Constrained iterative technique with embedded neural network for dual-polarization radar correction of rain path attenuation

A new stable backward iterative technique to correct for path attenuation and differential attenuation is presented here. The technique named, neural network iterative polarimetric precipitation estimator by radar (NIPPER), is based on a polarimetric model used to train an embedded neural network, constrained by the measurement of the differential phase along the rain path. Simulations are used to investigate the efficiency, accuracy, and the robustness of the proposed technique. The precipitation is characterized with respect to raindrop size, shape, and orientation distribution. The performance of NIPPER is evaluated by using simulated radar volumes scan generated from S-band radar measurements. A sensitivity analysis is performed in order to evaluate the expected errors of NIPPER. These evaluations show relatively better performance and robustness of the attenuation correction process when compared with currently available techniques.