Detection of Electromagnetic Radiations Sources at the Switching Time Scale Using an Inverse Problem-Based Resolution Method—Application to Power Electronic Circuits This paper presents a detailed study about the characterization of electromagnetic near-fields emitted from power electronic circuits. The approach proposed here provides a radiation model, which verifies the uniqueness of the solution using an inverse problem-based resolution method. For this purpose, the characterization method has been combined with an experimental mapping realized at the switching time scale. This mapping is based on a Fourier analysis applied on the measurements acquired from the device under study. It allowed us to trace a prior radiations map over the device under study. This radiation map is then exploited in order to validate the uniqueness of the final result, a result that has been obtained by solving the inverse problem using a genetic algorithm (GA) to optimize the highly nonlinear fitness function. The precision and the accuracy of the method have been validated by the superposition of the obtained radiation models over the circuit topology. The correlation between the sources parameters (position, orientation, geometrical parameters and currents) and the parameters corresponding to the components located at the regions defined by the radiation map has been successfully performed. This allowed us to confirm that the physical reality is respected, and to assume that we have reached the uniqueness of the solution.