Efficient techniques in the analysis of array elements for electromagnetic remote sensing

Extensive use of Electromagnetic simulation codes is made in the design of antennas for sensing arrays. Although these codes can achieve very high levels of accuracy they are demanding and hence time consuming in terms of computational load. This restricts their usefulness as design tools and suggests that faster calculation methods should be sought to enhance the design process. Efficient techniques were investigated to reduce the analysis time for broadband element design. Firstly, the interpolation method has been employed in the calculation of wide-angle active reflection coefficients of a TSA element as spatial dependence. A data compression ratio of 9 has been achieved while retaining the obtainable accuracy from EM simulators. Secondly, neural network models have been developed to establish the complicated relationship between the ARC of the TSA elements and key design parameters. When properly trained, the performance of a TSA element can be predicted easily using the trained neural network models and it runs extremely fast compared to EM simulators.