A Sliding Window RLS - like Adaptive Algorithm for Filtering - stable NoiseMurat

We introduce a sliding window adaptive RLS-like algorithm for ltering-stable noise. Unless previously introduced stochastic gradient type algorithms, the new adaptation algorithm minimizes the L p norm of the error exactly in a sliding window of xed size. Therefore, it behaves much like the RLS algorithm in terms of convergence speed and computational complexity compared to previously introduced stochastic gradient based algorithms which behave like LMS algorithm. It is shown that the new algorithm achieves superior convergence rate at the expense of increased computational complexity.