Parameter Identification of Static Friction Based on An Optimal Exciting Trajectory

[摘要] In this paper, we focus on how to improve the identification efficiency of friction parameters in a robot joint. First, the static friction model that has only linear dependencies with respect to their parameters is adopted so that the servomotor dynamics can be linearized. In this case, the traditional exciting trajectory based on Fourier series is modified by replacing the constant term with quintic polynomial to ensure the boundary continuity of speed and acceleration. Then, the Fourier-related parameters are optimized by genetic algorithm(GA) in which the condition number of regression matrix is set as the fitness function. At last, compared with the constant-velocity tracking experiment, the friction parameters from the exciting trajectory experiment has the similar result with the advantage of time reduction.

[发布机构] State Key Lab of Digital Manufacturing Equipment and Technology, Huazhong University of Science and Technology, Wuhan, Hubei; 430074, China^1

[效力级别] 无线电电子学

[关键词] Condition numbers;Constant velocities;Exciting trajectories;Fitness functions;Friction parameters;Quintic polynomials;Regression matrices;Static friction models