2023-01-22
Beam-Steering for Narrow Beamwidth 120 GHz
Antenna Array Using Deep Learning for Radar Application
By
Progress In Electromagnetics Research C, Vol. 129, 35-50, 2023
Abstract
The narrow beam-width 120 GHz industry, scientific, and medical band compact substrate integrated waveguide (SIW) driven antenna's design and characterization are discussed in this study. A low-cost fabrication is ensured by the employment of a single RO4350B substrate layer with SIW feeding. A transition from SIW to a rectangular waveguide is made for measuring purposes. The radiation pattern has been measured. By determining the right feeding phases for the 20 elements, a Deep Neural Network (DNN) is used to softly compute the beam steering. The weighted hybrid Modified Gravitational Search Algorithm (MGSA) - Particle Swarm Optimization (PSO) approach and neural network with back-propagation technique are utilized to beam-steer by anticipating the appropriate feeding phases of the antenna array elements. To evaluate the effectiveness of the approaches, a number of sample instances are given that beam-steer the pattern in a variety of directions. In addition to allowing for the establishment of crucial analytical equations for the synthesis of antenna arrays, the neural network synthesis method also offers a great deal of flexibility between the system parameters in input and output, which makes the synthesis possible due to the explicit relationship given by them. The conventional technique of the phased array is compared with our DNN model for implementing beam steering.
