TY - JOUR
T1 - Joint Average Fusion and Spatial Registration
T2 - GA-Based Distributed Multitarget Tracking
AU - Li, Guchong
AU - Zhang, Yuchen
AU - Li, Tiancheng
AU - Wang, Xueqian
AU - Li, Gang
AU - He, You
N1 - Publisher Copyright:
© 1965-2011 IEEE.
PY - 2026
Y1 - 2026
N2 - In this article, we address distributed multitarget tracking (DMTT) in spatially unregistered sensor networks (uSNs), which necessitates the joint estimation of multitarget states and intersensor spatial registration (SR) parameters. To this end, we propose a computationally efficient DMTT algorithm within the probability hypothesis density (PHD) filtering framework. Specifically, by exploiting geometric average (GA) fusion, we develop a recursive SR estimation scheme referred to as PHD-GA-SR. An exponential Gaussian mixture implementation is further derived for the SR posterior, while the local single-target probability density functions are represented by GMs. In addition, Kullback–Leibler divergence bounded pruning and maximum a posteriori estimate are investigated, and the computational complexity of the proposed approach is analyzed. Simulation results demonstrate that the proposed PHD-GA-SR achieves a favorable tradeoff between estimation accuracy and computational efficiency in both static and dynamic uSN scenarios.
AB - In this article, we address distributed multitarget tracking (DMTT) in spatially unregistered sensor networks (uSNs), which necessitates the joint estimation of multitarget states and intersensor spatial registration (SR) parameters. To this end, we propose a computationally efficient DMTT algorithm within the probability hypothesis density (PHD) filtering framework. Specifically, by exploiting geometric average (GA) fusion, we develop a recursive SR estimation scheme referred to as PHD-GA-SR. An exponential Gaussian mixture implementation is further derived for the SR posterior, while the local single-target probability density functions are represented by GMs. In addition, Kullback–Leibler divergence bounded pruning and maximum a posteriori estimate are investigated, and the computational complexity of the proposed approach is analyzed. Simulation results demonstrate that the proposed PHD-GA-SR achieves a favorable tradeoff between estimation accuracy and computational efficiency in both static and dynamic uSN scenarios.
KW - Distributed multi-target tracking
KW - Gaussian mixture
KW - geometric average
KW - probability hypothesis density
KW - spatial registration
UR - https://www.scopus.com/pages/publications/105048609586
U2 - 10.1109/TAES.2026.3725668
DO - 10.1109/TAES.2026.3725668
M3 - 文章
AN - SCOPUS:105048609586
SN - 0018-9251
VL - 62
SP - 16395
EP - 16406
JO - IEEE Transactions on Aerospace and Electronic Systems
JF - IEEE Transactions on Aerospace and Electronic Systems
ER -