The research on new adaptive beamforming algorithms

Abstract

In this paper, we propose a novel semi-blind algorithm and a novel transform domain algorithm. The feasibility, stability, complexity and other characteristics of the semi-blind algorithm were analyzed. The novel semi-blind algorithm integrates LMS Least mean square and CMA (Constant modulus algorithm). It uses the LMS algorithm, which is simple to implement and not computationally intensive, but with CMA initialization in order to ensure desirable robustness properties. Then the new transform domain least mean-square adaptive algorithm employing a new adaptive step size control technique is proposed. The variable step size approximates a theoretical optimal one based on a proposed cost function. As a result, considerable improvement in convergence speed is attained in early stages of convergence with small misadjustment near the optimum. Simulation experiments clearly demonstrate the enhanced convergence characteristics of the proposed algorithm.Index Terms: semi-blind algorithm, beamforming ,transform domain , new adaptive step size, variable step size . INTRODUCTIONThe ever-increasing demand for increased capacity in wireless communications services has led to developments of new technologies that exploit space selectivity. This is done through smart-antenna arrays [1-2] and the associated adaptive beamforming algorithms [3][4][5][6][7][8].Adaptive beamforming algorithm is the core of smart antenna technology, now it has been proposed based on the main methods which are non-blind algorithm including least mean square (LMS) [9], normalized least mean square (NLMS) sample matrix inversion method (SMI), recursive least squares (RLS) [10]. The main research directions of the blind algorithm contain the constant modulus algorithm (CMA) [11], decision-oriented algorithm (DDA), de-spread re-spread algorithm, the algorithm based on the cyclic and finite symbol set algorithm and so on. However, they can not meet the high stability, convergence speed, and little amount of calculation at the same time.Common transform domain LMS adaptive filtering is a very popular alternative technique to LMS-type algorithms when the input signal is colored [12][13][14]. The transform domain LMS algorithm attempts to decorrelate the input signal where gains in convergence speed are obtained only after power normalization of the transformed input vector is applied to reduce the eigenvalue disparity of the transformed signal autocorrelation matrix [14]. However, little attention has been paid to the fact that the transform domain LMS inherits the LMS algorithm limitation of having to establish a tradeoff between convergence speed and misadjustment level [15].In this paper, we focus on a new semi-blind algorithm and a new transform domain LMS algorithm with an adaptive step size control method. The first algorithm not only can make full use of the fast convergence of LMS algorithm, small error and high system resource utilization of CMA algorithm, but also can improve the speed and accuracy of the beamforming of next communication system. ...

Computer‐aided experimental studies for adaptive beamforming algorithms and array processing in engineering education

“…However, error signal does not depend on any reference signal. The CMA is given at below equations [33]:…”