基于自适应差分进化算法的腿履四足机器人稳定性优化  

Stability optimization of legged quadruped robot based on adaptive differential evolution algorithm

作　　者：袁泽浩 张铂轩 刘洪昆 王扬威 YUAN Zehao;ZHANG Boxuan;LIU Hongkun;WANG Yangwei(College of Mechanical and Electrical Engineering,Northeast Forestry University,Haerbin 150040,Heilongjiang,China)

机构地区：[1]东北林业大学机电工程学院,黑龙江哈尔滨150040

出　　处：《农业装备与车辆工程》2024年第6期120-123,共4页Agricultural Equipment & Vehicle Engineering

摘　　要：针对腿履复合式四足机器人处于Trot步态时,由于大腿复合结构部分摆动不稳定所引起的重心偏移问题,采用基于自适应差分进化算法优化。利用Hopf模型中的振荡器构建中央模式发生器网络拓扑结构;借助自适应差分进化算法,通过多次迭代可迅速找到输出模型的最佳参数组合,以减少重心偏移量和机体姿态角,解决了重心偏移导致的稳定性问题。通过ADAMS和MATLAB进行仿真实验验证提出的优化方法,结果表明:该方法能够高效地寻找到全局最佳参数,使腿履复合四足机器人能够展现出优秀的运动性能。In response to the problem of center of gravity deviation caused by unstable swinging of the thigh composite structure when the leg and foot composite quadruped robot is in trot gait, an optimization method based on adaptive differential evolution algorithm is adopted to address this issue. Constructing a Central Pattern Generators network (CPG) topology using oscillators in the Hopf model;with the help of adaptive differential evolution algorithm, the optimal CPG parameter combination for the output model was quickly found through multiple iterations to reduce center of gravity offset and body attitude angle , solving the stability problem caused by center of gravity offset. The proposed optimization method was validated through simulation experiments using ADAMS and MATLAB, and the results showed that the method could efficiently find the global optimal parameters, enabling the leg and foot composite quadruped robot to exhibit excellent motion performance.

关 键 词：腿履复合四足机器人 Hopf模型 节律运动 差分进化算法 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]