基于Hopf振荡器的六足机器人步态CPG模型设计

CPG model design based on hopf oscillator for hexapod robots gait

利用中枢模式发生器实现六足机器人爬行步态是运动仿生的关键。建立机器人坐标系,基于D-H参数求解正运动学;采用Hopf振荡器设计多腿耦合模型;构建由6个CPG单元组成的环形CPG网络拓扑结构,每个CPG单元由2个耦合的Hopf振荡器组成,分别输出髋关节、踝关节运动信号;采用膝踝映射函数方法,将踝关节输出信号映射为踝关节和膝关节角度轨迹,从而降低网络中振荡器个数;通过改变耦合系数保证相邻振荡器的相位互锁,输出稳定平滑信号;搭建实物样机进行步态测试。仿真和实验表明,CPG网络相位差稳定,可实现六足机器人三角步态下的平稳行走,爬行速度约为6.45 cm/s。

The key to bionic motion is a central pattern pod. Firstly, the coordinate system of the robot was set generator （CPG） , which realizes the crawl gait of a hexa- up and the associated forward kinematics were solved based on D-H parameters. Hopf oscillators were then adopted into the design of coupling models involving multiple legs. A CPG ring topology structure was established using six CPG units, with each CPG unit consisting of two coupled Hopf oscillators, which output the hip and ankle joint signals, respectively. In order to control each joint （ of a hex- apod robot）, a knee-ankle mapping function was used. The function mapped the output of the ankle to joint angles for both the knees and ankles. The number of oscillators in the CPG network was reduced using this method. Mean- while, the coupling coefficient was changed to guarantee the phase interlock of adjacent oscillators and give a stable and smooth signal. Finally, a physical prototype was constructed for testing the robotic gait. The simulations and test results show that this CPG network has a stable phase difference, which ensures that hexapod robot can walk stably in a triangular gait and a crawling speed of approximately 6.45 cm/sec can be achieved.