Please wait a minute...
浙江大学学报(工学版)
浙江大学学报(工学版)  2022, Vol. 56 Issue (5): 967-976    DOI: 10.3785/j.issn.1008-973X.2022.05.014
计算机与控制工程     
重尾非高斯定位噪声下鲁棒协同目标跟踪
陈小波(),陈玲,梁书荣,胡煜
江苏大学 汽车工程研究院，江苏 镇江 212013
Robust cooperative target tracking under heavy-tailed non-Gaussian localization noise
Xiao-bo CHEN(),Ling CHEN,Shu-rong LIANG,Yu HU
Automotive Engineering Research Institute, Jiangsu University, Zhenjiang 212013, China
 全文: PDF(2293 KB)   HTML
摘要：

针对定位数据的统计特性未知且易受异常值干扰而影响协同目标跟踪性能的问题，提出一种重尾非高斯定位噪声下的鲁棒协同目标跟踪方法. 该方法假设定位噪声服从多元学生t-分布，建立联合估计目标状态与定位噪声参数的贝叶斯模型. 针对目标状态与噪声分布参数相互耦合而难以计算联合后验分布的问题，应用变分贝叶斯推断原理和平均场理论对后验分布进行解耦，将目标状态与定位噪声参数的联合后验分布估计问题转化为最优化问题，以交替优化的方式实现系统参数的在线递推估计. 对提出的协同目标跟踪方法进行测试. 仿真结果表明，当定位数据中存在未知的野值噪声时，提出的协同跟踪算法具有较好的鲁棒性.
关键词： 协同目标跟踪;  学生t-分布;  变分贝叶斯推断;  野值噪声;  鲁棒性    
Abstract:

The statistical properties of the localization data are unknown and the localization data are susceptible to outlier interference, which affects the cooperative target tracking performance. Aiming at the problem, a robust cooperative target tracking method for heavy-tailed non-Gaussian localization noise was proposed. It was assumed that the localization noise followed multivariate Student’s t-distribution and a joint Bayesian estimation model of target state and localization noise parameters was constructed. To overcome the difficulty of computing joint posterior distribution owing to the coupling of target state and noise distribution parameters, the variational Bayesian inference and the mean-field theory were applied to decouple the joint posterior distribution and convert the problem of joint estimation of target state and localization noise parameters into a optimization problem. Alternative optimization was implemented to achieve recursive estimation of system parameters. The proposed cooperative target tracking method was evaluated experimentally. Simulation results showed that when unknown outliers exist in localization data, the proposed algorithm has better tracking robustness.
Key words: cooperative target tracking    Student’s t-distribution    variational Bayesian inference    outlier noise    robustness
收稿日期: 2021-06-05 出版日期: 2022-05-31
CLC:  TP 301.6  
基金资助: 国家自然科学基金资助项目(61773184)；国家重点研发计划资助项目(2018YFB0105000)；江苏省六大人才高峰高层次人才资助项目(JXQC-007)
作者简介: 陈小波（1982—），男，研究员，博导，从事智能交通研究. orcid.org/0000-0001-9940-1637. E-mail: xbchen82@gmail.com
服务　　
把本文推荐给朋友
加入引用管理器
E-mail Alert
作者相关文章　　
陈小波
陈玲
梁书荣
胡煜

引用本文:

陈小波,陈玲,梁书荣,胡煜. 重尾非高斯定位噪声下鲁棒协同目标跟踪[J]. 浙江大学学报(工学版), 2022, 56(5): 967-976.

Xiao-bo CHEN,Ling CHEN,Shu-rong LIANG,Yu HU. Robust cooperative target tracking under heavy-tailed non-Gaussian localization noise. Journal of ZheJiang University (Engineering Science), 2022, 56(5): 967-976.

链接本文:

https://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2022.05.014    或    https://www.zjujournals.com/eng/CN/Y2022/V56/I5/967

图 1  不同 $ \lambda $值下学生t-分布的概率密度函数
图 2  平移与旋转坐标转换示意图
图 3  协同目标跟踪的概率图模型
图 4  协同目标跟踪示意图
图 5  不同在线跟踪算法的误差对比
算法 ARMSE
$ {\varphi _1} $/m $ {\chi _1} $/% $ {\varphi _2} $/(m·s?1) $ {\chi _2} $/%
KF-ST 0.185 01 ? 0.097 41 ?
EKF-CT 0.171 99 7.0 0.093 06 4.5
VB-CT 0.160 40 13.3 0.087 43 10.2
VB-RCT 0.152 80 17.4 0.083 61 14.2
表 1  不同算法的跟踪误差性能分析
图 6  不同野值比率下的位置ARMSE
图 7  不同异常程度下，CV的非高斯分布定位噪声示意图
图 8  从不同分布采样野值数据时，不同算法的跟踪误差
分布 异常程度 pARMSE/m
KF EKF-CT VB-CT VB-RCT
$S\left( { {\boldsymbol{v} }_k^2;{\boldsymbol{0}},{\boldsymbol A},15} \right)$ $ {\boldsymbol A} = 15{\boldsymbol{R}} $ 0.185 01 0.166 35 0.151 25 0.153 42
$ {\boldsymbol A} = 25{\boldsymbol{R}} $ 0.185 01 0.171 99 0.160 40 0.152 80
$ {\boldsymbol A} = 35{\boldsymbol{R}} $ 0.185 01 0.175 22 0.169 33 0.152 39
$U\left( { {\boldsymbol{v} }_k^2; - {\boldsymbol{a}},{\boldsymbol{a}}} \right)$ ${\boldsymbol{a}} = 15{\boldsymbol{R} }$ 0.187 60 0.167 49 0.148 83 0.155 28
${\boldsymbol{a}} = 30{\boldsymbol{R} }$ 0.187 60 0.174 86 0.158 82 0.154 78
${\boldsymbol{a}} = 45{\boldsymbol{R} }$ 0.187 60 0.178 31 0.164 19 0.154 12
$L\left( { {\boldsymbol{v} }_k^2;{\boldsymbol{0}},{\boldsymbol{b}}} \right)$ ${\boldsymbol{ b}} = 10{\boldsymbol{R} }$ 0.187 60 0.163 32 0.153 40 0.155 02
${\boldsymbol{b}} = 15{\boldsymbol{R} }$ 0.187 60 0.168 40 0.156 86 0.154 72
${\boldsymbol{b}} = 20{\boldsymbol{R} }$ 0.187 60 0.171 82 0.160 33 0.154 44
表 2  从不同分布采样野值数据时，不同算法的ARMSE
图 9  不同遗忘因子下的位置ARMSE
图 10  跟踪误差随变分迭代次数的变化
1 裴汉林 基于深度学习的自动驾驶环境感知技术研究[J]. 内燃机与配件, 2021, (4): 193- 194
PEI Han-lin Research on autonomous driving environment perception technology based on Deep Learning[J]. Internal Combustion Engine and Parts, 2021, (4): 193- 194
doi: 10.3969/j.issn.1674-957X.2021.04.089
2 RANDESH 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
doi: 10.1109/TIV.2019.2938110
3 DHAMANAM N, KATHIRVELU M, RAO T G Review of environment perception for intelligent vehicles[J]. International Journal of Engineering and Management Research, 2019, 9 (2): 13- 17
4 石章松, 刘志坤, 吴中红. 目标定位跟踪方法与实践[M]. 北京: 电子工业出版社, 2019.
5 SAWARAGI Y, KATAYAMA T Performance loss and design method of Kalman filters for discrete-time linear systems with uncertainties[J]. International Journal of Control, 1970, 12 (1): 163- 172
doi: 10.1080/00207177008931830
6 NATTAPOL A, KUNRUTAI P, BENIAWAN T, et al A novel adaptive resampling for sequential Bayesian filtering to improve frequency estimation of time-varying signals[J]. Heliyon, 2021, 7 (4): e06768
doi: 10.1016/j.heliyon.2021.e06768
7 BEAL M J. Variational algorithms for approximate Bayesian inference [D]. London: University College London, 2003.
8 SARKKA S, NUMMENMAA A. Recursive noise adaptive Kalman filtering by variational approximations [J]. IEEE Transactions on Automatic Control. 2009, 54(3): 596-600.
9 沈忱, 徐定杰, 沈锋, 等 基于变分推断的一般噪声自适应卡尔曼滤波[J]. 系统工程与电子技术, 2014, 36 (8): 1466- 1472
SHEN Chen, XU Ding-jie, SHEN Feng, et al Generalized noises adaptive Kalman filtering based on variational inference[J]. Systems Engineering and Electronics, 2014, 36 (8): 1466- 1472
doi: 10.3969/j.issn.1001-506X.2014.08.03
10 余东平, 何谢, 齐扬阳, 等 基于变分贝叶斯推理的多目标无源定位算法[J]. 南京邮电大学学报:自然科学版, 2018, 38 (2): 48- 53,59
YU Dong-ping, HE Xie, QI Yang-yang, et al Variational Bayesian inference based multi-target device-free localization algorithm[J]. Journal of Nanjing University of Posts and Telecommunications: Science and Technology, 2018, 38 (2): 48- 53,59
11 ZHU H, LEUNG H, HE Z A variational Bayesian approach to robust sensor fusion based on Student-t distribution[J]. Information Sciences, 2013, 221: 201- 214
doi: 10.1016/j.ins.2012.09.017
12 HUANG Y L, ZHANG Y G, LI N, et al A novel robust student's t-based Kalman filter[J]. IEEE Transactions on Aerospace and Electronic Systems, 2017, 53 (3): 1545- 1554
doi: 10.1109/TAES.2017.2651684
13 喻尚, 杨艳, 张正轩, 等 车载无线通信技术浅析[J]. 汽车电器, 2021, (3): 46- 47,50
YU Shang, YANG Yan, ZHANG Zheng-xuan, et al Discussion on vehicle wireless communication technology[J]. Autoelectric Parts, 2021, (3): 46- 47,50
doi: 10.3969/j.issn.1003-8639.2021.03.017
14 KEVVAN A Joint use of DSRC and C-V2X for V2X communications in the 5.9 GHz ITS band[J]. IET Intelligent Transport Systems, 2020, 15 (2): 213- 224
15 HUANG C, HU P, LIAN J Online optimum velocity calculation under V2X for smart new energy vehicles[J]. Transactions of the Institute of Measurement and Control, 2021, 43 (10): 2368- 2377
doi: 10.1177/0142331221997280
16 KIRILL K, ALEXANDER M, BORIS M Robust data fusion of UAV navigation measurements with application to the landing system[J]. Remote Sensing, 2020, 12 (23): 3849
doi: 10.3390/rs12233849
17 陈宇波, 宋迎春 非高斯噪声下的车载GPS信号定位算法[J]. 中南大学学报:自然科学版, 2010, 41 (4): 1462- 1466
CHEN Yu-bo, SONG Ying-chun A Bayes filter algorithm with non-Gaussian noises based on location of vehicular GPS[J]. Journal of Central South University: Science and Technology, 2010, 41 (4): 1462- 1466
18 胡淼淼, 敬忠良, 董鹏, 等 基于T分布变分贝叶斯滤波的SINS/GPS组合导航[J]. 浙江大学学报:工学版, 2018, 52 (8): 1482- 1488
HU Miao-miao, JING Zhong-liang, DONG Peng, et al Variational Bayesian filtering based on Student-t distribution for SINS/GPS integrated navigation[J]. Journal of Zhejiang University: Engineering Science, 2018, 52 (8): 1482- 1488
19 BISHOP C M. Pattern recognition and machine learning[M]. Cambridge: Springer, 2006.
20 XIAO G C, SONG X L, CAO H T, et al Augmented extended Kalman filter with cooperative Bayesian filtering and multi-models fusion for precise vehicle localisations[J]. IET Radar, Sonar and Navigation, 2020, 14 (11): 1815- 1826
doi: 10.1049/iet-rsn.2020.0155
[1] 沈麒,赵琰,周晓炜,袁晓冉. 结合结构与梯度的图像哈希算法[J]. 浙江大学学报(工学版), 2020, 54(8): 1525-1533.
[2] 隋昊,覃高峰,崔祥波,陆新江. 基于误差均值与方差最小化的鲁棒T-S模糊建模方法[J]. 浙江大学学报(工学版), 2019, 53(2): 382-387.
[3] 杨春宁, 方家为, 李春, 葛晖. 基于稳定性判据的高超声速复合控制方法[J]. 浙江大学学报(工学版), 2017, 51(2): 422-428.
[4] 李滔, 王士同. 增量式0阶TSK模糊分类器及鲁棒改进[J]. 浙江大学学报(工学版), 2017, 51(10): 1901-1911.
[5] 杨慧琳, 黄智刚, 刘久文, 杜元锋. 基于核模糊C均值指纹库管理的WIFI室内定位方法[J]. 浙江大学学报(工学版), 2016, 50(6): 1126-1133.
[6] 赵婵媛, 陆志强, 崔维伟. 考虑随机故障的流水线调度问题前摄优化方法[J]. 浙江大学学报(工学版), 2016, 50(4): 641-649.
[7] 朱光明, 蒋荣欣, 周凡, 田翔, 陈耀武. 带测量偏置估计的鲁棒卡尔曼滤波算法[J]. 浙江大学学报(工学版), 2015, 49(7): 1343-1349.
[8] 马腾, 赵兴忠, 高博青, 吴慧. 自由曲面形状和拓扑联合优化研究[J]. 浙江大学学报(工学版), 2015, 49(10): 1946-1951.
[9] 张成,李志安,高博青,董石麟. 基于H∞理论的网壳结构鲁棒性分析[J]. J4, 2013, 47(5): 818-823.
[10] 单艳玲, 高博青. 基于连续体拓扑优化的网壳结构鲁棒构型分析[J]. J4, 2013, 47(12): 2118-2124.
[11] 王秀君, 胡协和. 一种改进的单神经元PID控制策略[J]. J4, 2011, 45(8): 1498-1501.
[12] 金良, 王维锐, 刘哲, 石浩然. 集成干扰观测器的BP-PID在炉温控制上的应用[J]. J4, 2011, 45(10): 1761-1765.
[13] 茅僰 应祖光 朱位秋. 不确定结构非线性随机最优控制的鲁棒性[J]. J4, 2008, 42(7): 1101-1105.