Generalized Maximum Total Correntropy Adaptive Filtering Algorithm

In the errors-in-variables model which both input and output signals are contaminated by noise, the total least squares algorithm has been widely used. But in the case of impulse noise interference, its convergence performance will deteriorate. Therefore, in order to deal with the impulse noise-contaminated errors-in-variables model, this paper combines the generalized maximum correntropy criterion with the total least squares estimation method, and proposes a robust generalized maximum total correntropy adaptive filtering algorithm. Through the comparison of algorithm simulation results, it can be concluded that the proposed algorithm has better convergence performance and robustness under impulse noise environment, and can effectively suppress the existence of impulse noise.