一种用于四足机器人逆运动学求解的改进天牛须算法研究  

An Improved Beetle Antennae Search Algorithm of Solving Inverse Kinematics of Quadruped Robot

作　　者：邱斌 冯科 王云霄 QIU Bin;FENG Ke;WANG Yunxiao(Guangxi key Laboratory of Embedded Technology and Intelligent System,Guilin 541004,Guangxi,China;School of Computer Science and Engineering,Guilin University of Technology,Guilin 541004,Guangxi,China)

机构地区：[1]广西嵌入式技术与智能系统重点实验室,广西桂林541004 [2]桂林理工大学计算机科学与工程学院,广西桂林541004

出　　处：《机械科学与技术》2025年第4期601-608,共8页Mechanical Science and Technology for Aerospace Engineering

基　　金：广西重点研发计划(桂科AB23075163,桂科AB23026034);桂林理工大学博士科研启动基金项目(GUTQDJJ2014042);桂林理工大学大学生创新创业训练计划(202210596341)。

摘　　要：针对求解四足机器人逆运动学时传统方法运算量大、稳定性差且收敛耗时长等问题,提出一种适用于12自由度四足机器人的改进天牛须算法。该算法首先在每次迭代时约束天牛朝向转角范围,提高算法的空间搜索能力;其次,引入适应度的历史信息进行步长自适应更新,设置与目标精度关联的基本步长进行局部搜索,提升算法的后期收敛性能;最后,构造适应度函数描述机器人足端位姿误差,通过最小化适应度求解逆运动学问题。仿真结果表明,所提算法在求解12自由度四足机器人的逆运动学时具有较好的收敛性,求解精度高。In order to solve the drawbacks of large computation,poor stability and slow convergence of traditional method used in the inverse kinematics of 12⁃DOF(twelve⁃degree⁃of⁃freedom)quadruped robot,an improved beetle antennae search algorithm is proposed.Firstly,the algorithm constrains the rotation angle of beetle′s orientation for each iteration to improve the searching ability of the algorithm in multi⁃dimensional space.Secondly,the historical information of fitness is introduced for adaptive update step size,and the minimum step size associated with the target accuracy is set for fine search for enhancing the anaphase convergence performance.Finally,a fitness function is constructed to describe the pose error of the quadruped robot,and the inverse kinematics of the quadruped robot is solved by minimizing the fitness.Numerical results show that the present method has better convergence,higher accuracy when solving the inverse kinematics problem of 12⁃DOF quadruped robot.

关 键 词：四足机器人 逆运动学 适应度 天牛须算法 

分 类 号：TP27[自动化与计算机技术—检测技术与自动化装置]