Research on Voice Signal Acquisition and Recovery Algorithm Based on Compressive Sensing

This paper introduces a method based on compressive sensing to acquire voice signals from distributed wireless sensors. The method uses compressive sensing technology to sample, transmit and recover data to reduce the sampling rate and further compress the data. As the voice signals in the discrete cosine transform domains are approximately sparse, this paper builds a JSM-1 based model and uses the random Gaussian measurement matrix to measure the signal and recovers the signal using SOMP algorithm. The method can be applied to Internet of things in agriculture to monitor wireless sensor signals, which can significantly lower hardware costs and reduce power consumption. What's more, the method can prevent the network structure damage caused by unbalanced power consumption and can be widely used.