Peter Sapaty and Joaquim Filipe (Eds.) Multi-Agent Robotic Systems Proceedings of the 3rd International Workshop on Multi-Agent Robotic Systems MARS 2007 In conjunction with ICINCO 2007 Angers, France, May 2007 INSTICC PRESS Portugal ii Volume Editors Peter Sapaty The University of Aizu Aizu-Wakamatsu, Japan and Joaquim Filipe Polytechnic Institute of Setúbal / INSTICC Setúbal, Portugal 3rd International Workshop on Multi-Agent Robotic Systems Angers, France, May 2007 Peter Sapaty and Joaquim Filipe (Eds.) Copyright © 2007 INSTICC PRESS All rights reserved Printed in Portugal ISBN: 978-972-8865-86-6 Depósito Legal: 257880/07 iii Foreword This volume contains all the papers presented at MARS-2007 (3rd International Workshop on Multi-Agent Robotic Systems), held in Angers/France, in May 2007, as a satellite workshop of the ICINCO-2007 conference (International Conference on Informatics in Control, Automation and Robotics) which was co-organized by INSTICC (Institute for Systems and Technologies of Information, Control and Communication) and the University of Angers. The 3rd International Workshop on Multi-Agent Robotic Systems intends to be a forum for the debate of issues concerning all kinds of theoretical and applied studies combining the interdisciplinary areas of multi-agent systems and robotics, with communicating agents, flexibly materialized in either software or hardware - the latter as mobile robots. The workshop papers included in these proceedings spread across a number of topic areas including the following: autonomous and emergent formations, swarm robotics, cooperative robotics, human-robot communication and coordination, mobile robot teams, maps and navigation, robotic soccer and other types of multi-agent system aspects related to robotics, like agent-oriented design, evolution and self-recovery, behavior planning and task allocation, distributed administration and maintenance, mission reliability and simulation, artificial intelligence, targets search, sensing, convergence, and others. After a double blind review process a total of 9 papers out of 17 submitted were selected for oral presentations, arranged in 3 sessions of a single track program; 3 were selected as poster presentations. We would like to express our sincere gratitude to all the authors, who provided the rich material discussed at the workshop, and the members of the Program Committee who have reviewed and assessed the scientific merit of each submitted paper, thus ensuring high quality standards. Last but not least, thanks to Marina Carvalho for her effort in the secretariat support, and to Bruno Encarnação and Vitor Pedrosa for producing the proceedings. Joaquim Filipe Polytechnic Institute of Setúbal / INSTICC, Portugal Peter Sapaty Institute of Mathematical Machines and Systems, NAS, Ukraine iv v Workshop Chairs Peter Sapaty The University of Aizu Aizu-Wakamatsu, Japan and Joaquim Filipe Polytechnic Institute of Setúbal / INSTICC Setúbal, Portugal Program Committee Ronald Arkin, Georgia Tech, U.S.A. Nicolas Bredeche, LRI, France Alfred Bruckstein, Technion - IIT, Israel Luca Cernuzzi, Universidad Católica, Paraguay Anibal Castilho Coimbra de Matos, Faculdade de Engenharia da Universidade do Porto (FEUP), Portugal Urbano Nunes, University of Coimbra, Portugal Michel Occello, Université Pierre-Mendès-France, France Mohan Sridharan, University of Texas at Austin, U.S.A. Edward Tunstel, NASA Jet Propulsion Laboratory, U.S.A. Paul Vogt, University of Edinburgh, U.K. vi vii Table of Contents Foreword.................................................................................................... iii Workshop Chairs ...................................................................................... v Program Committee ................................................................................. v Full Papers A Hybrid Dynamic Task Allocation Approach for a Heterogeneous Multi-Robot System.............................................. 3 Yan Meng and Kashyap Shah A Bio-inspired Multi-Robot Coordination Approach......................... 14 Yan Meng, Olorundamilola Kazeem and Jing Gan To Add with Caution —Decreasing a Swarm Robotics’ Efficiency by Imprudently Enhancing the Robots’ Capabilities................... 24 Yaniv Altshuler, Israel A. Wagner and Alfred M. Bruckstein Cooperative Collision Avoidance between Multiple Robots based on Bernstein-Bézier Curves............................................................. 34 Igor Skrjanc and Gregor Klancar A Complexity Theory Approach to Evolvable Production Systems ............................................................................................... 44 Regina Frei, José Barata and Giovanna Di Marzo Serugendo Multi-Robot Task Allocation with Tightly Coordinated Tasks and Deadlines using Market-based Methods................................ 54 Robert Gaimari, Guido Zarrella and Bradley Goodman viii BESA-ME: Framework for Robotic MultiAgent System Design................................................................................................. 64 David M. Flórez, Guillermo A. Rodríguez, Juan M. Ortiz and Enrique González Robustness Against Deception in Unmanned Vehicle Decision Making................................................................................................ 74 William M. McEneaney and Rajdeep Singh Ultrasound Sensor Array for Robust Location .................................... 84 José N. Vieira, Sérgio I. Lopes, Carlos A. C. Bastos and Pedro N. Fonseca Posters An Elementary Communication Framework for Open Co-operative RoboCup Soccer Teams .......................................... 97 Luís Mota and Luís Paulo Reis Towards a Generic Anticipatory Agent Architecture for Mobile Robots................................................................................... 102 Noury Bouraqadi and Serge Stinckwich An Experiment in Distributed Visual Attention.................................. 106 P. Bachiller, P. Bustos, J. M. Cañas and R. Royo Author Index ............................................................................................. 113 FULL PAPERS A Hybrid Dynamic Task Allocation Approach for a Heterogeneous Multi-Robot System Yan Meng and Kashyap Shah Department of Electrical and Computer Engineering Stevens Institute of Technology, Hoboken, NJ 07030, USA [email protected], [email protected] Abstract. In this paper, we propose a communication-efficient hybrid task scheduling algorithm for a heterogeneous multi-robot system under dynamic unknown environment, where each robot makes its own decision through com- municating with others as well as checking a global task status queue. The pro- posed hybrid algorithm takes advantage of centralized approaches to improve the overall system efficiency and distributed approaches to reduce the commu- nication overhead, which automatically leads to a reasonable reduction of power consumption. This algorithm avoids unnecessary communication by broadcasting global information which is in everybody’s interest and mean- while limits specific information which is in interest of some specific robots only. In addition, each robot would dynamically allocate the task to robots which are capable and most available. This feature makes the system robust against communication failures and robot failures. Simulation results demon- strate the efficiency and robustness of the proposed approach. 1 Introduction With growing need of building reliable real-time applications coupled with advance- ment of high-speed networks and high-performance computers, in the past decade heterogeneous multi-robot systems have been increasingly used for many real-time applications, like urban search and rescue, surveillance, hazardous materials detec- tion, and reconnaissance, in which the correctness of the systems depend not only on the results of a computation, but also on the time which these results are produced [1]. To achieve real-time performance of such a complex system, an efficient task alloca- tion and coordination among the team members is required. Vali Veloso stated in [2] that team performance can be drastically increased if the team coordinates well and the information is being shared by all teammates in a multi-robot environment. Dynamic task allocation for multi-robot systems under dynamic environment is a challenging problem, which aims to efficiently finish all of unknown tasks as fast as possible while keep the cost as low as possible. Although some algorithms have been proposed to tackle this problem, such as auction-based algorithm like MURDOCH [3][4], behavior-based algorithm like ALLIANCE [5][6], and instantaneous greedy scheduler based approaches, all of these available methods have a great deal of broadcast communication overhead to share information with all of team members. 4 Some of available algorithms are only good for a homogeneous robot team with one global task like mapping or exploration of an area, and some of them don’t take con- sideration of system robustness in the case of communication failure or robot mal- functions. In this paper, we aim at investigating a task scheduling algorithm for a heteroge- neous multi-robot system under dynamic unknown environment. As we know, a cen- tralized approach consists of making all decisions in one place, where all the tasks to be performed are collected by a central scheduler. This centralized scheduler decom- poses tasks into programs of actions, order actions when necessary and assigns them to robots with respect to their capabilities, work loads, and locations.