A New Variable Step Size LMS Algorithm Based on Softsign Function

Abstract:
The traditional LMS algorithm is widely used in adaptive filtering algorithms, but it suffers from inconsistency between convergence speed and steady-state error, resulting in poor system performance. A new variable step size LMS algorithm is proposed to address the above shortcomings. The algorithm uses the Softsign function as the adjustment function of the step size factor, and introduces the perturbation of the absolute estimation error into the dynamic change process of the filter weight vector, so that the weight vector is constantly updated towards the optimal value, in order to enhance the convergence speed of the algorithm. Simulation results show that the algorithm outperforms existing variable step size LMS algorithms based on other functions in terms of convergence speed and steady-state error.