Research on Channel Equalization of Kernel Algorithm Based on Decorrelation Principle
Ying Wang · International Core Journal of Engineering · 2021
In this paper, the decorrelation least mean square (DLMS) algorithm is mapped to the kernel space, and the decorrelation kernel least mean square (DKLMS) algorithm is derived. Using the idea of decorrelation and normalization, the problem of slow convergence caused by the high correlation of input data is solved. The simulation experiment proved that the DKLMS algorithm has a faster convergence rate and better steady-state performance than the KLMS algorithm, and its application in nonlinear channel equalization is studied.