Alternative Extended Block Sparse Bayesian Learning for Cluster Structured Sparse Signal Recovery

Clustered sparse signals recovery with unknown cluster sizes and locations is considered in this paper. An improved alternative extended block sparse Bayesian learning algorithm (AEBSBL) is proposed. The new algorithm is motivated by the graphic models of the extended block sparse Bayesian learning algorithm (EBSBL). By deriving the graphic model of EBSBL, an equivalent cluster structured prior for sparse coefficients is obtained, which encourages dependencies among neighboring coefficients. With the sparse prior, other necessary probabilistic modelings are constructed and Expectation and Maximization (EM) is applied to infer all the unknowns. The alternative algorithm reduces the unknowns of EBSBL. Numerical simulations are conducted to demonstrate the effectiveness of the proposed method.