A speed convergence Least Squares Constant Modulus Algorithm for smart antenna beamforming

Abstract

Adaptive algorithms used in smart antenna systems have been thoroughly studied along the years. Well known algorithms like the LMS (Least Mean Squares) were improved and other more performant algorithms were created. An example is the VSSLMS (Variable Step Size LMS) algorithm, with Kwong and Aboulnasr variants. The LMS class of algorithms uses a reference signal in order to reach convergence. Other class of algorithms like the CMA (Constant Modulus Algorithm) does not use a reference signal and has a good performance for signals with a constant envelope. In this paper we propose a hybrid blind equalization algorithm based on the combination of the LMS and CMA algorithms plus a speed convergence module that provides a very fast convergence speed. We used this algorithm with a beamforming smart antenna network in order to point the main beam of the radiation pattern in certain directions.