基于改进灰狼算法的机器人激励轨迹优化

Optimization of Robot Excitation Trajectory Based on Improved Gray Wolf Optimizer

针对协作型机器人惯性参数辨识中激励轨迹设计问题,文章提出了一种采用改进灰狼算法(SGWO)用于优化设计激励轨迹参数方法。首先,用Newton-Euler递推法推导并建立了机器人最小惯性参数观测矩阵,并将观测矩阵条件数准则作为优化目标函数;其次,引入改进灰狼算法(SGWO).通过反向最优最差策略改善种群初始值,采用sigmoid函数优化收敛因子;最后,利用改进灰狼算法(SGWO)优化设计了满足多约束条件的周期傅里叶级数作为激励轨迹。实验结果表明,采用所提优化方法设计的激励轨迹可以充分激发机器人动力学特性,提高参数辨识的抗噪能力,为准确获取机器人动力学参数奠定基础。

To deal with the problem of excitation trajectory design in the identification of inertial parameters of collaborative robots.An improved Grey Wolf optimizer(SGWO)is proposed to optimize and design the excitation trajectory parameters.First,the minimum observation matrix of the robot dynamics model is derived by the Newton-Euler method,and the minimum condition number of the inertia matrix is used as the optimization objective function.Secondly,the improved gray wolf optimizer(SGWO)is introduced.The initial value of the population is optimized by the inverse optimal worst-case strategy,and the sigmoid function is used to optimize the convergence factor.Finally,the improved grey Wolf optimizer(SGWO)is used to optimize the periodic Fourier series satisfying multiple constraints as the excitation trajectory.The experimental results show that the excitation trajectory designed by the proposed optimization method can fully stimulate the dynamic characteristics of the robot,improve the anti-noise ability of parameter identification,and lay the foundation for accurately obtaining the dynamic parameters of the robot.