Joint Tracking and Identification of Ground Target with Complex Motion Using Heterogeneous Multi-Sensor Data

A novel and systemic joint target tracking and target identification (for short JTI) approach for ground targets with complicated dynamics using heterogeneous multi-sensor data. With the increase in the number of vehicles and the development of sensor technology, vehicles usually face complex motion (high maneuverability, high nonlinearity, etc.) and heterogeneous multi-sensor data are available. Furthermore, target tracking and target identification are usually coupled. Therefore, this paper proposes a novel JTI approach to jointly handle the ground target tracking and identification problem by fully utilizing the multi-sensor data information. Firstly, we formulate the multi-sensor JTI problem with complex motion by models. Then we propose to adopt a joint generalized Bayes risk by unifying tracking and identification risks. A JTI solution containing a tracker and an identifier is presented by taking their couplings and also the multi-sensor data information into account. Specifically, for tracking, we propose a multiple model unscented Kalman filter approach, which can not only the address the complex target motion, incorporate the multi-sensor data information and can also be easily handled mathematics. Finally, simulation results demonstrate that the proposed JTI approach can effectively take advantages of the coupling between ground target tracking and identification, as well as the useful information from multi-sensor data, and ultimately achieving the best overall performance.