Joint target recognition and tracking using class specific features

Recognition and tracking ground targets using measurements from ground moving target indicator (GMTI) and high range resolution (HRR) profiles are considered jointly to enhance both the discrimination and tracking capability for closely-spaced targets. We propose a feature extraction technique based on the target kinematic information from the target state estimator. The classifier matches aspect angle dependent target features for different target types and provides target type information to the tracker. The likelihood of target type is incorporated into the assignment cost to solve the measurement-to-track association problem and improve the tracking accuracy. Results using HRR profiles from the MSTAR data set and GMTI measurements show that both target recognition and tracking accuracy are significantly improved compared with those using the classifier and tracker alone.