基于BP神经网络的模型参考自适应姿态控制

Model Reference Adaptive Attitude Control Technology Based on BP Natural Network

采用模型参考自适应控制的基本设计框架,并通过BP神经网络对PID控制参数进行自主调节,实现飞行器的自适应姿态控制。利用参考模型输出、实际对象输出等信号作为训练信号,对所构建的三层BP神经网络进行权重更新。仿真结果表明,将BP神经网络应用于飞行器的自适应姿态控制中,能够实现PID控制器参数的自主调整,表明了BP神经网络优良的逼近性能。同时,该控制方案确保了飞行器姿态控制系统的性能指标,并且提高了工程设计的智能化水平。

In this paper,the basic design framework of model reference adaptive control is adopted,and the parameters of PID control are adjusted by BP neural network in order to realize the adaptive attitude control of aircraft.Using reference model output and actual object output as training signals,the weights of the three-layer BP neural network are updated.The simulation results show that the parameters of the PID controller can be adjusted autonomously by applying the BP neural network to the adaptive attitude control of the aircraft,which show the excellent approximation performance of the BP neural network.At the same time,the control scheme guarantees the performance index of the aircraft attitude control system and improves the intelligent level of engineering design.