The Maximum Correntropy Criterion-Based Robust Hierarchical Estimation Algorithm for Linear Parameter-Varying Systems with Non-Gaussian Noise.

The traditional identification algorithms are generally based on the quadratic criterion functions and perform well under the interference of Gaussian noise. However, their performances become worse when non-Gaussian impulsive noise is encountered in practical situations. By introducing the maximum correntropy criterion and utilizing its robustness against outliers, this paper develops a robust recursive algorithm for the linear parameter-varying system with the non-Gaussian noise. The algorithm is not only simple and easy to implement, but also can resist the negative influence of the non-Gaussian noise. Moreover, a maximum correntropy criterion-based robust hierarchical algorithm is presented to increase the calculative efficiency. Three simulation examples are provided to demonstrate the validity of the proposed algorithms. [ABSTRACT FROM AUTHOR]