Research On Baseband Signal Compression Recovery Algorithm Based On Bayesian Inference

As an important branch of Bayesian theory,Bayesian Inference is widely used in various signal recovery algorithms.In this paper,we mainly study the application of Bayesian inference in baseband signal compression recovery and baseband signal compression recovery with sparse noise in frequency domain.As a kind of digital signal in the digital communication,the baseband signal is more and more concerned,and its compression transmission is also an important research direction.However,there are few studies on the compression recovery of baseband signals.In this paper,by adding the finite alphabet character of the baseband signal,the mixed Gaussian a priori is applied to each element in the baseband signal,and the mean of each Gaussian distribution in the mixed Gaussian priori is the element in the finite alphabet.Based on this model,this paper proposes a baseband signal compression recovery algorithm by using Bayesian inference.The algorithm combines the mixed Gaussian model with the Bayesian inference to restore the baseband signal based on the multivariate finite alphabet set,and it can also recovers the baseband signal based on the complex finite character set.Experiments show that the algorithm has higher compression rate than traditional linear programming method,high recovery rate,robustness of algorithm,and obvious performance advantage.In addition,this paper also proposes a compression recovery algorithm when the baseband signal contains frequency domain noise.The algorithm combines the previously proposed baseband signal compression recovery algorithm with Bayesian compressive sensing to impose a priori on the baseband signal and frequency domain noise and uses Bayesian to infer the posterior distributions of the baseband signal and frequency domain noise so that it is possible to simultaneously recover the baseband signal and the frequency domain noise.Experiments show that the algorithm still has a good recovery success rate even when the frequency of noise is relatively large,and still has a good recovery accuracy when the noise power in the frequency domain is large.So the algorithm has obvious performance advantages.In addition,in order to solve the problem that the previously proposed baseband signal compression recovery algorithm can not be directly applied to the case of frequency domain noise,an algorithm for parallel processing of baseband signal and frequency domain noise is proposed by further exploring the a priori characteristics of baseband signal and spurious signal in frequency domain in this paper.Experiments show that the proposed algorithm can recover the baseband signal and the frequency domain noise at the same time,and it also has high recovery success rate.This method extends the application scenario of the previously proposed baseband signal compression recovery algorithm.