Diffusion proportionate affine projection sign algorithm for distributed estimation over network

This paper proposes a diffusion proportionate affine projection sign algorithm for distributed estimation of sparse vector over network. The algorithm is derived by minimizing l
₁
-norm intermediate error vector subject to a weighted constraint on the filter coefficients, where the positive definite weighting matrix is designed to accelerate the convergence of the nonzero coefficients for sparse vector estimation. Due to l
₁
-norm error minimization, the algorithm achieves robust performance against the impulsive interferences. Simulation results show that the proposed algorithm outperforms the existing distributed estimation strategies in terms of the convergence rate in various impulsive-interference environment.