Reinforcement learning for robust radar tracking
Quanyi Cheng, Lei Chen, K. Zhang, ZR Wang · IET conference proceedings. · 2021
In complicated environments with strong clutter and dense targets, radar tracking may encounter the issue of ambiguities in association. This paper proposes a reinforcement learning approach to reduce the number of misassociations in radar tracking. For each existing track, the characteristics of the target are learned to form a classifier, and then the classifier is employed for detection and association for this track. The parameters of each classifier will be updated using the feedback of the updated track. Experimental results on synthetic data and real data verify the effectiveness of our method.