Asymmetric Correntropy for Robust Adaptive Filtering

In recent years, correntropy has been successfully applied to robust adaptive filtering eliminate adverse effects of impulsive noises or outliers. Correntropy is generally defined as the expectation a Gaussian kernel between two random variables. This definition reasonable when error variables symmetrically distributed around zero. For case asymmetric distribution, symmetric however inappropriate and cannot adapt distribution well. To address this problem, in brief we propose new variant correntropy, named which uses an model function. addition, algorithm based on developed its steady-state convergence performance analyzed. Simulations are provided confirm theoretical results good proposed algorithm.