Research On Multi-frame Detection Techniques For High Frequency Radar Moving Targets

The fast development of electronic countermeasures lessens the RCS (radar crosssection), which deteriorates the detection capability of radar. Besides, the increase oftargets maneuverability worsens the radar tracking capability and lessens thesurveillance distant. For these problems, radar techniques need further updates to meetthe changes of environment and targets. However, radar target detecting and trackingprogresses to a dim target detecting and tracking situation due to these problems. Thedissertation researches the track-before-detect techniques in the radar dim targetdetecting and tracking, and proposes a revised dynamic programming algorithm for theradar multi-frame detection. The details of the dissertation summarizes as follows.Firstly, the dissertation introduces the radar simulation platform which incorporatesradar target state model and radar target measurement model, presents the radardetection theory, and discusses the construction of the radar ARD(azimuth-range-Doppler) data.Secondly, traditional TBD methods and their simulations are presented, includingdynamic programming (DP), Hough transform (HT) and particle filtering (PF). Twoalgorithms for the dynamic programming are proposed, i.e. DP with power and DP withstate likelihood. Simulations results are presented and discussed in each step of the DPwith power, DP with state likelihood and HT algorithm.Finally, a revised dynamic programming algorithm and its simulation are putforward. In this section, the dissertation begins with the CFAR (constant false alarm rate)techniques and combines CFAR technique with traditional dynamic programming toform this revised method. The comparative simulation results of the revised and theother three algorithms are carried out and analyzed. It can be concluded that the revisedalgorithm outperforms the other three algorithms both in detection capability andcomputational efficiency.All in all, this dissertation researches several algorithms in TBD techniques andproposes a revised DP algorithm which has an improvement in detecting andcomputational efficiency.