A study of compressed sensing based collision avoidance by multi-sensor CW radar data

摘要

Compressed sensing algorithms are studied for the purpose of detecting targets and estimating their motion states in a collision warning radar system which consists of 4 continuous-wave (CW) modules operating in the industrial, scientific, and medical (ISM) frequency band around 24GHz. A method for sparse representation and a discretization adjusted to the problem are investigated. Also, the capabilities of various estimation algorithms belonging to the two groups of basis and greedy pursuits are compared considering the fine discretization in the six dimensional solution space.