Abstract：In this paper we propose a learning mechanism for mobile robot navigation.The robot is controlled by a dynamic set of fuzzy rules and the rule set is learned using genetic algorithm. We use messy genetic algorithm to reduce the size and complexity of chromosome and niche genetic algorithm to increase the learning speed. We also take the kinematics model of wheeled mobile robot into account and use the speed of wheels directly as the output of fuzzy rules. After the rules have been learnt in simulated environment, they are tested in the globe vision system designed by us. Experimental results prove the learning mechanism is correct and feasible.
[1] M A Salichs,L Moreno.Navigation of Mobile Robots:Open Questions.Robotica,2000,18:227-234 [2] R A Brooks.A Robust Layered Control System For a Mobile Robot.IEEE Journal of Robotics and Automation,1986,RA-2(1):14-23 [3] M H Lim,S Rahadja.A GA Paradigm for Learning Fuzzy Rules.Fuzzy Sets Syst.1996,82:177-186 [4] Frank Hoffimann.Soft Computing Techniques for the Design of Mobile Robot Behaviors.Information Science,2000,122:241-258 [5] Seung-Ik Lee,Sung-Bae Cho.Emergent Behaviors of a Fuzzy Sensory-Motor Controller evolved by Genetic Algorithm.IEEE Tran Syst Man,Cybern.B 2001,31(6):919-929 [6] 杨杰,黄欣等.基于遗传算法的模糊规则自动生成.系统仿真学报,2000,12(2):159-161 [7] T L Seng,M B Khalid,R Yusof.Tuning of a Neuro-Fuzzy Controller by Genetic Algorithm.IEEE Tran Syst Man,Cybern B,1999,29:226-236 [8] Goldberg D E,et al.Messy Genetic Algorithms:Motivations,Analysis and First Result.Complex System,1989,3:493-530 [9] 罗青,吕恬生等.足球机器人仿真系统的研究与开发.计算机仿真,1999,4
