Oversampled Pulse Compression Based on Signal Modeling: Application to CONSERT/Rosetta Radar

The context of space missions implies very strong constraints on the design of the instruments. For instance, mass, data rates, power consumption, clock frequency, and accuracy are very limited. These limitations degrade the performance and the accuracy of the measurements. This paper is related to the study of signals sampled at a low frequency. More exactly, we propose an oversampled pulse compression method for radar signals, which allows to reconstruction the signals with a much higher sampling rate. Consequently, this method allows significantly improving the multipath amplitude and the delay estimation accuracy. The efficiency of our method is clearly demonstrated with real measurements. These latter correspond to calibration measurements performed on ground with the radar COmets Nucleus Sounding Experiment by Radio-wave Transmission. The latter allows performing the tomography in the transmission of 67P Churyumov-Gerasimenko comets nucleus in the frame of the Rosetta/European Space Agency space mission. The results show that the proposed method allows estimating the delays with accuracy better than 10% of the sampling period in the case of a sampling frequency value close to the Nyquist limit. The analysis of the measured signals shows how a small frequency aliasing power degrades the delay estimation performance. This problem can be resolved using an estimation method based on an end-to-end signal modeling.