A Robust Total Least Mean Square Algorithm For Nonlinear Adaptive Filter

The robust nonlinear adaptive filtering problem based on Volterra model is researched when the input and output observation data are both corrupted by noise in this paper. On the basis of minimizing Volterra total mean square error (VTMSE), a robust total least mean square adaptive filtering algorithm for nonlinear Volterra filter is proposed. The performance analysis demonstrates that the robust performance of the presented algorithm is nicer than other existed algorithms. And simulation results have also shown the prominent advantages of the presented algorithm, it can't only permit to use larger learning factor, but its convergence precision is remarkably higher than other algorithms under higher noise environments.