基于RBF神经网络的破障武器内模PID控制

IMC-PID Control of Obstacle-destroying Weapons Based on RBF Neural Network

针对破障武器伺服系统中存在的非线性、参数时变和不确定性等因素,提出了一种基于RBF神经网络的内模PID控制策略。通过RBF神经网络对内模PID控制器的唯一参数进行调节,同时对RBF神经网络采用LM算法进行训练。仿真结果表明,控制方法具有良好的系统动态品质、鲁棒性和抗干扰能力,能够有效地加快破障武器的调炮速度以及提高破障武器的破障精度。

In view of such factors as nonlinearity,parameter time-varying and uncertainty existing in the servo system of obstacle-destroying weapons,an IMC-PID control strategy based on RBF neural network was proposed.According to this method,the sole parameter of the IMC-PID controller was adjusted through RBF neural network,and the RBF neural network was trained with LM algorithm.Simulation results showed that the proposed control method had good system dynamic quality,robustness and anti-interference ability,and could effectively accelerate gun rotation of obstacle-destroying weapons and improve their accuracy in obstacle destroying.