在虛擬世界中有各式各樣的智慧型代理人(Intelligent Behavior Agent)執行著不同的工作(Tasks)。而在傳統的系統之下代理人都單獨從事自己的工作，然而當代理人之間的工作性質類似或是具有相依性的時候，這樣的作法就顯得沒有效率。若能夠建立起一套讓代理人互相能夠合作幫助的機制，則勢必能夠增進其工作效率；而合作制度的建立也就是代表著代理人不再獨自行動，而是會產生所謂群體(Group)或是隊伍(Team)的行動模式。
而在一個群體(隊伍)之中，每個代理人都有自己不同的角色定位，本論文便是以群體中身為決策者之代理人的行為模型建構來做探討；我們以“移動至特定目的地”為目標來做討論，著眼在建構出決策者的行為模型，此模型能夠評估環境的情勢並讓決策者做出對目標的規劃，在未來希望也能夠使用此模型對使用者做出分析以及評估，使用其分析結果來套用Bayesian Networks的計算來使決策者代理人能夠有近似於原使用者的行為模式。

With the goal as “moving to specified location”, we focus on building the behavior model of Decision Maker in a cooperative team. The Decision Maker in a cooperative team should evaluate the whole environment properly, and design policies for the team to accomplish the goal. 

We propose a hybrid method of Potential Field Methods (PFM) and Virtual Force Field Methods (VFF) here for environment evaluation and path planning. With the adjustment in parameters of evaluation, we can personalize the Decision Maker and make it close to its original user. In the future we shall analyze the behavior models of original users by their evaluation data based on this method, and try to apply this method to Bayesian Network mechanism for more advanced characteristics personalizing.

第一章 緒論	6
1.1  研究動機	6
1.2  研究目的	7
1.3  論文架構	7
第二章 相關背景知識與技術	9
2.1 互動式的虛擬環境	9
2.2 智慧型代理人	9
2.3 Bayesian Network	11
2.4 Potential Field	13
2.4.1 MAPS (Multi-Agent Planning System)	15
2.5 Virtual Force Field	17
2.5.1 CARMEL機器人實驗	18
第三章 系統架構	20
3.1問題描述	21
3.2 對環境中的物件做分類	21
3.3 目標的分解	23
3.4 決策者的行為流程	26
3.5 以PFM作為評估環境的機制	28
3.5.1 以PFM來評估環境	28
3.5.2 以VFF來評估環境	33
3.5.3 PFM與VFF的混合方式	41
第四章 實作說明與結果討論	47
4.1 系統之實作說明		47
4.1.1 模擬實作之目標	47
4.1.2 系統實作環境		47
4.1.3 實作界面說明		47
4.2 模擬結果與討論		48
第五章 結論及未來相關研究	55
相關文獻	57



圖目錄
圖2-1 IBA的架構以及與虛擬環境的互動關係............10
圖2-2 Bayesian Network範例圖 ......................12
圖2-3以右方球門產生吸引力所產生的位能場............16
圖2-4以週遭環境來判斷出射門的較佳路徑..............17
圖2-5機器人以VFF方法來決定行進路線................18
圖2-6 CARMEL以VFF做路徑規劃.......................19
圖3-1依照環境的評估來規劃行進路線..................24
圖3-2 根據規劃的路徑切割成1n段路徑當作子目標 ......25
圖3-3 決策者的行為流程圖 ...........................26
圖3-4 不同行為模式代理人所規劃出的不同路徑 .........31
圖3-5 以PFM在單純環境下做路徑規劃的效果............34
圖3-6 在某些環境下PFM的顛簸狀況....................35
圖3-7 嚴重的顛簸現象................................36
圖3-8 以PFM以及VFF規劃之路徑(一) .................38
圖3-9 以PFM以及VFF規劃之路徑(二)..................39
圖3-10 當Frep與Fatt在一直線上且方向相反的情形.........40
圖3-11 因合力被抵銷導致無法走到目標所在地的情形.....41
圖3-12 結合PFM以及VFF的演算流程 ...................43
圖3-13 PFM於狹窄路徑的位能問題 ......................44
圖3-14 地區P以及P周圍的四個障礙物 .................45
圖3-15 狹窄路徑以兩種不同方式來計算位能 .............46
圖4-1 實作程式介面 ..................................48
圖4-2 兩個敵人以及一個隊友的路徑規劃 ................49
圖4-3 PFM方法在角落產生的顛簸情形...................50
圖4-4 VFF作用力的抵消現象...........................51
圖4-5 三者在障礙物排列複雜情況之下的表現 ............52
圖4-6 對於狹窄路徑的表現 ............................53
圖4-7 目標物在封閉環境中的表現.......................54


表目錄
表3-1目標為「抵達某目的地」的物件分類表.................22
表3-2目標為「除去競爭對手」的物件分類表.................22
表3-3不同狀態的隊友產生不同影響力列表...................32
表3-4不同威脅性的敵人產生不同影響力列表.................32
表3-5 PFM規劃路徑的演算流程 .............................33
表3-6 VFF規劃路徑的演算流程 .............................37
表4-1 系統實作軟硬體列表.................................47

1.	[Andrews,83] J.R. Andrews, N. Hogan, “Impedance Control as a Framework for Implementing Obstacle Avoidance in a Manipulator” Control of Manufacturing Processes and Robotic Systems, Eds. Hardt, D. E. and Book, W., ASME, Boston, 1983, pp.243-151
2.	[Banerji,80] Banerji, Ranan B., Artificial intelligence : a theoretical approach, New York : North-Holland, 1980.
3.	[Biliang,03] Biliang Zhong, Qi Zhang, Yimin Yang, “Real Time Reactive Strategies Based on Potential Fields for Robot Soccer” International Conference on Robotics, Intelligent Systems and Signal Processing, October, 2003, pp.1052-1057
4.	[Borenstein,89] J. Borenstein, Y. Koren, “Real-Time Obstacle Avoidance for Fast Mobile Robots” IEEE Transactions on Systems, Man, and Cybernetics, vol.19, no.5, September/October, 1989, pp.1179-1187
5.	[Borenstein,90] J. Borenstein, Y.Koren, “Real-time Obstacle Avoidance for Fast Mobile Robots in Cluttered Environments” IEEE International Conference Robotics and Automation, 1990, pp.572-577
6.	[Chao,04] Hsiao-Chuan Chao, “Constructing an Intelligent Behavior Avatar in a Virtual World : A Self-Learning Model based on Reinforcement” Thesis report, University of Tamkang, Danshui, Taiwan, 2004.
7.	[Frank,03] Frank E. Schneider, Dennis Wildermuth, “A Potential Field Based Approach to Multi Robot Formation Navigation” IEEE International Conference on Robotics, Intelligent Systems and Signal Processing, October, 2003, pp.680-684
8.	[Hung,03] Haw-Yun Hung, “Building an intelligent Behavior Avatar in a Virtual World” Thesis report, University of Tamkang, Danshui, Taiwan, 2003.
9.	[Jensen,01] Finn V. Jensen, Bayesian Networks and Decision Graphs, Springer-Verlag, New York, 2001.
10.	[Khatib,86] O. Khatib, “Real-Time Obstacle Avoidance for Manipulators and Mobile Robots” International Journal of Robotics Research, vol.5, pp. 90-98, 1986
11.	[Koren,91] Y. Koren, J. Borenstein, “Potential Field Methods and Their Inherent Limitations for Mobile Robot Navigation” IEEE International Conference on Robotics and Automation, April, 1991, pp.1398-1404
12.	[Bayazit,03] Osmon Burchan Bayazit [Online], Washington University, Saint Louis, Available: http://www.cse.wustl.edu/~bayazit/courses/cse558/index.php
13.	[Park,02] M.G. Park, M.C. Lee, “Experimental Evaluation of Robot Path Planning by Artificial Potential Field Approach with Simulated Annealing” Society of Instrument and Control Engineers of Japan, August 5-7, 2002, pp.2190-2195
14.	[Poty,04] A. Poty, P. Melchjor, A. Oustaloup, “Dynamic Path Planning for Mobile Robots using Fractional Potential Field” First International Symposium on Control, Communications and Signal Processing,, 2004, pp.557-561
15.	[Prahlad,00] Prahlad Vadakkepat, Kay Chen Tan, Wang Ming-Liang, “Evolutionary Artificial Potential Fields and Their Application in Real Time Robot Path Planning” Congress of Evolutionary Computation, San Diego, California, July, 2000, pp.256-263
16.	[Prahlad,01] Prahlad Vadakkepat, Tong Heng Lee, Liu Xin, “Application of Evolutionary Artificial Potential Fields” Joint 9th IFSA World Congress and 20th NAFIPS International Conference, Vancouver, Canada, July 25-28, 2001, pp.2781-2785
17.	[Sng,02] Sng H.L., G. Sen Gupta, C.H. Messom “Strategy for Collaboration in Robot Soccer” Proceedings of the 2002 The First IEEE International Workshop on Electronic Design, Test and Applications, January, 2002, pp.347-354
18.	[Takashi,04] Takashi KATOH, Kensaku HOSHI, Norio SHIRATORI “On Agents’ Cooperative Behavior Based on Potential Field” International Conference on Advanced Information Networking and Application, vol.2, no.2, 2004, pp.149-154
19.	[Tews,99] A. Tews, G. Wyeth “Multi-Robot Coordination in the Robot Soccer Environment” Australian Conference on Robotics and Automation, March 30 - April 1, Brisbane, pp. 90-95