关键词: 机械臂, 动力学模型, Stribeck摩擦模型, 参数辨识, 混合迭代法

Abstract: Accurate identification of robot dynamics parameters is the key to the research of impedance control and trajectory tracking control,etc. Since the traditional identification methods cannot guarantee the physical feasibility of the dynamics parameters,resulting in the identification parameters cannot be used,a hybrid iterative method for the identification of robot dynamics parameters is proposed. The method is divided into two loops,the physical feasibility of the identified parameters is guaranteed by introducing linear matrix inequality and semi-positive definite programming in the inner loop,and then the inertia parameters are identified,and the friction parameters are identified by improving the friction torque estimation with the Stribeck friction model in the outer loop,which is computed in a cyclic manner until the convergence condition is satisfied. Finally,experiments are carried out on a six-degree-of-freedom tandem robotic arm,and the root-mean-square error value of the moments of the 6 joints is 17.24 % lower in the hybrid iterative method compared with the least-squares method,which practically improves the accuracy of the robot dynamics parameterization,and proves the validity of the hybrid iterative method.

Key words: robotic arm, dynamic model, Stribeck friction model, parameter identification, hybrid iterative method