A minimum-error entropy criterion with self-adjusting step-size (MEE-SAS)

In this paper, we propose a minimum-error entropy with self-adjusting step -size (MEE-SAS) as an alternative to the minimum-error entropy (MEE) algorithm for training adaptive systems. MEE-SAS has faster speed of convergence as compared to MEE algorithm for the same misadjustment. We attribute the self-adjusting step-size property of MEE-SAS to its changing curvature as opposed to MEE which has a constant curvature. Analysis of the curvature shows that MEE-SAS converges faster in noisy scenarios than noise-free scenario, thus making it more suitable for practical applications as shown in our simulations. Finally, in case of non-stationary environment, MEE-SAS loses its tracking ability due to the “flatness” of the curvature near the optimal solution. We overcome this problem by proposing a switching scheme between MEE and MEE-SAS algorithms for non-stationary scenario, which effectively combines the speed of MEE-SAS when far from the optimal solution with the tracking ability of MEE when near the solution. We demonstrate the performance of the switching algorithm in system identification in non-stationary environment.