基于Transformer的异类目标智能关联跟踪

doi:10.7527/S1000-6893.2024.31643

Abstract

Abstract:

Considering tracking different types of multiple targets， the current tracking system use fixed model parameters， which fail to fully exploit motion characteristics from different types of targets. This limitation can lead to decreased tracking accuracy and even track loss or mis-tracking. To address this issue， a Transformer-based intelligent data association and tracking method is proposed for different types of multiple targets. In the case of time sliding window， by integrating sparse attention and self-attention mechanisms， the target-type features and track association features are deeply explored and exploited from the historical radar data and short-time track data， to adaptively adjust the process noise covariance matrix. This is beneficial to realize the refined motion modeling of different types of targets and acquire the reliable association between every radar echo and each target track， thereby producing high-quality tracking outputs for different types of multiple targets. Simulation results in the typical scenario of tracking different types of multiple targets with varying clutter densities demonstrate that， compared with the joint probabilistic data association filter， multiple hypothesis tracking， probability hypothesis density filter， and intelligent data association tracking with fixed model parameters， the proposed method achieves superior tracking accuracy and stability.

Key words: different types of multiple targets tracking, data association, short-time track classification, process noise covariance adjustment, Transformer network

CLC Number:

V243.2

Yao LU, Zihao LI, Zhunga LIU, Yanbo YANG. Intelligent data association and tracking of different types of multiple targets based on Transformer[J]. Acta Aeronautica et Astronautica Sinica, 2025, 46(17): 331643.

Figures/Tables 16

Fig.1

Table 1

List of initial values of different types of multiple targets

参数	目标1	目标2	目标3	目标4	目标5
$x$ /m	5 000	5 000	5 000	5 000	5 000
$x ˙$ /（m·s^-1）	-14	-14	-14	-14	-14
$y$ /m	2 900	2 800	2 600	2 650	2 850
$y ˙$ /（m·s^-1）	-17	12	0	15	15
转弯率	0	0	0	0.17	0.12

Table 1

Fig.2

Table 2

Table 3

Fig.3

Fig.4

Fig.5

Fig.6

Fig.7

Table 4

Comparison of average running times of different methods based on 50 Monte Carlo realizations

算法	运行时间/s
算法	$λ$ = $10 - 5$	$λ$ = $10 - 4$	$λ$ = $10 - 3$
JPDAF	0.64	1.30	17.52
MHT	4.93	1.62	7.98
本文算法	5.10	6.09	8.43

Table 4

Fig.8

Table 5

Fig.9

Fig.10

Table 6

References 31

[1]	刘雷，刘大卫，王晓光，等. 无人机集群与反无人机集群发展现状及展望［J］. 航空学报， 2022， 43（）： 726908.
	LIU L， LIU D W， WANG X G， et al. Development status and prospect of UAV cluster and anti-UAV cluster［J］. Acta Aeronautica et Astronautica Sinica， 2022， 43（Sup 1）： 726908 （in Chinese）.
[2]	王传云，苏阳，王琳霖，等. 面向反制无人机集群的多目标连续鲁棒跟踪算法［J］. 航空学报， 2024， 45（7）： 329017.
	WANG C Y， SU Y， WANG L L， et al. Multi-object continuous robust tracking algorithm for anti-UAV swarm［J］. Acta Aeronautica et Astronautica Sinica， 2024， 45（7）： 329017 （in Chinese）.
[3]	褚昭晨，宋韬，金忍，等. 基于视觉图像的空对空多无人机目标跟踪［J］. 航空学报， 2024， 45（14）： 629379.
	CHU Z C， SONG T， JIN R， et al. Vision-based air-to-air multi-UAVs tracking［J］. Acta Aeronautica et Astronautica Sinica， 2024， 45（14）： 629379 （in Chinese）.
[4]	柳超，王月基. 对海探测雷达多目标跟踪技术综述［J］. 雷达学报， 2021， 10（1）： 100-115.
	LIU C， WANG Y J. Review of multi-target tracking technology for marine radar［J］. Journal of Radars， 2021， 10（1）： 100-115 （in Chinese）.
[5]	RANGESH A， TRIVEDI M M. No blind spots： Full-surround multi-object tracking for autonomous vehicles using cameras and LiDARs［J］. IEEE Transactions on Intelligent Vehicles， 2019， 4（4）： 588-599.
[6]	ZHOU M， ZHANG J J， PAPANDREOU-SUPPAPPOLA A. Multiple target tracking in urban environments［J］. IEEE Transactions on Signal Processing， 2016， 64（5）： 1270-1279.
[7]	FORTMANN T， BAR-SHALOM Y， SCHEFFE M. Sonar tracking of multiple targets using joint probabilistic data association［J］. IEEE Journal of Oceanic Engineering， 1983， 8（3）： 173-184.
[8]	王树亮，毕大平，阮怀林. 认知雷达波形自适应数据关联跟踪算法［J］. 宇航学报， 2017， 38（12）： 1331-1338.
	WANG S L， BI D P， RUAN H L. Waveform self-adaption data association algorithm for cognitive radar tracking［J］. Journal of Astronautics， 2017， 38（12）： 1331-1338 （in Chinese）.
[9]	CHONG C Y. An overview of machine learning methods for multiple target tracking［C］∥ 2021 IEEE 24th International Conference on Information Fusion （FUSION）. Piscataway： IEEE Press， 2021.
[10]	LI X R. Tracking in clutter with strongest neighbor measurements［J］. IEEE Transactions on Automatic Control， 1998， 43（11）： 1560-1578.
[11]	SONG T L， LEE D G. A probabilistic nearest neighbor filter algorithm for m validated measurements［J］. IEEE Transactions on Signal Processing， 2006， 54（7）： 2797-2802.
[12]	何友，修建娟，刘瑜，等. 雷达数据处理及应用［M］. 4版. 北京：电子工业出版社， 2022： 138.
	HE Y， XIU J J， LIU Y， et al. Radar data processing with applications［M］. 4th ed. Beijing： Publishing House of Electronics Industry， 2022： 138 （in Chinese）.
[13]	WILLIAMS J L. Marginal multi-bernoulli filters： RFS derivation of MHT， JIPDA， and association-based member［J］. IEEE Transactions on Aerospace and Electronic Systems， 2015， 51（3）： 1664-1687.
[14]	WILLETT P， RUAN Y， STREIT R. PMHT： Problems and some solutions［J］. IEEE Transactions on Aerospace and Electronic Systems， 2002， 38（3）： 738-754.
[15]	SONG T L， LEE D G， RYU J. A probabilistic nearest neighbor filter algorithm for tracking in a clutter environment［J］. Signal Processing， 2005， 85（10）： 2044-2053.
[16]	BAR-SHALOM Y， TSE E. Tracking in a cluttered environment with probabilistic data association［J］. Automatica， 1975， 11（5）： 451-460.
[17]	HE S M， SHIN H S， TSOURDOS A. Distributed joint probabilistic data association filter with hybrid fusion strategy［J］. IEEE Transactions on Instrumentation and Measurement， 2020， 69（1）： 286-300.
[18]	QIAN H， YANG X， ZHANG X， et al. PLA-JPDA for indoor multi-person tracking using IR-UWB radars［C］∥IEEE Radar Conference. Piscataway： IEEE Press， 2020.
[19]	杨峰，王永齐，梁彦，等. 基于概率假设密度滤波方法的多目标跟踪技术综述［J］. 自动化学报， 2013， 39（11）： 1944-1956.
	YANG F， WANG Y Q， LIANG Y， et al. A survey of PHD filter based multi-target tracking［J］. Acta Automatica Sinica， 2013， 39（11）： 1944-1956 （in Chinese）.
[20]	申屠晗，薛安克，周治利. 多传感器高斯混合PHD融合多目标跟踪方法［J］. 自动化学报， 2017， 43（6）： 1028-1037.
	SHENTU H， XUE A K， ZHOU Z L. Multi-sensor Gaussian mixture PHD fusion for multi-target tracking［J］. Acta Automatica Sinica， 2017， 43（6）： 1028-1037 （in Chinese）.
[21]	徐洋，徐晖，安玮，等. 一种利用信号幅度信息的光学传感器像平面GM-PHD滤波算法［J］. 宇航学报， 2012， 33（4）： 478-486.
	XU Y， XU H， AN W， et al. A GM-PHD filter using signal amplitude information for image plane of optical sensor［J］. Journal of Astronautics， 2012， 33（4）： 478-486 （in Chinese）.
[22]	SONNTAG V， LE CAILLEC J M， PERES A， et al. Transformer-based state estimation for tracking： Maneuvering target and multi-target capabilities［C］∥2024 IEEE Radar Conference （RadarConf24）. Piscataway： IEEE Press， 2024.
[23]	ZHANG Y S， LI G， ZHANG X P， et al. A deep learning model based on transformer structure for radar tracking of maneuvering targets［J］. Information Fusion， 2024， 103： 102120.
[24]	CHUNG Y N， CHOU P H， YANG M R， et al. Multiple-target tracking with competitive hopfield neural network based data association［J］. IEEE Transactions on Aerospace and Electronic Systems， 2007， 43（3）： 1180-1188.
[25]	LIU H J， ZHANG H， MERTZ C. DeepDA： LSTM-based deep data association network for multi-targets tracking in clutter［C］∥2019 22th International Conference on Information Fusion （FUSION）. Piscataway： IEEE Press， 2019.
[26]	LI W N， YANG A L， ZHANG L Z. Improved data association algorithm for airborne radar multi-target tracking via deep learning network［C］∥IGARSS 2022-2022 IEEE International Geoscience and Remote Sensing Symposium. Piscataway： IEEE Press， 2022.
[27]	GAO C， LIU H W， VARSHNEY P K， et al. Signal structure information-based data association for maneuvering targets with a convolutional Siamese network［C］∥2021 CIE International Conference on Radar （Radar）. Piscataway： IEEE Press， 2021.
[28]	MEINHARDT T， KIRILLOV A， LEAL-TAIXÉ L， et al. TrackFormer： Multi-object tracking with transformers［C］∥2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2022.
[29]	李文娜，张顺生，王文钦. 基于Transformer网络的机载雷达多目标跟踪方法［J］. 雷达学报， 2022， 11（3）： 469-478.
	LI W N， ZHANG S S， WANG W Q. Multitarget-tracking method for airborne radar based on a transformer network［J］. Journal of Radars， 2022， 11（3）： 469-478 （in Chinese）.
[30]	PINTO J， HESS G， LJUNGBERGH W， et al. Next generation multitarget trackers： Random finite set methods vs transformer-based deep learning［C］∥2021 IEEE 24th International Conference on Information Fusion （FUSION）. Piscataway： IEEE Press， 2021.
[31]	WEI X W， LIN Y R， ZHANG L N， et al. Transformer-based multi-target tracking with Bayesian perspective［C］∥2024 27th International Conference on Information Fusion （FUSION）. Piscataway： IEEE Press， 2024： 1-7.

参数	数值
输入维度	4
输出维度	2
层数	3
注意力头数	8
隐藏层维度	64
Dropout率	0.1

参数	数值
输入维度	滑窗大小×目标数×量测数
输出维度	滑窗大小×目标数×量测数
层数	5
注意力头数	5
Dropout率	0.1

类型	平均OSPA/m
2层	16.40
2层+稀疏注意力	16.61
5层	15.11
5层+稀疏注意力	14.23
10层	15.79
10层+稀疏注意力	15.29
PHD滤波	75.59

类型	运行时间/s
2层	7.54
2层+稀疏注意力	7.21
5层	8.72
5层+稀疏注意力	8.51
10层	10.80
10层+稀疏注意力	10.58
PHD滤波	74.91

Intelligent data association and tracking of different types of multiple targets based on Transformer

RichHTML

PDF (PC)

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 16

References 31

Related Articles 9

Recommended Articles

Metrics

Comments

[1]	Zhaochen CHU, Tao SONG, Ren JIN, Defu LIN. Vision-based air-to-air multi-UAVs tracking [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(14): 629379-629379.
[2]	ZHANG Tianyu, ZHENG Jian, TIAN Zhuoer, RONG Yingjiao, GUO Yunfei, SHENTU Han. Target existence probability based multi-radar track-to-track fusion algorithm [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2019, 40(8): 322848-322848.
[3]	WANG Shuliang, BI Daping, RUAN Huailin, ZHOU Yang. Cognitive radar data association algorithm based on visual attention [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2018, 39(6): 321828-321828.
[4]	WANG Cong, WANG Haipeng, XIONG Wei, HE You. Data association algorithm based on least square fitting [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2016, 37(5): 1603-1613.
[5]	Fan Bingyi;Li Jianxun;Liu Tan. Target Oriented Probabilistic Multi-hypothesis Tracker Algorithm [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2010, 31(12): 2373-2378.
[6]	Zhu Ziqian;. A Universal Track Initiation Model [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2009, 30(3): 497-504.
[7]	WANG Jie-gui;LUO Jing-qing. Data Association Algorithm Based on Fuzzy Synthetic Evaluation of Multiple Features in Multi-Target Passive Tracking [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2004, 25(2): 172-175.
[8]	FAN Tao;YANG Chen-yang;LI Shao-hong. SPATIAL DOMAIN MULTI-RESOLUTION DIM TARGET TRACKING [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2001, 22(S1): 75-79.
[9]	Wang Peide;Zhou Hongren;Jing Zhongliang. RESIDUAL FILTER AND ITS APPLICATION TO TRACKING MULTIPLE MANEUVERING TARGETS IN DENSE MULTI-RETURN ENVIRONMENT [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 1988, 9(5): 281-289.