Robust Feature-Aided Tracking and Identification by Exploiting Joint Target Kinematics and Identification Information

"MRLets Technologies, Inc.proposes an innovative approach for robust feature-aided tracking (FAT) and identification by effectively exploitingjoint target kinematics and identification (ID) information. It has been proposed and proven by thePI that {\it moving target} ID information which is reflected in ``features" is tightly coupledwith target kinematics information. To design an effective and robust feature-aided tracker, one shouldjointly explore both target kinematics and ID information spaces.This proposal aims at: (1) effective information exploitation of target kinematics and ID spaces at a level that has never been explored before; (2) novel feature extraction based on local motion; and (3)design of a robust feature-aided tracker -- joint multiple hypothesis tracking and identification (JMHTI). The proposed research embarks upon the first research of this kind indiscovering the information coupling mechanism between target kinematics and ID spaces, andthe key is the introduction of global motion and local motion. Local motion carries certain kinematics informationand rich ID information and therefore, naturally serves as an informationcoupling mechanism. One of the challenges is the separation of local motion fromcomposite motion information delivered by GMTI, HRR and range/Doppler sensors.We proposed to use a waveletsdomain statistic processing technique based on the hidden Markov trees (HMT) andstructure invariants approa