摘要
对于检测管道槽口宽度的机器人,其摩擦性能直接影响到测量精度和运行效率。通过研究管道机器人在减速过程中的运动特性,利用改进遗传算法对其摩擦参数进行辨识。首先对Stribeck摩擦模型参数进行辨识,误差小于5%,有效避免了局部最优解的问题;然后运用自适应遗传算法跟踪实验所测摩擦力曲线,拟合情况良好。仿真和实验结果表明:改进遗传算法的参数辨识精度高、辨识速度快,所建立的摩擦数学模型可用于计算实时摩擦力,对管道机器人性能的提升具有实际意义。
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.
作者
刘鹏
赵言正
闫维新
LIU Peng;ZHAO Yanzheng;YAN Weixin(Institute of Robotics,Shanghai Jiao Tong University,Shanghai 200240,China)
出处
《中国民航大学学报》
CAS
2018年第6期54-58,共5页
Journal of Civil Aviation University of China
基金
国家自然科学基金项目(61273342
51475305)
关键词
遗传算法
参数辨识
摩擦模型
管道机器人
genetic algorithm
parameter identification
friction model
pipeline robot
