Robust Estimation Method of $K$-Distribution Shape Parameter

Abstract

Parameter estimation for the $K$ -distribution plays an important role in marine radar and sonar applications. However, an outlier in the observation is an obstacle to achieving accurate and robust estimation solution of the $K$ -distribution parameter. If not handled properly, the result of estimation may be disturbed severely. To address this problem, this paper proposes a shape parameter estimation algorithm based on robust statistics. In contrast to the conventional method of moment (MoM) approach, the proposed algorithm can reduce the interference of the outlier significantly, meanwhile retaining a high estimation efficiency. Simulation results confirm the good performance of the proposed method.