이족 보행 로봇을 위한 계층형 중추패턴발생기 개발

Abstract

Central Pattern Generator (CPG) developed from nervous system is used for a trajectory generation method of periodic motion of biped robot. When the CPG is applied to the robot, the gains of oscillator have to be changed whenever moving conditions are changed such as speed. To solve this problem, hierarchical CPG model is proposed in this paper. It consists of Slave Oscillator (SO), Master Oscillator (MO), and User’s Command (UC). SOs generate each joint trajectory and MO controls frequency and amplitude of SOs. UC can control MO to start, end, and change of all motion like velocity. The generated joint trajectory by SO is changed by effect of high command like UC or MO. Therefore proposed CPG can be easily use to a robot because effect of 게인 s is decreased to generate joint trajectory, even if walking conditions of the robot are changed. To prove that, a planar biped robot was simulated to walk by using proposed CPG model.