Research On Distributed Adaptive Channel Estimation Algorithm Based On Compressed Sensing In WSNs

With the development of low-power wireless communication technologies,Wireless Sensor Networks(WSNs)technology has emerged and attracted widespread attention.At present,the main challenges of signal processing in WSNs are channel estimation,node clock synchronization,denoising of information transmited between nodes,and distributed parameter estimation.This paper mainly studies the application of distributed estimation algorithm based on compressed sensing in channel estimation.Firstly,the two kinds of centralized and distributed processing technologies in WSNs are introduced,and the superiority of distributed-collaborative cooperation is described.Introduced a wide range of applications of distributed algorithms.Then introduced the mathematical model of channel estimation and the common LMS channel estimation algorithm.Secondly,distributed distributed collaboration models are introduced,focusing on diffusion-based collaborative strategies.Therefore,the diffuse LMS(D-LMS)algorithm under the Metropolis convergence criteria is introduced,which is one of distributed adaptive channel estimation algorithms.This paper analyzes and verifies the convergence of D-LMS.In recent years,high-dimensional data has often appeared in relevant fields of academia and industry.With the continuous improvement of application requirements,the dimensions of the parameters to be estimated are also increasing,and the use of traditional algorithm processing often brings higher complexity and greater energy consumption,as well as lower convergence speed and lower estimation accuracy.Considering that the sensor nodes in the WSNs have the characteristics of energy,storage space and limited computing power,the distributed estimation system must be improved.This paper makes full use of the sparsity of the channel and introduces a powerful tool for compressed sensing in the framework of distributed channel estimation.Compressed sensing-based distributed channel estimation system reduces the data dimension for adaptive estimation to reduce the node's computational complexity.The feasible basis of the new system comes from the reconstruction theory of compressed sensing under noisy measurement.Finally,in the framework of distributed channel estimation based on Compressed Sensing,the sensing matrix is optimized to increase the reconstruction accuracy of the channel coefficients.The matrix optimization algorithm has simplified the steps of the traditional optimization algorithm and combines the column vector normalization and the column mutual coherence coefficient optimization into an optimization problem.Simulations verify the advantages of the new algorithm in terms of matrix average mutual coherence and channel coefficient reconstruction accuracy.In this paper,the three technologies of distributed cooperation,adaptive signal processing and compressed sensing are organically combined and applied to channel estimation.The main contributions are as follows:1.D-LMS algorithm is used for channel estimation in WSNs,and analysis its convergence;2.Combining Compressed sensing into a distributed channel estimation system,becoming CS-DCE system;3.Optimizing the sensing matrix in the CS-DCE system and improve the CS-DCE reconstruction performance;4.Simulations verify the performance of the CS-DCE system corresponding to the proposed measurement matrix optimization algorithm.