The adaptive algorithm plays an important role in ensuring the stability and performance of adaptive systems. In particular, in the case of time-varying systems where parameter invariance cannot be assumed, the performance of the adaptive algorithm becomes more important. The least mean square (LMS) algorithms and the recursive least square (RLS) algorithms have been widely used for the adaptive identification of time-varying systems. In order to apply them to time-varying systems, there have been many studies on the variation types of LMS and RLS algorithms.
Despite the efforts of many researchers, the study on improved adaptive algorithms with faster convergence rates, lower computational complexity and more improved tracking performance still remains an important task for scholars.
Based on the concept of distance in the parameter space, Kim Kwang Ho, a section head at the Faculty of Automation Engineering, has proposed a real-time identification algorithm and compared it with NLMS (Normalized LMS) and RLS. Through the comparison, he has found that the proposed algorithm shows desirable convergence and tracking performance as an adaptive algorithm for rapidly time-varying systems.
The numerical simulation results demonstrate that the proposed algorithm is more effective than other algorithms in adaptive identification for rapidly time-varying systems.
© 2021 Kim Chaek University of Technology