Neural network based switch beam smart antenna

A neural network based switched beam smart antenna has been proposed. A feed-forward back propagation artificial neural network (FFBPANN) was applied in antenna beam switching. The network takes samples of the calibration of the footprint of the base station, in terms of azimuth and elevation angle, and maps it to the radiation pattern of different base station antenna arrays covering the area. It was treated as a classification problem. The developed methodology was tested for a 4×4 Butler matrix and the result obtained shows an improvement in the performance of beam switching. The developed network can be used to switch different antennas of the base stations depending on the position of the target.