Gradient based technique for electromagnetic layered earth model data inversion

The primary aim of this study is to improve the computational efficiency of full-waveform inversion algorithms for ground-penetrating radar applications. This can be achieved if an exact explicit solution for the gradient is known. This valuable information permits to compute the Jacobian and al lows for improved convergence. The problem of local minima and non-uniques can be approached by using different starting models and comparing the different results, assessing the corresponding data misfit functions, and model interpretations using geological or geotechnical knowledge when conceivable.