Simultaneous Optimization of Robot Structure and Control System Using Evolutionary Algorithm
The simultaneous optimization of a robot structure and control system to realize effective mobility in an outdoor environment is investigated. Recently, various wheeled mechanisms with passive and/or active linkages for outdoor environments have been developed and evaluated. We developed a mobile robot having six active wheels and passive linkage mechanisms, and experimentally verified its maneuverability in an indoor environment. However, there are various obstacles in outdoor environment and the travel ability of a robot thus depends on its mechanical structure and control system.We proposed a method of simultaneously optimizing mobile robot structure and control system using an evolutionary algorithm. Here, a gene expresses the parameters of the structure and control system. A simulated mobile robot and controller are based on these parameters and the behavior of the mobile robot is evaluated for three typical obstacles. From the evaluation results, new genes are created and evaluated repeatedly. The evaluation items are travel distance, travel time, energy consumption, control accuracy, and attitude of the robot.Effective outdoor travel is achieved around the 80th generation, after which, other parameters are optimized until the 300th generation. The optimized gene is able to pass through the three obstacles with low energy consumption, accurate control, and stable attitude.
simultaneous optimization evolutionary algorithm mobile robot rough terrain
Masanori Sato Kazuo Ishii
Kyushu University, Fukuoka, 819-0385,Japan Kyushu Institute of Technology, Fukuoka, 808-0196,Japan
国际会议
The 3rd International Conference of Bionic Engineering(第三届国际仿生工程会议ICBE10)
珠海
英文
185-190
2010-09-14（万方平台首次上网日期，不代表论文的发表时间）