A novel distributed bearing‐only target tracking algorithm for underwater sensor networks with resource constraints
Wei Zhao,Xuan Li,Zhouqi Pang,Chengpeng Hao
DOI: https://doi.org/10.1049/rsn2.12554
2024-03-23
IET Radar Sonar & Navigation
Abstract:The authors introduce an innovative distributed bearing‐only target tracking algorithm tailored for resource‐constrained underwater sensor networks. It operates within the diffusion cubature information filter framework, effectively fusing non‐linear bearing measurements and conducting state estimation. During the incremental update stage, the algorithm employs the Posterior Cramer‐Rao Lower Bound as a tracking performance metric and proposes a strategy for selecting neighbouring nodes. In the diffusion update stage, it utilises a multi‐threshold event triggering mechanism and proposes an adaptive convex combination weight, demonstrating superior performance in Monte Carlo simulations by significantly reducing energy consumption during information exchange while preserving tracking accuracy. Through theoretical analysis, the asymptotic unbiasedness and convergence of the algorithm have been proven. Underwater sensor networks hold immense potential for advancing the field of underwater target tracking, yet they encounter significant resource constraints stemming from energy storage and communication methods. In order to balance tracking accuracy and energy consumption, the authors present a distributed bearing‐only target tracking algorithm that can be used in underwater sensor networks with resource constraints. Anchored in the diffusion cubature information filter framework, this algorithm achieves fusion for non‐linear bearing measurements and state estimation. During the incremental update stage, individual nodes leverage the Posterior Cramer‐Rao Lower Bound as a metric for tracking performance. Subsequently, a strategy for selecting neighbouring nodes is introduced, ensuring tracking accuracy while efficiently kerbing energy consumption. In the diffusion update stage, a multi‐threshold event triggering mechanism is employed to partially diffuse the intermediate estimation. Additionally, an adaptive convex combination weight is proposed for cases involving partial diffusion. Through theoretical analysis, the asymptotic unbiasedness and convergence of the algorithm have been proven. Through Monte Carlo simulation experiments, the authors verify that the algorithm is superior to existing algorithms. Furthermore, the algorithm significantly reduces energy consumption in information interaction, minimising tracking accuracy loss.
telecommunications,engineering, electrical & electronic