噪声有源控制的简化模糊神经网络方法

Active noise control using a simplified fuzzy neural network

研究了非线性噪声有源控制的问题,给出了一种简化的模糊神经网络控制方法。控制系统只有一个输入,需要比较少的参数,因此降低了计算负荷。系统的收敛速度明显快于全局逼近神经网络。

In this paper,the nonlinear active noise control(ANC) is discussed.An adaptive nonlinear noise control approach,using a simplified fuzzy neural network(SFNN) is derived.Only one input is required in the control system,fewer parameters are utilized and the computing load is reduced greatly by using the time delay output.An on-line learning algorithm based on the error gradient descent method is proposed.Since the SFNN is a local approximated neural network,the convergence rate of the SFNN is faster than that of the global approximated neural network.