利用加強式進化學習之模糊控制器於移動式機器人導航

Mobile Robot Navigation Using Fuzzy Controller with Reinforcement Evolutionary Learning

楊惟傑(Wei-Chieh Yang)

指導教授 ： 陳政宏

國立虎尾科技大學/電機資訊學院/電機工程系碩士班/碩士(2016年)

摘要

移動式機器人的自動導航在機器人研究領域是一大挑戰。在本研究中，我們將自動導航分成兩個行為。一個是使用以貿易為基礎之帝國主義競爭演算法(Trade-based Imperialist Competitive Algorithm, TICA)於進化式模糊控制器(Evolving Fuzzy Controller, EFC)來完成導航，稱為目標找尋行為。另外一個是由我們設計的兩個模糊邏輯控制器(Fuzzy Logic Controller, FLC)用來避開障礙物，稱為沿牆行為。兩個模糊邏輯控制器分別是，左沿牆模糊邏輯控制器(Left Wall-following Fuzzy Logic Controller, LWFLC)與右沿牆模糊邏輯控制器(Right Wall-following Fuzzy Logic Controller, RWFLC)。在導航任務中，EFC的輸入是移動式機器人量測到的距離與目標之間的角度，輸出則是左右車輪的速度。本論文提出的TICA演算法通過增加國際貿易的概念改善原帝國主義競爭演算法(Imperialist Competitive Algorithm)的同化階段，使其提高了全域搜索的能力。其中我們藉由增強式學習方法定義了一個適應值函數，並藉由該函數來評估EFC在導航任務中的效能。最後本論文將以EFC在導航任務上的效能比較TICA和其它以族群為基礎的進化式演算法，來證明我們提出的方法可以在未知環境中完成自動導航。

關鍵字

進化式模糊控制器 ； 導航控制 ； 機器人運動控制 ； 帝國主義競爭演算法 ； 沿牆控制器 ； 增強式學習

並列摘要

One of the challenging aspects in mobile robots is the ability to automatic navigation. In this study, we used two behavior to complete navigation task. One is evolving fuzzy controller (EFC) learning by trade-based imperialist competitive algorithm (TICA) in order to achieve mobile robot navigation, namely goal-seeking behavior. In the other behavior, we designed two fuzzy logic controller for obstacle avoiding, left wall-following fuzzy logic controller (LWFLC) and right wall-following fuzzy logic controller (RWFLC), namely wall-following avoid-obstacle behavior. During the navigation task, the EFC inputs are the measured distance and angle between the mobile robot and the goal, the outputs of EFC are the speeds of left wheel and right wheel. The proposed TICA that modified the assimilation phase of imperialist competitive algorithm (ICA) by adding international trade concept that increase the global search ability. A fitness function was defined by reinforcement learning method to evaluate the EFC performance in the navigation task. To demonstrate the performance of the EFC-TICA, the method was compared with other population-based algorithms through efficiency of the navigation task.

並列關鍵字

Evolving fuzzy controller ； navigation control ； robot motion control ； imperialist competitive algorithm ； wall-following controller ； reinforcement learning

參考文獻

[1] L. E. Parker, “Current research in multi-robot systems,” Artificial Life and Robotics, vol. 7, pp. 1-5, 2003.

[2] H. Seraji and A. Howard, “Behavior-based robot navigation on challenging terrain: a fuzzy logic approach,” IEEE Trans. on Robotics and Automation, vol. 18, no. 3, pp.308-321, Jun. 2002.

[3] F. Abdessemed, K. Benmahammed, and E. Monacelli, “A fuzzy-based reactive controller for a non-holonomic mobile robot,” Robotics and Autonomous System, vol. 47, no. 1 pp. 31-46, May.2004.

[4] D. K. Pratihar, K. Deb, and A. Ghosh, “A genetic-fuzzy approach for mobile robot navigation among moving obstacles,” International Journal of Approximate Reasoning, vol.20, no.2 pp.145-172, Feb. 1999.

[6] C. F. Juang and Y. C. Chang, “Evolutionary-group-based particle-swarm-optimized fuzzy controller with application to mobile-robot navigation in unknown environments,” IEEE Trans. on Fuzzy Systems, vol. 19, no. 2 pp. 379-392, Apr. 2011.