Pedestrian recognition in automotive radar sensors

Automotive radar sensors in the 24 GHz frequency domain measure target range and Doppler frequency simultaneously even in multiple target situations. The all weather capability of radar sensors is an important argument for their application in modern cars and in driver assistant systems. The measurement accuracy and target resolution depends on the applied waveform. Therefore the Multi Frequency Shift Keying (MFSK) and the Chirp Sequence waveform will be considered and compared in this paper. The classical radar sensor measures target range, radial velocity and azimuth angle. Additionally the radar sensor should be able to distinguish between different objects like pedestrians and cars in automotive applications. In this case the target recognition quality depends mainly on the signal-to-noise ratio (SNR), the signal features and their measurement accuracy, respectively. In this paper the quality of the target recognition process will be quantitatively measured and compared for two different radar waveforms.