基于虚拟样机的双足机器人多目标步态规划研究  

Research on Multi-objective Gait Planning of Biped Robot Based on Virtual Prototype

作　　者：张彦凯 王雪松 金雨斌 张东升[2] Zhang Yankai;Wang Xuesong;Jin Yubin;Zhang Dongsheng(Lanzhou Longneng Electric Power Science&Technology Ltd.,Lanzhou 730000,China;Xi'an Jiaotong University,Xi'an 710000,China)

机构地区：[1]兰州陇能电力科技有限公司,甘肃兰州730000 [2]西安交通大学,陕西西安710000

出　　处：《系统仿真学报》2024年第12期2984-2992,共9页Journal of System Simulation

摘　　要：针对双足机器人步态规划中难以兼顾个性化和性能的问题,提出了一种基于虚拟样机的多目标步态优化方法。根据中国人的身体结构设计了平面欠驱动双足机器人缩比样机,采用辨识方法获得机器人精确惯性参数;建立了机器人虚拟样机,设计了包括速度、能耗、稳定性三项机器人行走的优化目标,采用改进NSGA-II算法解出了机器人多目标步态规划问题Pareto最优解集;通过仿真实验获得了不同速度下的符合人体行走规律且具有多样性的步态。A multi-objective gait optimization method based on virtual prototype is proposed to address the difficulty of balancing personalisation and performance in gait planning for bipedal robots.A scale prototype of a planar underactuated biped robot is created according to the body structure of Chinese people,and an identification approach is used to determine the robot's exact inertial parameters.A virtual prototype of the robot is created,and three optimization goals—speed,energy use,and stability are developed.Using the enhanced NSGA-II algorithm,the Pareto optimal solution set for the robot multi-objective gait planning issue is produced.Numerous gaits that conform to the law of human walking and have diversity at different speeds are obtained through the simulation experiment.

关 键 词：双足机器人 步态规划 多目标优化 虚拟样机 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]