Compressed sensing based target tracking using raw radar measurements

Track before detect (TBD) based approaches declare the presence of targets and their corresponding tracks through jointly processing several consecutive scans based on the targets' kinematics. A TBD algorithm could improve track accuracy and follow low signal-to-noise ratio (SNR) targets at the price of an increase of the computational complexity. In this work, a novel TBD algorithm is proposed in a compressed sensing based framework. Different from the existing TBD algorithms that adopt hypothesis test using the radar measurement, the proposed algorithm reconstructs the whole radar scenario (DOA-Doppler plane) for each range gate at consecutive scans using compressed sensing method. Temporal tracking is performed inside each scan (in the three dimensional space comprised of range, DOA and Doppler), as well as among consecutive scans. The proposed algorithm avoids the nonlinear problem since it detect the targets directly in the reconstructed three dimensional range-DOA-Doppler space. The proposed algorithm could track and detect multiple targets at the same time. Moreover, the proposed algorithm could track and detect the fast moving targets captured by a successive illuminations during one scan.