A Data-Reuse Regularized Recursive Least-Squares Adaptive Filtering Algorithm

The recursive least-squares (RLS) algorithm is very popular for adaptive filtering applications, mainly due to its fast convergence rate. The regularized version of this algorithm owns improved robustness features, especially in noisy environments. On the other hand, this robust behavior could influence the tracking capability of the algorithm and its overall convergence performance. To reach a proper balance between these performance criteria, a data-reuse regularized RLS algorithm is proposed in this paper. The specific data-reuse parameter of the resulting algorithm can be used as an additional control mechanism, thus reducing the influence of tuning the main convergence parameter of the RLS filter, which is the forgetting factor. Simulation results obtained in the framework of echo cancellation sustain the performance characteristics.