Neural network solution of the circular loop antenna radiation problem

The circular cylindrical antenna is a simple, inexpensive, versatile, and very popular antenna type, which has received much attention due its wide range of applications. The exact values of the radiated near and far electromagnetic fields have recently been analytically evaluated, in terms of complex series involving Legendre functions of the second kind and half-integral order. The inverse problem of determining the loop antenna parameters (radius and current) causing specific field levels at one or more points of interest is even more complex. In order to find the solution, avoiding the associated lengthy and time-demanding mathematical analysis, we apply artificial neural network modeling. The proposed models consist of a feedforward back-propagation and a radial basis neural network trained with theoretical data. The results obtained are found to be in perfect agreement with the exact theoretical data.