A wavelet-based radial-basis function neural network approach to the inverse scattering of conducting cylinders

A new approach, based on the radial-bias function neural network (RBF-NN) combined with wavelet transform, is presented for the estimation of the locations and radii of conducting cylindrical scatterers. The discrete wavelet transform coefficients of the electric-field values scattered by the cylinder are fed into the RBF-NN, whose outputs are the location and the radius of the cylinder. The efficiency of the proposed approach is compared with the approach where the field values are directly used. The performance of the wavelet-based approach for noisy field measurements is also investigated.