基于模糊归一化神经网络的开环子带主动消噪

将无延时开环子带自适应滤波结构应用到传统的宽带主动消噪系统中，并提出开环子带结构中的X滤波预补偿方法，弥补了滤波器延迟对系统的影响，提高了系统的收敛速度，降低了计算复杂度。此外，利用神经网络来精细模拟实际噪声的高度非线性性，提出模糊归一化收敛步长调整方法以控制开环方案带来的额外误差能量，并给出了数学证明。对实际噪声的仿真结果显示，给出的主动消噪系统和算法具有更快的收敛速度和更低的稳态误差，达到了更好的降噪效果。

Open-loop Subband Active Noise Control Using Fussy Normalized Neural Network

Active noise control improves the control capability of the low frequency noise,and is widely used in aspects such as headphone design and the construction of low noise environment. A delayless open-loop subband adaptive filter structure into conventional wideband ANC system is introduced. A filtered-X subband pre-compensation approach is presented,which controls the delay,increases the convergence rate and reduces the computational complexity. In addition,the neural network is used to model the non -linear...