A fast-converging adaptive FIR technique for channel equalization

Modern advanced hardware technology has made possible the implementation of sophisticated algorithms. The Complex Block Least Mean Square (LMS) algorithm has been widely used in adaptive filtering applications. However, the major drawback of this technique is its dependence on the appropriate choice of the step size. This paper presents the Complex Block Conjugate-gradient LMS algorithm with optimal Individual adaptation of parameters, CBCI-LMS. The proposed technique generates the optimal individual step size for each coefficient of the Finite Impulse Response (FIR) filter at each iteration. In addition, the conjugate gradient principle is employed to find the orthogonal update directions for the adaptive filter coefficients. The performance of the CBCI-LMS is tested for adapting a channel equalizer. The simulation results show that the CBCI-LMS exhibits the faster convergence compared with the Complex Block LMS and the recently proposed CBC-LMS, while maintaining comparable accuracy.