摘要
多关节高柔性机械臂中谐波减速器作为一种复杂的传动装置,其运动学和动力学特性具有高度的非线性,会导致机械臂在控制过程中产生不可预测的误差和振动,从而影响其控制的稳定性。为此,提出基于柔性稳定函数的多关节高柔性机械臂控制研究。采用GA-RBF神经网络分析多关节高柔性机械臂控制率失稳的原因,考虑到关节柔性滚弯运动中存在的不确定性因素,设计了稳定函数,并引入速度特征量、跟踪误差等对控制率有影响的不确定性关节参数;采用GA-RBF神经网络训练采集的参数,得到可以描述关节轨迹跟踪误差和机械臂运动速度的柔性稳定函数;将速度特征量、位移特征量、方向特征量和加速度特征量这些参数代入稳定函数中得到稳定的控制率函数,基于反馈机制比较理想位置轨迹与实际位置轨迹之间的差异,实现多关节高柔性机械臂的有效控制。试验数据证明:所提方法在测试点2和测试点9的关节角速度分别在0 rad/s和-0.4 rad/s上下波动,关节角位置和输出力矩均在0 rad,0 N·m上下波动,可有效提高不确定性影响下机械臂的稳定性和控制精度。
The harmonic reducer in the multi-joint high-flexibility manipulator,as a complex transmission device,has highly non-linear kinematic and dynamic characteristics,which causes unpredictable errors and vibrations in the control process of the manipulator,thereby affecting its control stability.Therefore,in this article efforts are made to explore the control of the multi-joint high-flexibility manipulator based on the flexible stability function.The GA-RBF neural network is used to identify why the control rate of the multi-joint high-flexibility manipulator is instable.With the uncertainties in joint flexible rolling and bending motions taken into consideration,a stability function is designed;such uncertain joint parameters as speed feature quantity and tracking error that affect the control rate are introduced.The GA-RBF neural network is used to train the collected parameters;a flexible stability function that describes the joint’s trajectory-tracking errors and the manipulator’s speed is worked out.By substituting the parameters of speed feature quantity,displacement feature quantity,direction feature quantity,and acceleration feature quantity into the stability function,a stable control rate function is obtained.Based on the feedback mechanism,the difference between the ideal position trajectory and the actual position trajectory is compared;as a result,the multi-joint high-flexibility manipulator is controlled effectively.Experimental data shows that the joint angular speed fluctuates around 0 rad∕s and-0.4 rad∕s at the test point 2 and the test point 9,respectively.The joint angular position and the output torque also fluctuate around 0 rad and 0 N·m.This study ensures higher stability and control accuracy of the manipulator with uncertainties.
作者
王保勇
胡永
WANG Baoyong;HU Yong(Henan Light Industry Vocational College,Zhengzhou 450052;School of Information Engineering,Xizang Minzu University,Xianyang 712082)
出处
《机械设计》
北大核心
2025年第8期174-182,共9页
Journal of Machine Design
基金
国家自然科学基金项目(U1708258)。
