自适应噪声抵消算法优化与FPGA实现

Adaptive noise cancellation algorithm optimization and FPGA implementation

针对自适应噪声抵消算法在FPGA上资源占用和运算速度的问题,提出了一种降低运算量的符号LMS算法,分析其较于传统LMS算法的优点,设计了一种基于状态机和符号LMS算法的自适应噪声抵消器,首先采用高性能音频编解码芯片WM8731对含噪音频信号和噪声信号采样,然后噪声抵消器对所得的两路信号进行自适应滤波去除噪声信号,最后用音频芯片输出去除噪声后的音频信号.实验结果表明噪声抵消器可达到快速有效的噪声抵消效果.

To deal with the problem of resource occupation and computing speed of adaptive noise cancellation algorithm on FPGA,signum-LMS algorithm that can reduce the computation complexity is proposed after analyzing its advantages over the traditional LMS algorithm,and then an adaptive noise canceller based on state machine and signum-LMS algorithm is designed.First,the audio codec chip WM8731 is used to sample the speech signal with noise and noise signal,and then the canceller based on FPGA runs the signum-LMS algorithm to reduce the noise signal,finally,the audio codec output the clean speech signal.Experimental results show that the noise canceller can reduce the noise quickly and effectively.