Compressed Sensing Approach for Fast Collision Warning Radar

A compressed sensing approach is utilized to estimate positions and velocities of moving targets from multi-sensor continaous-wave (CW) radar data. A possible field of application is in the area of collision warning systems. In such eases, either speeds of the targets are high or their distances to the system are short dictate high demands on the processing time. Compressed sensing is seen as a suitable approach to the problem since it operates on sparse data and has the potential to significantly save computing time. In this anti-collision application, only few hazardous targets are assumed which seems to be an adequate condition for compressed sensing. The prototype comprises 4 CW sensors working in the 24 GHz industrial, scientific and medical (ISM) frequency band. Simulated data with ideal point scatterers and a metallic sphere as well as measured data with a sphere are chosen to test the algorithm. Comparisons of the results verify the capability of the proposed method to detect moving targets and to estimate their motion states.