Data censoring with set-membership algorithms (original) (raw)

This paper addresses challenges related to data censoring in iterative online processing using Set-Membership Filtering (SMF) algorithms. The focus is on the development of the SM-NLMS (Set-Membership Normalized Least Mean Squares) technique that effectively updates parameters while managing incomplete data sets. By introducing the parameter γ, the method aims to optimize the update rates for various input signal types, demonstrating improved performance in scenarios involving outlier signals and adaptive systems.