Analysis of pedestrian's Doppler characteristics using MIMO CW radar

The objective of this study was to describe the utilization of multiple-input multiple-output (MIMO) radar for Doppler estimation which could be applied to automotive radar application to improve pedestrian classification. Our focus in this study is to compare Doppler's estimation results between MIMO radar and single-input single output (SISO) radar. Furthermore, we propose two features extraction from radar measurements for classification targets. The Doppler measurement was carried out using a pseudo MIMO Doppler radar setup using electromechanical scanner and Spectrum analyzer over a center frequency of 4.7 GHz in a laboratory, observing two types of target; human and conductive sphere. The results revealed that the micro-Doppler of the targets using orthogonal MIMO waveform marked a bigger margin between the two types of target when observed in a scatter plot classification compared to SISO measurement.