Detection of multi-people micro-motions based on range–velocity–time points

摘要

The micro-Doppler effect is a useful signature for classifying various human behaviours. However, most micro-Doppler researches assume that only a single moving target exists during the observation. Their works lack in separating micro-motion features from multi-movers. When more than one target is present, their performance will deteriorate heavily. To address this issue, the authors design a new 3D (three-dimensional) model, range-velocity-time points, to separate and describe multi-mover micro-motions measured by the ultra-wideband radar. These 3D points contain the range-velocity-time information simultaneously. By dividing points in the 3D space instead of single Doppler domain, micro-Doppler signatures of each target can be separated effectively. Multi-people motion simulation results verify the effectiveness of the authors' method.