Microstrip antenna design using neural networks optimized by PSO

Microstrip antennas became very attractive due to their features like small dimensions, simplicity, conformability and low cost of production. The production these antennas for wireless communication is focused to design efficient, wide band, antennas that easily can be incorporated into various systems. Design parameters of microstrip antenna need to be calculated in very high accuracy to ensure proper functioning. Design of the rectangular or circular microstrip patch antenna requires high accuracy of calculations, and it is not so easy task. The known models (transmission line model or cavity model) give less accurate results. Simulations based on full-wave analysis techniques, such as CST Studio SuiteTM, enable more accurate analysis of microstrip antenna characteristics, but they suffer from time consuming, intensive computations. Demand for fast and enough accurate computations can be satisfied using neural network models. This work studies microstrip patch antenna design using artificial neural network optimized by PSO (Particle Swarm Optimization) algorithm. According to neural network design results an antenna was made. The calculated and neural network results are compared with measured antenna values.