TY - GEN
T1 - SOM Neural Network Based Gaussian Mixture PHD Algorithm for Multi-Sensor Multi-Target Tracking
AU - Zhang, Yujie
AU - Zheng, Hongwei
AU - Zhang, Zheng
AU - Yu, Jianglong
AU - Hua, Yongzhao
AU - Dong, Xiwang
AU - Ren, Zhang
N1 - Publisher Copyright:
© 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
PY - 2023
Y1 - 2023
N2 - In the multi-sensor multi-target tracking (MSMTT) problem, the matching of the measurements and the targets will generate a huge computational burden, resulting in an unsatisfactory real-time performance of the maneuvering targets tracking. In order to reduce the computational burden, this paper proposes a self-organizing feature map (SOM) neural network based Gaussian mixture probability hypothesis density algorithm (SOM-GMPHD). Firstly, a distributed filtering MSMTT algorithm based on SOM neural network is proposed. The distributed SOM-GMPHD algorithm (DSOM-GMPHD) has two fusion steps. Secondly, to further reduce the computational complexity, a centralized SOM-GMPHD algorithm (CSOM-GMPHD) with only one-step fusion is proposed. The computational complexity analysis of the existing MSMTT algorithms (DGMPHD and CGMPHD) and the proposed SOM-GMPHD algorithms are carried out in this paper. Finally, the effect of the proposed algorithms is evaluated in the simulation experiment.
AB - In the multi-sensor multi-target tracking (MSMTT) problem, the matching of the measurements and the targets will generate a huge computational burden, resulting in an unsatisfactory real-time performance of the maneuvering targets tracking. In order to reduce the computational burden, this paper proposes a self-organizing feature map (SOM) neural network based Gaussian mixture probability hypothesis density algorithm (SOM-GMPHD). Firstly, a distributed filtering MSMTT algorithm based on SOM neural network is proposed. The distributed SOM-GMPHD algorithm (DSOM-GMPHD) has two fusion steps. Secondly, to further reduce the computational complexity, a centralized SOM-GMPHD algorithm (CSOM-GMPHD) with only one-step fusion is proposed. The computational complexity analysis of the existing MSMTT algorithms (DGMPHD and CGMPHD) and the proposed SOM-GMPHD algorithms are carried out in this paper. Finally, the effect of the proposed algorithms is evaluated in the simulation experiment.
KW - Gaussian mixture PHD
KW - Multi-sensor multi-target tracking
KW - SOM neural network
UR - https://www.scopus.com/pages/publications/85151166713
U2 - 10.1007/978-981-19-6613-2_318
DO - 10.1007/978-981-19-6613-2_318
M3 - 会议稿件
AN - SCOPUS:85151166713
SN - 9789811966125
T3 - Lecture Notes in Electrical Engineering
SP - 3276
EP - 3285
BT - Advances in Guidance, Navigation and Control - Proceedings of 2022 International Conference on Guidance, Navigation and Control
A2 - Yan, Liang
A2 - Duan, Haibin
A2 - Deng, Yimin
A2 - Yan, Liang
PB - Springer Science and Business Media Deutschland GmbH
T2 - International Conference on Guidance, Navigation and Control, ICGNC 2022
Y2 - 5 August 2022 through 7 August 2022
ER -