关键词: 遗传算法, 参数辨识, 摩擦模型, 管道机器人

Abstract: For robots used for detecting the width of pipeline notch, the friction between robot and pipeline would directly affect the accuracy and efficiency of measurement. The solution is to identify the friction parameters using improved genetic algorithm by studying the deceleration characteristics of the pipeline robot. Simulation and experimental results are as follows: firstly, parameters of Stribeck friction model are identified, of which the error rate is less than 5%, indicating the problem of local optimization could be successfully solved. Then, the adaptive genetic algorithm is applied to fit the experimental data. Through the research, it is concluded that the improved genetic algorithm has a high accuracy and is fast for parameter identification. The established friction mathematical model can be used to calculate real-time friction, which is of great significance for advancing the performance of pipeline robot.

Key words: genetic algorithm, parameter identification, friction model, pipeline robot