基于粒子群算法的机器人动力学参数辨识  

Nonlinear Dynamic Identification of Robotic Manipulators Based on Particle Swarm Optimization Method

作　　者：傅霞君 袁建军[1] 王帅[1] 王宁[1] 张伟军[1] 

机构地区：[1]上海交通大学机械与动力工程学院,上海200240

出　　处：《机电一体化》2017年第2期3-8,64,共7页Mechatronics

摘　　要：提出了一种对机器人机械臂各关节的动力学参数进行离线识别的新方法。该方法在实现途径上简单有效,无需附加传感器。通过控制机械臂各待辨识关节按照周期性傅里叶级数轨迹运动,读取关节电流以获取关节力矩值便可实现对机械臂连杆动力学参数进行识别。该识别算法通过将待识别参数分为线性部分和非线性部分,并对其分别采用线性最小二乘法和粒子群算法进行解算。实现过程简单,不仅能识别机械臂的动力学参数,还能识别动摩擦系数和黏滞摩擦系数。通过在新松七自由度协作机器人上进行实验,验证了该方法的可行性和有效性。A new identification method to identify the parameters of robotic manipulators is proposed. It uses the torque exerted by the control motors of the robot joints. The settings are simple and doesn＇t use additional sensors like force sensor or intertia measurement unit. In the process of identification, the torque exerted by each joint during per- forming Fourier periodic excited trajectories is used. It divides the parameters into a linear part and a non-linear part, and uses linear least square （LLS） parameter estimation and particle swarm optimization （PSO） method to compute them respectively. It can identify both the dynamic parameters and the viscous friction coefficients as well as the cou- lomb friction coefficients. A SIASUN 7-axis flexible robot is used to experimentally validate our proposal. The comparison of the predicted torque values and true values of the joints confirms the effectiveness of the method.

关 键 词：参数辨识 线性最小二乘法 粒子群算法 机器人 机械臂 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]