Page 157 - 201903

第 38 卷 第 3 期 李嶷等： 低信噪比条件下多节点声呐目标跟踪算法 439

得到目标轨迹；如果检测门限设置过低，那么单节点 [3] Coraluppi S, Carthel C. Multi-hypothesis sonar track-
声呐可能得到大量目标轨迹虚警预报。但是，采用 ing[C]. Proceedings of the 7th International Conference
on Information Fusion, Stockholm, Sweden, 2004.
本文的算法，对信噪比进行置信度等级划分，并对多
[4] Balasingam B, Baum M, Willett P. Random finite set par-
节点数据进行表决融合后，可得到目标轨迹，且轨迹 ticle filter for source enumeration and direction-of-arrival
拟合效果较好，轨迹拟合误差小。 tracking using sonar arrays[C]. 2017 20th International
Conference on Information Fusion, Xi’an, China, 2017:
1.988
1–6.
1.986 [5] Coraluppi S, Carthel C. Distributed tracking in multi-
1.984 static sonar[J]. IEEE Transactions on Aerospace and Elec-
tronic Systems, 2005, 41(3): 1138–1147.
y/10 6 m 1.980 B [6] Munafò A, Furfaro T, Ferri G, et al. Supporting AUV
1.982
A
localisation through next generation underwater acoustic
1.978 C networks: results from the field[C]. IEEE/RSJ Interna-
tional Conference on Intelligent Robots and Systems 2016,
1.976
Daejeon, South Korea, 2016: 1328–1333.
1.974
1.35 1.40 1.45 1.50 1.55 1.60 [7] Coraluppi S, Carthel C. Performance limits of real-time
x/10 m contact-based tracking[C]. IEEE OCEANS 2007, Ab-
5
erdeen, UK, June 2007: 1–6.
图 7 目标拟合轨迹示意图
[8] Lepage K D, Goldhahn R, Alves J, et al. Autonomous
Fig. 7 Fitting tracks of target
networked anti-submarine warfare research and develop-
ment at CMRE[C]. IEEE OCEANS 2015, Genova, Italy,
3 结论 May 2015: 1–6.
[9] Blanding W, Willett P, Coraluppi S. Sequential ML for
本文采用分级滤波方法，首先根据目标的运动 multistatic sonar tracking[C]. IEEE OCEANS 2007, Ab-
特性分别对各节点的亮点数据进行轨迹判定，剔除 erdeen, UK, 2007: 1–6.
[10] Grimmett D, Coraluppi S. Contact-level multistatic sonar
孤立的野点；然后利用多节点协同探测优势，在考虑
data simulator for tracker performance assessment[C].
各段轨迹置信度高低的情况下，采用多节点投票表 IEEE 2006 9th International Conference on Information
决方法对轨迹进行判断，保留综合性能较优的轨迹 Fusion, Florence, Italy, 2006: 1–7.
段。算法通过单节点声呐到多节点声呐的逐级滤波 [11] Braca P, Goldhahn R, Ferri G, et al. Distribution infor-
mation fusion in multistatic sensor networks for under-
处理，最终剔除了大部分亮点数据，有效减少了虚警
water surveillance[J]. IEEE Sensor Journal, 2016, 16(11):
概率。 4003–4014.
本文采用的 “结合置信度水平的表决融合” 算 [12] Papa G, Braca P, Horn S, et al. Adaptive Bayesian
法既考虑了原始回波信号的信噪比，又考虑了轨迹 tracking with unknown time-varying sensor network per-
formance[C]. IEEE 2006 9th International Conference on
的连续性和置信度水平，最后还从多节点声呐决策
Acoustic Speech Signal Process, Brisbane, QLD, Aus-
的角度进行了分析。从文中可看出，在低信噪比条 tralia, 2015: 2534–2538.
件下，单节点声呐对目标进行定位跟踪时的误差大， [13] Sildam J, Lepage K D. Ambiguity reduction of under-
轨迹断续，很难得到目标的真实轨迹。但是利用多 water targets in framework of topic modeling[C]. IEEE
2005 18th International Conference on Information Fu-
个声呐节点进行探测，并采用本文提出的数据融合
sion, Washington, DC, USA, 2015: 2017–2024.
算法后，可以充分利用各节点的探测优势，提高检测 [14] Ferri G, Munafò A, Alves J, et al. A data-driven con-
概率和目标定位跟踪精度。 trol strategy in synergy with continuous active sonar for
littoral underwater surveillance[C]. IEEE OCEANS 2016
参 考 文 献 MTS/IEEE Monterey, Monterey, CA, USA, 2016: 1–7.
[15] Li Y, Chen X H, Sun C Y. A tracking method based on
[1] Meyer F, Braca P, Willett P, et al. A scalable algorithm particle filter for multistatic sonar system[C]. IEEE Inter-
for tracking an unknown number of targets using multiple national Conference on Signal Processing, Communication
sensors[J]. IEEE Transactons on Signal Processing, 2017, & Computing, Guilin, China, 2014: 162–165.
65(13): 3478–3493. [16] Yang L, Yang L, Ho K C. Moving target localization in
[2] Coraluppi S, Grimmett D. Multistatic sonar tracking[C]. multistatic sonar using time delays, Doppler shifts and ar-
Proceedings of the SPIE Conference on Signal Processing, rival angles[C]. IEEE International Conference on Acous-
Sensor Fusion, and Target Recognition XII, Orlando FL, tics, Speech and Signal Processing, New Orleans, LA,
USA, 2003, 5096: 399–410. USA, 2016: 3399–3403.