Multi-target tracking via dynamic-programming based track-before-detect
Wei Yi, Mark R. Morelande, Lingjiang Kong, Jianyu Yang, Xiaobo Deng · 2012
This paper addresses the multi-target tracking problem through the use of dynamic programming based track-before-detect (DP-TBD). The usual way of implementing DP-TBD in a multi-target scenario is to adopt a multi-target state, which is the concatenation of individual target states. But this method involves the solution of a high-dimensional joint maximization which is usually computationally prohibitive. Besides, it suffers from the unknown number of targets since the dimension of the multi-target state has to be determined before DP integration. In this work, via utilizing the structure of target dyamics, a multi-target DP-TBD algorithm is developed to approximately solve the joint maximization in a efficient and accurate manner. Also, a generalized detection procedure is adopted in a way that the dimension of the multi-target state is no longer needed be to predetermined, therefore single-target and multi-target scenarios are handled in the same framework. Simulation example shows that the proposed algorithm can efficiently and reliably track multiple targets even targets are in proximity.