Automatic locomotion pattern generation for modular robots

Locomotion is considered as most basic function of robots. In the case of ordinary robots, they are not needed to change locomotion pattern because their configurations are constant. For self-reconfigurable modular robots, since they can change their configurations, locomotion patterns must be prepared in advance and changed by each configuration. There are two types of locomotion used for modular robots. One is locomotion by self-reconfiguration which is realized by using its reconfiguration capability. The other is locomotion using many degrees of freedoms of the configuration, e.g. walking, crawling and rolling, where the connection relationship between modules is constant. In this paper, we focus on the latter type of locomotion. Actually to design locomotion pattern suited for each configuration analytically or manually by human is difficult because it includes many DOFs. To solve this problem, we propose an automatic locomotion pattern generation method using neural oscillator a...

• Type: Conference
• Year: 2003
• Where: ICRA
• Authors: Akiya Kamimura, Haruhisa Kurokawa, Eiichi Yoshida, Kohji Tomita, Satoshi Murata, Shigeru Kokaji
• Keywords: ICRA 2003, Locomotion Pattern, Modular Robots, Robotics, Self-reconfigurable Modular Robots