Solving inversion problem for refractivity estimation using Artificial Neural Networks

Atmospheric refractivity index is one of the most important variable that effects the propagation direction of radio waves. This means a radar or communication system can show unexpected behaviour and performance depending on this variable. Estimation of refractivity characteristics of atmosphere is possible by using radar clutter data. This method is called refractivity from clutter (RFC). RFC is a nonlinear inversion problem. In this work, Artificial Neural Networks are studied to solve inversion problem for refractivity estimation. A training data set had to be prepared to represent ducts with refractivity parameters. Fortunately, learning and generalization capability of Neural Networks (NN) is very helpful in this point, so a good mapping of refractivity parameters can be enough for solving inversion problem.