Title :
An adaptive Volterra filtering algorithm with reduced parameters and kernels combination
Author :
Li Fei-xiang ; Zhao Zhi-jin ; Zhao Zhi-dong
Author_Institution :
Telecommun. Sch., Hangzhou Dianzi Univ., Hangzhou, China
Abstract :
A Volterra adaptive filter based on discrete cosine transform (DCT) and kernels combination is proposed. In accordance with the problem that the computational complexity of Volterra adaptive filtering algorithm increases by power series, the quadratic kernels are transformed into a diagonal matrix by DCT, so that the complexity of the filtering algorithm is reduced. In addition, the correlativity of the input signals is reduced, too. And the same order kernels of the Volterra filter are taken parallel combination by the mixing parameters, so the performance of the algorithm is significantly improved. Simulation results show that the proposed algorithm has faster convergence rate, lower steady-state error, better tracking capabilities and better noise robustness.
Keywords :
adaptive filters; computational complexity; convergence; discrete cosine transforms; matrix algebra; nonlinear filters; DCT; adaptive Volterra filtering algorithm; computational complexity; convergence rate; diagonal matrix; discrete cosine transform; kernels combination; mixing parameters; noise robustness; parallel combination; parameter reduction; power series; quadratic kernels; signal correlativity; steady-state error; tracking capabilities; α -stable distribution; Volterra filter; combination of kernels; discrete Cosine transform (DCT); mixing parameter;
Conference_Titel :
Signal Processing (ICSP), 2012 IEEE 11th International Conference on
Conference_Location :
Beijing
Print_ISBN :
978-1-4673-2196-9
DOI :
10.1109/ICoSP.2012.6491655