Beamforming Of Smart Antenna For Cellular Communication Using Variable Step Size Algorithms And Machine Learning Algorithms

Abstract

Antenna technology is one of the pivotal technologies for the improvement of the newlineperformances of wireless communication. Proper use of antenna techniques, enhances newlinespectrum utilization, reduces interference, and increases security. In fourth generation newline(4G) wireless communication International Telecommunication Union (ITU) has newlinesuggested to use Smart antennas. A Smart antenna is an antenna array of any type of newlineantennas, like dipole, microstrip which uses digital signal processor. A Smart antenna newlineafter determination of direction of arrival (DOA), produces radiation beam towards the newlinedesired user and produces null towards the undesired interferer. The 4G wireless newlinecommunication exploits multiple-input multiple-output (MIMO) system where multiple newlineantennas (not in the form of array) are used at the transmitter and receiver ends to newlineemploy the diversity reception. In 5G wireless communications, to provide services to a newlinehuge number of users, the massive MIMO (MMIMO) system is used, where newlinebeamforming of smart antennas plays the central role. The beamforming of Smart newlineantenna provides very narrow targeted radiation beam. In 6G wireless communications, newlinephased array or smart antenna are used for sky and sky-to-terrestrial network newlineconnections via satellite, high altitude platform (HAP) and unmanned aerial vehicles newline(UAV). The beamforming algorithms play a pivotal role for the formation of radiation newlinebeam in a smart antenna. In this thesis, different types of adaptive signal processing newlinealgorithms are used for the beamforming of smart antennas and the performances are newlineinvestigated. Initially, the least mean square (LMS), recursive least square (RLS) and newlinesample matrix inversion (SMI) algorithms are used for the beamforming of smart newlineantennas. Then, to obtain better results for reduction of side lobe levels (SLLs), variable newlinestep-size algorithms are used for beam generation. The research work is extended to the newlinebeamforming using machine learning algorithms, where, better results compared to the newlinevariable step-size algorithms are achieved