Fast Diffusion Minimum Generalized Rank Norm Based on QR Decomposition

The outliers or impulsive noise is prominent in practical wireless sensor networks due to saturation effects, non-linearities, malfunction of sensors, environmental abnormalities, etc. The classical diffusion algorithms based on mean square error cost function are sensitive to outliers and their performance degrades in the presence of outliers in either desired data or the input data. A fast diffusion minimum generalized rank norm based on QR decomposition (FDGR-QR) is proposed, which is robust against outliers in both desired and input data and has faster convergence than the state of the art algorithms. Different outlier percentages distribution across the network is considered for the simulations based experiments. The proposed algorithm is validated for both stationary and non-stationary parameter estimation scenarios.