基于神经网络参数优化的迭代学习控制算法

An Iterative Learning Control Algorithm of Parameters Optimization Based on Neural Network

针对迭代学习控制用于轨迹跟踪时存在收敛速度慢的问题,提出用RBF网络优化迭代控制器参数的算法。在每一次迭代学习过程之后利用RBF网络对当次输出的数据进行优化计算,拟合出最优的学习增益,使迭代学习算法具有较快的收敛速度,在单关节机器人中进行仿真验证了方法的有效性。

An iterative learning control algorithm of parameters optimization based on the RBF network is proposed for the slow convergence speed of an iterative learning controller in the trajectory tracking.After each iterative learning process,the RBF network is used to optimize the current output data, fit the optimal learning gains,then make the iterative learning algorithm have the quicker convergence rate.The simulation result of single-joint robot shows the efficiency of the method.