Multi-sensor Measurement Information Fusion for Maneuvering Target Recognition and Tracking

Abstract

This paper proposes multi-sensor information fusion algorithm for maneuvering target recognition and tracking based on mobile sensor networks (MSNs). In order to improve target recognition accuracy, the modified evidence combination rule is designed to obtain the identity information of target based on the Dempster-Shafer (DS) algorithm. The target recognition process is designed by two stages, namely, sensing stage and fusion stage. During the first stage, each sensor node in the MSNs can obtain the feature information of the target, which makes up the measurement information. In the information fusion stage, each sensor combine the information exchanged from neighboring nodes with local measurement information to obtain the high-precise identity information of the maneuvering target based on the improved DS algorithm. Then, the identity information can be used to obtain the accurate state estimation of the maneuvering target. For the purpose of obtaining the accurate state estimation of target, the target tracking process is designed by local filtering stage and fusion stage. Finally, a numerical simulation is designed to illustrate the effectiveness of the algorithm.