QR-based robust diffusion strategies for wireless sensor networks using minimum-Wilcoxon-norm

Impulsive noise and interference are always present in the environment in addition to additive white Gaussian noise that corrupts the measured data. In such cases, conventional adaptive estimation algorithms based on least squares error cost function provides poor performance in estimating the optimum parameter. In order to alleviate this shortcoming, a robust diffusion strategy based on QR decomposition and the Wilcoxon norm is proposed. The proposed QR-based diffusion minimum-Wilcoxon-norm (DMWN) provides faster convergence than the DMWN. To demonstrate the efficacy of the algorithm, simulations are carried out with different percentage of outliers in the desired data and found to be robust against traditional methods. Moreover, the condition for convergence in mean is analysed, and the algorithm is observed to be stable.