改进型BP神经网络MRAS控制系统

在常规BP神经网络模型参考自适应控制器基础上采用改进型BP神经网络作为辨识器和控制器,组成新的模型参考神经网络自适应控制系统,利用改进型BP神经网络的优点弥补传统自适应方法的不足,使系统具有更强的鲁棒性,收敛更快,逼近精度更高的优点。仿真结果表明,该系统比传统BP神经网络模型参考自适应系统具有更好的稳定性和更快的响应速度。

Improved BP neural network MRAS control system

In conventional BP neural network model reference adaptive controller based on BP neural network as a modified identification device and controller, the formation of a new model reference adaptive neural network control system, using the improved BP neural network advantages make up for the deficiencies of traditional adaptive method, make the system more robust, faster convergence, the advantages of approximation precision higher. The simulation results show that the system than traditional BP neural network model reference adaptive system has better stability and faster response speed.