Adaptive beamforming of MIMO system using optimal steering vector with modified neural network for channel selection

Abstract

The forming of adaptive beam can improve the throughput of the system to a great extent by means of matching the parameters of transmitters to that of the wireless channels that are time-variant. The quality of the channel state is very crucial to the adaptive forming. The Multiple-Input Multiple-Output (MIMO) systems are known to provide some very significant gains in the spectral efficiency as well as its reliability. This has been based on an assumption that the transmitter and the receiver will have knowledge of the coefficients of the channels. In reality, however, they will have to be estimated or sometimes predicted. There are some popular methods that are used for the estimation of the channel which is made by means of using the pilot symbols and also the Space-Time Block Codes (STBCs). Both these methods will not avail the time-learning channels even during the transmission of some meaningful data. In this work, a light weight neural network is proposed for the channel selection. The proposed Artificial Neural Network (ANN) is duly optimized with the Particle Swarm Optimization (PSO) and the Bacterial Foraging Optimization (BFO)-based algorithms for enhancing the predictions. The method is popular and is robust adaptive as a beamforming technique and optimized weights will be used for training the ANN effectively. The results when compared prove the advantages of the techniques proposed.