Sparse Reconstruction Off-grid OFDM Time Delay Estimation Algorithm Based on Bayesian Automatic Relevance Determination

Abstract

Aiming at the problem of single measurement vector (SMV) OFDM signal delay estimation in complex environment, a sparse reconstruction off-grid delay estimation algorithm based on Bayesian automatic correlation determination (OGBARD) is proposed. The algorithm adopts Bayesian framework, from the perspective of further mining useful information, introduces asymmetric autocorrelation to determine a priori parameters and off-grid parameters, and integrates them into the parameter estimation process, which improves the delay estimation accuracy under SMV and low signal-to-noise ratio (SNR) condition. Firstly, based on the channel frequency response (CFR) estimation value of the OFDM signal physical layer protocol data unit, the algorithm constructs a sparse representation model with off-grid parameters, and then introduces the probability hypothesis of noise vector, off-grid parameters and sparse coefficient vector in the model. Finally, based on Bayesian inference, the expectation maximization (EM) algorithm is used to estimate the hyperparameters. Simulation experiments show that the proposed algorithm has better estimation performance than the comparison algorithm and is closer to the Cramer–Rao Bound (CRB).