A Knowledge-based System for Developing Cognitive Map Based on Geographical Map A Thesis Submitted in Partial Fulfillment of the Requirements for the Master Degree in Computer Science By Ali Mohammad Rawashdeh Supervisor Dr. Hussein H. Owaied Faculty of Information Technology Middle East University Amman, Jordan Dec. 2010 I II III IV إهداء V Acknowledgment First and foremost, last but not least, thanks be to Allah for my life, you made my life bountiful, may your name be exalted, honored, and glorified. I wish to thank my supervisor ―Dr. Hussein H. Owaied‖. He has been supporting me since the days I began working on my thesis. He has supported me not only by providing a research assistantship over almost the last semester, but also academically and emotionally through the rough road to finish this thesis. Great thanks to him I had the opportunity to work with him on this field. He helped me come up with the thesis topic and guided me over almost a year of development. He gave me the moral support and freedom I needed to move on. His help, stimulating suggestions, encouragement and with his wisdom he helped me through the whole time of preparing and writing this thesis. This thesis, indeed, grew out of a series of dialogs with him It is difficult to overstate my gratitude to Faculty of Information Technology membership at Middle East University with their enthusiasm, inspiration and great effort to explain things clearly and simply. My sincere gratitude is attributed to the decent and most respectful examining committee for their efforts in advance to discuss this thesis. My deepest gratitude goes to my family for their unflagging love and supporting through out my life, this thesis is simply impossible without them. I am indebted to my mother and father for their inspiration and supports. I can‘t ask for more from my mother as she is simply perfect. I have no suitable word that can fully describe her ever lasting love to me. Last but not least, I thank all my brothers and sisters who were with me all the time, I have to thank them for their withstanding and understanding me, I would like to express my gratitude for my sister‘s husband ―Omar Samara‖ for all the nights he spent to put all his efforts to help me, and great thanks for all students at the Middle East University and to all those who encouraged me and gave me the possibility to complete this thesis. VI Table of Content Chapter One 1. Introduction 1.1. Overview………………………………………………………. 1 1.2. Geographic Information System (GIS) ……………………….. 3 1.3. Global Position System (GPS) ...…………………………….. 4 1.4. Artificial Intelligent and the Cognitive Map ………………….. 4 1.5. Problem Statement …………………………………………... 7 1.6. Motivations …………………………………………………… 8 1.7. Objectives …………………………………………………….. 9 1.8. Contributions ………………………………………………… 10 1.9. Study Boundaries ……………………………………………. 11 1.10. Thesis Structure ……………………………………………… 12 Chapter Two 2. Literature Survey and Related works 2.1. Literature Survey ………………………………….…………. 13 2.1.1. Geographic maps …………………………………………….. 13 2.1.2. Cognitive map ………………………………………………... 15 2.1.3. Intelligent Agent ……………………………………………... 17 2.1.4 Intelligent Agent mapping …………………………………... 19 2.2. Related Works ……………………………………………….. 20 Chapter Three 3. Methodology Used 3.1. Overview ……………………………………………………… 24 3.2. Geographical Map ………………………………………..…… 24 3.3. Coordinate Grids …………………………………..………….. 26 3.4. Scaling ………………………………………………………… 27 3.5. Spatial Knowledge Acquisition ……………………..………... 27 VII 3.5.1 Landmark knowledge ………………………..……………….. 28 3.5.2 Procedural Knowledge ………………………………………... 28 3.5.3 Survey Knowledge …………………………….……………… 29 3.5.4 Methodology Used for Data Acquisition ……………………... 30 Chapter Four 4. Proposed Model 4. Overview …………………………………………………….. 31 4.1. Google Mapping and GIS.…………………………………… 31 4.2. Image processing ………………….…………………............. 32 4.3. The Architecture of Proposed Model…………………………. 33 4.3.1. User Interface ………………………………………………... 34 4.3.2. Inference Engine ……………………………………………... 35 4.3.3. Data Base ……………………………………………………... 35 4.3.4. Knowledge Base ……………………………………………… 36 4.3.4.1 Landmark knowledge as Declarative Representation …............ 44 4.3.4.2 Procedural Knowledge Representation as Rules ..……............. 46 4.3.4.3 Survey Knowledge Representation as Blackboard …………... 48 4.3. Threshold Module …………………………………………… 49 4.4. Road Extracting Module ……………………………………. 50 4.5. Skelton Module ………………………………………………. 52 4.6. Localization …………………………………………………... 55 4.6.1. Start Point ……………………………………………………. 55 4.6.2. End Point ……………………………………………………... 56 4.6.3. Nearest Point …………………………………………………. 58 4.7. Path Finding Module ……………………..…………………... 59 4.8. Shortest Path Module ………………………………………... 63 4.9. Orientation …………………………………………………… 66 VIII 4.10. Developing Cognitive Map Module ………………………….. 68 Chapter Five 5. Conclusion and Future Work 5.1. Conclusion ……………………………………………………. 73 5.2. Future Work …………………………………………………. 74 IX List of Figures Fig.2.1 The Architecture of Intelligent Robot as Knowledge-Based System…. 19 Fig. 3.1 The Coordinate Grids of the word map ……...………..……………… 26 Fig. 3.2 Map scaling details application ……………...……………………….. 27 Fig. 3.3 Rout map ………………………..…………......……………………... 29 Fig. 4.1 An overview of the proposed model …..………………..……………. 34 Fig. 4.2 Metadata sample using XML…………………………………………. 36 Fig. 4.3 The hierarchical organization of spatial knowledge ..………………… 38 Fig.4.4 Fuzzy Cognitive Map produced………………………………………. 43 Fig.4.5 basic rule to construct the relations 44 Fig.4.6 Simulation of geographical map ……………………………………… 46 Fig. 4.7 Threshold pseudo code ……………………..………….……………... 50 Fig. 4.8 Map result after thresholding …………...………………..…………… 50 Fig. 4.9 Different patterns of road width and seed point ……………………… 51 Fig. 4.10 Filling Algorithm pseudo code ………...……………….…………….. 51 Fig. 4.11 Map result after Filling ………...……………………….…………….. 52 Fig. 4.12 Example of transition ………………...………………….…………… 53 Fig. 4.13 Example of connection ………………….……………….…………… 53 Fig. 4.14 Zhang-Ssuen thinning algorithm pseudo code …………...…………... 54 Fig. 4.15 Map result after skeletonization ………………….………..…………. 54 Fig. 4.16 finding start and end point centers pseudo code ……………………… 56 Fig. 4.17 Landmark symbol details …………….…………………….………… 57 Fig. 4.18 localization API segment code …………..…………………………. 57 X Fig. 4.19 Start and End point using Google Maps API…………………………. 58 Fig.4.20 Start and end point estimation using nearest point……………………. 59 Fig. 4.21 Distinct facilities of pixels …………..………………………………... 60 Fig. 4.22 Path Finding Algorithm Pseudo Code ……………….…….…………. 61 Fig. 4.23 A cross point and the eight neighbors characteristics ...……..……….. 62 Fig. 4.24 comparison process pseudo code……………………………………… 63 Fig. 4.25 First path detected ……….……………………………………..……... 64 Fig. 4.26 Second path detected ………..………………………………………... 64 Fig. 4.27 Shortest path selected ………..………………………………….……. 65 Fig.4.28 Segment code of rout generation on Google Maps API………………. 65 Fig.4.29 Common Directions…………………………………………………… 66 Fig.4.30 Direction translation in the natural language………………………….. 67 Fig.4.31 Angels Translated into Directions…………………………………….. 68 Fig. 4.32 Map contains many landmarks ……………………………………….. 70 Fig. 4.33 Cognitive map with hospitals only …………………………………… 70 Fig. 4.34 Cognitive map for land mark with shortest path ……….…………….. 71 Fig.4.35 Steps for finding landmark and path using CM……………………….. 71 Fig.4.36 Final cognitive map using Google Maps……………………………… 72 Fig. 5.1 Multi path …………...………….…………………………………….. 74 Fig. 5.2 Multi direction ……………………………………………..…………. 74 XI List of Tables Table 3.1 Map scaling details depends on application ……………..…………… 27 Table 4.1 Landmarks sample attributes ………………………………………… 35 Table 4.2 Estimated strength values between landmarks....................................... 39 Table 4.3 Highest influencin values between landmarks………………………... 39 Table 4.4 Seperation influence values…………………………………………… 40 Table 4.5 Example of influencer and infuenced landmarks……………………... 41 Table 4.6 Summary of relations between landmarks……………………………. 42 Table 4.7 Filtered influencing values……………………………………………. 42 Table 4.8 List of abbreviations names used in FCM……………………………. 43 Table 4.9 Categorizing landmarks………………………………………………. 44 Table 4.10 Example of metadata in degree reading………………………………. 46 Table 4.11 List of landmark symbols……………………………………………... 58 XII Abstract This thesis presents a knowledge-based system for developing cognitive map from geographical map. The system consists of three parts; the user interface, the inference engine, and the knowledge base. The development of cognitive map processes ware based on the geographical map and the philosophy of human vision. These processes have been described in a set of mathematical algorithms and implemented as software modules. The mathematical algorithms have constituted the two parts, inference engine and the knowledge base of the knowledge-based system. This thesis introduces many applications for the system and gives an example that produces a very good result. The present thesis concentrates on the knowledge for only one layer of the geographical map and the future work will include the knowledge for multilayer. XIII انخﻻصت حقذو هزي انشسبل وظبو معشفت نبىبء انخشٌطت اﻻدساكٍت مه انخشٌطت انجغشافٍت . انىظبو ٌحخوي عهى ثﻻثت اجضاء: وسٍهت اﻻحصبل، محشك اﻻسخذﻻل و قبعذة انمعﻻفت. طوس بىبء انخشٌطت اﻻدساكٍت بىبءً عهى انخشٌطت انجغشافٍت و فهسفت انىظش عىذ اﻻوسبن. وصفج عمهٍت بىبء انخشٌطت اﻻدساكٍت ببنخفصٍم كمجموعت مه انخواسصمٍبث انشٌبضٍت و طبقج عهى هٍئت وحذاث بشمجٍت ، ان