A variable step-size lmp algorithm for heavy-tailed interference suppression in phased array radar

Abstract

A new variable step-size Least Mean p-norm (VSS-LMP) algorithm is proposed for phased array radar application with space-time adaptive processing to combat heavy-tailed non-Gaussian clutters. The algorithms automatically change the step size according to the estimated p-th and (2p - 2)-th moments of the error, where 1 ≤ p ≤ 2. The algorithm is evaluated via a space-slow-time STAP example and the excess Mean Square Error (MSE) and misadjustment results show that the proposed VSS-LMP converges fast and reaches lower steady-state error than the fixed stepsize LMP. It also provides a better compromise between convergence speed and low steady state error than existing VSS Least Mean Square (LMS) algorithms in both Gaussian and Compound K clutter environments.