Efficient Initialization for the Adaptive LMS Beamforming Algorithm

• Published in: International journal of applied evolutionary computation Vol. 13; no. 2; pp. 1 - 10
• Main Author: Aounallah Naceur
• Format: Journal Article
• Language: English
• Published: Hershey IGI Global 22.12.2022
• Subjects: Adaptive algorithms; Algorithms; Antenna arrays; Antenna radiation patterns; Antennas; Basic converters; Beamforming; Complexity; Covariance matrix; Matrix algebra; Smart antenna systems
• ISSN: 1942-3594; 1942-3608
• DOI: 10.4018/IJAEC.315635

In spite of the great progress in research work related to the smart antenna field, obtaining an efficient beamforming technique with low complexity, fast converge, and better other performance remains the preferred objective of most researchers. The present work proposes a new version of least mean square (LMS) approach for the beamforming of smart antenna array. The novelty of the proposed algorithm versus its basic version is focalized in its dependence on a new initialization technique, whose aim is to accelerate convergence speed and maintain, at the same time, the algorithm simplicity. The central idea of the proposed technique, which is named new initialized LMS (NI-LMS), is to compute an initial weight vector using only a diagonal matrix extracted from the spatial auto-covariance matrix. Simulation examples are carried out on linear antenna array to demonstrate and validate the effectiveness of the new method. In addition, the computational complexity of the new proposition is analyzed and compared to that of the conventional LMS beamforming approach.