Robust augmented Volterra adaptive filtering.

• This paper proposes a new nonlinear system model for signal processing and develops its corresponding adaptive algorithm. • This paper also gives a variant version of the proposed algorithm to reduce its complexity. • The performance analysis is derived and verified by simulations. • The superior behavior is demonstrated by examples of NSAEC and nonlinear channel identification. In the field of nonlinear signal processing, Volterra filter is generally used as an effective tool. The utilization of the augmented model can make the filter maintain its merit in both circular and non-circular signals. From this point of view, this paper proposes a new online nonlinear system model called the augmented Volterra model. To combat impulsive noise, we introduce the complex correntropy to the Volterra filter, and propose the augmented Volterra recursive maximum complex correntropy criterion (A-VRMCCC) algorithm. Considering the computational cost of the augmented Volterra model, we propose the decomposable method to effectively reduce its computational load. Additionally, the performance of A-VRMCCC has also been studied. Finally, simulation results validate the correctness of the performance analysis, and show that the newly proposed algorithms have a competitive advantage over other algorithms. [ABSTRACT FROM AUTHOR]