Radar signal simulation by stochastic modification of its experimental records

Signal simulation is a very important tool in radar system design. Due to the complexity of radar signal structure the traditional approach to simulation needs calculation of the signal samples using an appropriate simulation algorithm. The definition of the algorithms should be preceded by tedious statistical analysis of the available signal data but the result of the simulation is not always satisfactory. This paper presents new approach to signal simulation that is based on manipulation of complete experimental signal records. In the proposed method of simulation the time consuming calculations of signal samples are replaced by generation of new signal samples on the model of the recorded ones. It is achieved by appropriate modification of real signal data respecting invariance of important signal feature(s) and is implemented by stochastic processing of the signal records. The identification of characteristic signal feature(s) is sine qua non condition of successful signal simulation. The proposed approach increases the efficiency of traditional method of simulation implementation useful among others for signal data bases construction.