronto, Canada Kazuhiro Kosuge, The University of Hong Kong August 7-10, 2022, Guilin, Guangxi, China Paolo Dario, Scuola Superiore Sant'Anna, Italy Masayoshi Tomizuka, UC Berkeley, USA Mario A. Rotea, University of Massachusetts, USA Ju-Jang Lee, KAIST, Korea Ren C. Luo, National Taiwan University, China Yu Yao, Harbin Engineering University, China Yanrong Li, UESTC, China Huadong Yu, CUST, China Qingxin Yang, Tianjin University of Tech., China General Chairs: Hideyuki Hirata, Kagawa University, Japan Simin Li, Guangxi Univ. of Science and Tech., China Co-sponsors: IEEE R & A Society, Guangxi University of Science and Technology, Kagawa University General Co-Chairs: Shigeki Sugano, Waseda University, Japan Technical Co-sponsors: CAA, RSJ, JSME, SICE, JSPE, TJUT, HIT, HEU, UEC, UESTC, CUST William R. Hamel, University of Tennessee, USA James K. Mills, Univ. of Toronto, Canada Qiang Huang, Beijing Institute of Technology, China Index Sergej Fatikow , University of Oldenburg, Germany Call for Papers Darwin G. Caldwell, Italian Institute of Tech., Italy The 2022 IEEE International Conference on Mechatronics and Automation (IEEE Yuxin Zhao, Harbin Engineering University, China ICMA 2022) will take place in Guilin, Guangxi, China from August 7-10, 2022. Lixin Dong, City University of Hong Kong, China Zhan Yang, Suzhou University, China Guilin is a prefecture-level city in the northeast of China's Guangxi Zhuang Wubin Xu, Guangxi Univ. of Science and Tech., China Autonomous Region. It is situated on the west bank of the Li River and borders Enzeng Dong, Tianjin University of Tech., China Hunan to the north. Its name means "forest of sweet osmanthus", owing to the Yuanqing Xia, Beijing Institute of Technology, China large number of fragrant sweet osmanthus trees located in the region. The city has Program Chair: Jin Guo, Beijing Institute of Technology, China long been renowned for its scenery of karst topography. Program Co-Chairs: As the host city of IEEE ICMA 2022, Guilin not only provides the attendees with Jian Li, Guangxi Univ. of Science and Tech., China a great venue for this event, but also an unparalleled experience in Chinese history Hideyuki Sawada, Waseda University, Japan and culture. You are cordially invited to join us at IEEE ICMA 2022 in Guilin. Ryuma Niiyama, The University of Tokyo, Japan Kevin Lynch, Northwestern University, USA The objective of ICMA 2022 is to provide a forum for researchers, educators, Cecilia Laschi, Scuola Superiore Sant’Anna, Italy engineers, and government officials involved in the general areas of mechatronics, Stefan Byttner, Halmstad University, Sweden robotics, automation and sensors to disseminate their latest research results and Wan Kyun Chung, POSTECH, Korea exchange views on the future research directions of these fields. Yili Fu, Harbin Institute of Technology, China Anqi Qiu, National Univ. of Singapore, Singapore The topics of interest include, but not limited to the following: Jian Huang, Huazhong Univ. of Science and Tech., China - Intelligent mechatronics, robotics, biomimetics, automation, control systems, Xiufen Ye, Harbin Engineering University, China - Opto-electronic elements and Materials, laser technology and laser processing Organizing Committee Chairs: - Elements, structures, mechanisms, and applications of micro and nano systems Shuxiang Guo, Kagawa University, Japan Guangxing Tan, Guangxi Univ. of Science and Tech., China - Teleoperation, telerobotics, haptics, and teleoperated semi-autonomous systems Xueshan Gao, Beijing Institute of Technology, China -Sensor design, multi-sensor data fusion algorithms and wireless sensor networks Organizing Committee Co-chairs - Biomedical and rehabilitation engineering, prosthetics and artificial organs Youping Sun, Guangxi Univ. of Science and Tech. - Control system modeling and simulation techniques and methodologies Aiguo Ming, U. of Electro-Communications, Japan Tutorials/Workshop Chairs: - AI, intelligent control, neuro-control, fuzzy control and their applications Guangjun Liu, Ryerson University, Canada - Industrial automation, process control, manufacturing process and automation Huosheng Hu, University of Essex, U.K Yuwei Lu, Guangxi Univ. of Science and Tech., China Contributed Papers: All papers must be submitted in PDF format prepared Invited/Organized Session Chairs: Hiroyuki Nabae, Tokyo Institute of Technology, Japan strictly Guofu Wang, Guangxi Univ. of Science and Tech., China following the IEEE PDF Requirements for Creating PDF Documents for IEEE Meisuke Morishima, Osaka University, Japan Xplore. The standard number of pages is 6 and the maximum page limit is 8 pages Yongfu Li, City University of Hong Kong, China with extra payment for the two extra pages. See detailed instructions in the Yasuhisa Hirata, Tohoku University, Japan conference web site. All papers accepted by IEEE ICMA 2022 will be indexed by Awards Committee Co-chairs Hong Zhang, Southern U. of Science and Tech., China EI and included in IEEE Xplore®. Extensions of selected papers will be Qinxue Pan, Beijing Institute of Technology, China published in a regular or a special issue of the journals of IJMA. Publications Chair: Organized Sessions: Proposals with the title, the organizers, and a brief statement Liwei Shi, Beijing Institute of Technology, China of purpose of the session must be submitted to an OS Chair by March 20, 2022. Publicity Chairs: Tutorials & Workshops: Proposals for tutorials and workshops that address Jinjun Shan, York University, Canada Qiang Fu, Tianjin University of Technology, China related topics must be submitted to one of the Tutorial/Workshop Chairs by May 1, Xiankun Lin, Guangxi Univ. of Science and Tech., China 2022. Finance Chairs: Important Dates: Jian Guo, Tianjin University of Technology, China April 10, 2022 Full papers and organized session proposals Hidenori Ishihara, Kagawa University, Japan Ying Wen, Guangxi Univ. of Science and Tech., China May 1, 2022 Proposals for tutorials and workshops Local Arrangement Chairs: May 15, 2022 Notification of paper and session acceptance Xiaolong Yang, Guangxi Univ. of Science and Tech., China June 1, 2022 Submission of final papers in IEEE PDF format Yan Zhao, Hebei University of Technology, China Secretariats: For detailed up-to-date information, please visit the IEEE ICMA conference Z. Yang, C. Li, X. Li, Kagawa University, Japan C. Lv, H. Yin, A. Li, Beijing Institute of Tech., China website at: L. Zheng, Guangxi Univ. of Science and Tech., China http://2022.ieee-icma.org

IEEE ICMA 2021 International Conference

本会議は公益財団法人 NSK メカトロニクス技術高度化財団により助成された

This Conference Is Sponsored by NSK. Foundation for the Advancement of

Mechatronics