A novel adaptive variable forgetting factor RLS algorithm

Abstract

This paper investigates the variable forgetting factor least squares method and establishes a nonlinear functional relationship between the forgetting factor and the error signal based on the Sigmoid function. The algorithm overcomes the conflict problem of steady-state performance and dynamic performance of the fixed forgetting factor RLS algorithm. The forgetting factor gradually increases during the gradual convergence of the algorithm, which ensures the algorithm’s steady-state performance while accelerating the algorithm’s tracking speed and convergence speed. At the same time, this paper also analyzes the rules of the parameters α and β and the effects of the parameters α and β on the performance of the RLS algorithm. Finally, computer simulations are conducted, and the results are consistent with the theoretical analysis, confirming that the algorithm outperforms the traditional RLS algorithm.