基于鱼群算法的仿人机器人步态优化  

Research on the gait optimization of humanoid robots based on the fish swarm algorithm

作　　者：卢嘉敏[1] 杨宜民[1] 付根平[1] 

机构地区：[1]广东工业大学自动化学院,广东广州510090

出　　处：《智能系统学报》2013年第6期526-531,共6页CAAI Transactions on Intelligent Systems

摘　　要：针对现有步态优化方法中存在的缺点和不足,分析了仿人机器人腿部各关节对步态稳定性的影响,在径向和侧向平面内以机器人步态的ZMP稳定裕度和整体能耗为参数构造目标函数,利用鱼群算法对基于样条插值方法规划的仿人机器人步态进行优化.仿真实验表明,该方法不仅能够获得ZMP稳定裕度大、能耗小的仿人机器人平滑步态,而且相对采用遗传算法对仿人机器人步态优化具有更好的全局搜索能力和更快的寻优速度.This paper focuses on the shortcomings and deficiencies existing in the present gait optimization method. The influence caused by the leg joints of a humanoid robot on the gait stability is analyzed. In the radial and lateral plane, the ZMP stability margin of the gait of a robot and the integral energy consumption are taken as parameters to form a target function, the fish swarm algorithm is applied to optimize the gait of the humanoid robot programmed on the basis of the spline interpolation method. The simulation experiment shows that, by the use of this method, not only may the smooth gait of a humanoid robot with a large ZMP stability margin and small energy consumption be obtained, but also the global search capability has been found to be better and the optimization speed is faster, in com- parison to the optimization of the inheritance algorithm used for the gait optimization of a humanoid robot.

关 键 词：仿人机器人 步态优化 鱼群算法 能耗 ZMP稳定裕度 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]