基于参量阵系统的神经网络PID逆控制及仿真研究  

Neural network PID inverse control and simulation research based on parametric array system

作　　者：吴新龙[1] 陈敏[1] 赵亮[1] 

机构地区：[1]电子科技大学航空航天学院,成都611731

出　　处：《声学技术》2013年第6期511-514,共4页Technical Acoustics

基　　金：青年科学基金项目(60901062)

摘　　要：针对参量阵系统这样一个强非线性声学系统的声音输出扭曲问题,提出了神经网络自适应逆控制策略。首先介绍了参量阵系统的逆模型,并由参量系统模型和它的逆模型构成一个伪线性系统。控制系统中的BP网络自适应调节PID(Proportional Integral Derivative)的三个控制参数,并利用PID控制参量阵系统。通过对MATLAB建立的复合逆控制模型进行仿真研究。仿真结果表明:神经网络PID控制具有较高的控制精度和适应性,可以获得良好的控制效果。A neural network adaptive inverse control strategy was proposed to control the distortion of parametric array system which is a strong nonlinear acoustic system. The inverse model of parametric array system was introduced. A pseudo-line system was generated from the parametric array model and it＇s inverse model. Three parameters of PID were adaptively controled by the BP network and the system was controled by PID. Simulation research was carried out by establishing the combined inverse control model with MATLAB. Simulation results showed that the neural network PID control had good accuracy and adaptability such that good control effect could be achieved.

关 键 词：参量阵 神经网络 系统辨识 逆控制 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]