Modified particle filtering‐based robust estimation for a networked control system corrupted by impulsive noise
Xuehai Wang, Feng Ding
Abstract
Abstract The non‐Gaussian characteristic of the impulsive noise significantly degrades the performance of the ‐norm‐based identification algorithms. To overcome the negative effects of impulsive noise to system identification, this article develops a modified particle filtering‐based robust multi‐innovation gradient (MPF‐MIG) algorithm for a networked control system corrupted by impulsive noise. The proposed algorithm is formulated based on a continuous logarithmic mixed p ‐norm cost function, which can generate an adjustable gain that adapts to the data quality. A modified particle filtering is designed to estimate the unknown true output, which is corrupted by an impulsive noise without explicit probability density function. The simulation examples exhibit that the MPF‐MIG algorithm has better robustness and higher estimation accuracy than the conventional multi‐innovation gradient algorithm.