基于遗传算法优化的仿人足球机器人步态规划研究

摘要

Aiming at the features of the RoboCup match and the requirement of fast walking for humanoid socer robot, a planning based on GA optimized was presented. Firstly, a walking movement cycle of the humanoid soccer robot was analyzed. During the walking cycle, the movement of the hip was continuous and the ankle was intermittent. So a cubic spline interpolation and a high order polynomial interpolation were used to plan their trajectories respectively. Secondly, the objective function was constructed using zero moment point (ZMP) stability margin as a parameter. In order to obtain a stable gait with larger ZMP stability margin, the planned trajectory was optimized based on the genetic algorithm (GA). Simulation results show that the humanoid soccer robot can walk more stably with the gait planned by the presented approach.%针对机器人足球比赛的特点及实现仿人足球机器人稳定快速步行的要求,提出了一种基于遗传算法优化的步态规划方法.首先,分析了仿人足球机器人一个完整步行周期的运动过程,根据步行中髋关节运动连续和踝关节运动间断的特点,分别采用三次样条插值与高次多项式插值进行了轨迹插值;其次,以零力矩点(ZMP)稳定裕度为参数构造了目标函数;最后,利用遗传算法(GA)对规划的步态进行了优化,从而得到了ZMP稳定裕度相对较大的稳定步态.仿真结果表明,该方法规划的步态能实现仿人足球机器人更稳定地步行.