Nonconvex Extended Object Tracking Method Based on Set Inversion

To address the problem of target tracking under nonlinear bounded measurements, this paper proposes a nonconvex extended object tracking method based on set inversion. The method first establishes a nonconvex extended object model using the adjacency matrix. A tree-structured constraint propagation algorithm then shrinks the node state prediction set and constructs the set inversion problem based on measurement equation constraints. The shrunk set serves as the initial interval for the set inversion algorithm, which is used to solve the set inversion problem to obtain the optimal state intervals of the extended object nodes. Simulation and experimental results demonstrate the effectiveness of the proposed algorithm in tracking nonconvex extended object.