Data-Driven Back Projection for Target Scattering Modeling VIA Radar Sensor Network

Modeling target scattering accurately from sensors is crucial for imaging and recognizing non-cooperative objects. This work proposes a data-driven back projection (DDBP) method for target scattering modeling. It performs BP on the data of multiaspect high-range resolution profiles (HRRPs) acquired from a radar sensor network within individual snapshots. The advantage of DDBP lies in its independence from Doppler or micro-Doppler features, ensuring robust imaging regardless of the target's motion state and exhibiting satisfactory performance even in time-critical application scenarios. In comparison with ISAR, the DDBP further exploits the spatial scattering energy integration characteristics, leading to enhanced imaging performance. Field experiment results validate the effectiveness of the proposed approach.