交通监测毫米波雷达作为智能交通系统的重要组成部分，能够准确地提供道路车辆的位置、速度等信息，雷达数据经多目标跟踪处理后可实时获取车辆运动轨迹，用以实现路段流量统计等功能。但因交通场景的复杂性与车辆目标分布的密集性，传统多目标跟踪算法运用在交通监测场景时会出现跟踪稳定性差、多目标航迹合并等问题，严重影响车流量统计精度。因此，开展交通监测毫米波雷达多目标跟踪算法研究具有十分重要的意义。

针对交通监测场景中车辆目标较多且位置分布密集，导致跟踪过程中存在漏跟踪与误跟踪的问题，提出一种改进的交通监测毫米波雷达多目标跟踪算法。通过分析交通场景下雷达量测点迹分布特征，利用多帧数据积累的方法提高车辆目标点迹密度并孤立噪声产生的虚假点迹；基于欧式距离度量点间距离，并结合雷达多维参数测量精度差别，对高精度特征参数进行加权处理以提高密集间隔目标的类间距离；利用多项式函数对点间距离分布进行拟合，并依据函数凹凸性求拟合函数极值点，实现DBSCAN算法参数的自适应求解，进而利用聚类算法实现多目标点迹的凝聚处理；基于极大后验估计原理，在滤波算法中引入Sage-Husa噪声估计器对量测噪声进行时序更新，降低量测噪声对目标跟踪精度的影响；依据点迹凝聚处理得到的多帧点迹类别信息与航迹质量建立关联准则，同批次数据关联时将同类点迹与相应航迹进行关联，各批次数据间关联时构建雷达量测数据与目标航迹外推点间统计距离矩阵，优先关联高质量航迹，实现车辆目标的稳定跟踪。

基于改进的多目标跟踪算法与毫米波雷达硬件平台，对实际道路交通内车辆目标进行跟踪系统测试，并依据车辆跟踪轨迹数据计算路段车流量信息，平均流量统计精度可达95%。测试结果表明，改进算法在实际道路交通场景下对车辆目标具有较稳定的跟踪效果，对交通信息智能化采集领域应用具有一定的参考价值。

As an important part of intelligent transportation system, traffic monitoring millimeter-wave radar can accurately provide the location, speed and other information of road vehicles. After multi-target tracking processing, the radar data can obtain the vehicle movement track in real time, which can be used to realize traffic flow statistics and other functions. Due to the complexity of traffic scenes and the high-density of vehicle target distribution, the traditional multi-target tracking algorithm in traffic monitoring scenes will have poor tracking stability, multi-target track combination and other problems, which seriously affect the accuracy of vehicle flow statistics. Therefore, it is of great significance to research the multi-target tracking algorithm of traffic monitoring millimeter wave radar.

An improved multi-target tracking algorithm for millimeter-wave radar in traffic monitoring was proposed to solve the problem of missing track and false track in the tracking process due to the large number of vehicle targets and dense location distribution in traffic monitoring scene. By analyzing the distribution characteristics of radar measurement plots in traffic scenes, the multi-frame data accumulation method is used to improve the density of vehicle target plots and isolate the false plots generated by noise. Based on the Euclidean distance measurement method and combined with the difference of radar multi-dimensional parameters measurement accuracy, the high-precision characteristic parameters are weighted to improve the interclass distance of densely spaced targets. The polynomial function was used to fit the distance distribution between plots, and the extreme point of the fitting function was obtained according to the concavity and convexity of the function, so as to realize the adaptive solution of the parameters of DBSCAN algorithm, and then the clustering algorithm was used to realize the multi-object plots centroid processing. Based on the principle of maximum posteriori estimation, the Sage-Husa noise estimator was introduced into the filtering algorithm to update the measured noise and reduce the influence of the measured noise on the target tracking accuracy. Based on data category from plots centroid and track quality information to establish correlation rule, in the same batch data associated with the category of plots between tracks, the statistical distance matrix between the radar measurement data and the target track extrapolation point is constructed when the data of each batch is correlated, and the high-quality track is associated with priority to achieve the stable tracking of vehicles.

Based on the improved multi-target tracking algorithm and the millimeter wave radar hardware platform, the tracking system is tested on multiple batches of vehicle targets within actual road traffic, and the average statistical accuracy of traffic flow reaches 95%. The test results show that the proposed algorithm has good tracking effect and traffic flow counting accuracy for vehicle targets in real road traffic scenarios, which is of reference value for the application in the field of intelligent traffic information collection.

[1] Kaffash S, An Truong N, Zhu J. Big data algorithms and applications in intelligent transportation system: A review and bibliometric analysis [J]. International Journal of Production Economics, 2021, 231.

[2] Guan K, He D P, et al. 5-GHz Obstructed Vehicle-to-Vehicle Channel Characterization for Internet of Intelligent Vehicles [J]. IEEE Internet of Things Journal, 2019, 6(1): 100-110.

[3] Emami P, Elefteriadou L, Ranka S. Long-Range Multi-Object Tracking at Traffic Intersections on Low-Power Devices [J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23(3): 2482-2493.

[4] Guerrero-Ibanez J, Zeadally S, Contreras-Castillo J. Sensor Technologies for Intelligent Transportation Systems [J]. Sensors, 2018, 18(4).

[5] Wang K, Xiong H, Zhang J, et al. SenseMag: Enabling Low-Cost Traffic Monitoring Using Noninvasive Magnetic Sensing [J]. IEEE Internet of Things Journal, 2021, 8(22): 16666-16679.

[6] Lamas-Seco J J, Castro P M, Dapena A, et al. Multi-loop inductive sensor model for vehicle traffic applications [J]. Sensors and Actuators a-Physical, 2017, 263: 580-592.

[7] Husain A A, Maity T, Yadav R K. Vehicle detection in intelligent transport system under a hazy environment: a survey [J]. IET Image Processing, 2020, 14(1): 1-10.

[8] Gao Z, Zhai R F, Wang P F, et al. Synergizing Appearance and Motion With Low Rank Representation for Vehicle Counting and Traffic Flow Analysis [J]. IEEE Transactions on Intelligent Transportation Systems, 2018, 19(8): 2675-2685.

[9] Chen Y, Zhao D, Er M J, et al. A novel vehicle tracking and speed estimation with varying UAV altitude and video resolution [J]. International Journal of Remote Sensing, 2021, 42(12): 4441-4466.

[10] Liu X, He R. Application of Laser Sensor in Urban Vehicle Type Detection [J]. Tehnicki Vjesnik-Technical Gazette, 2018, 25(2): 622-626.

[11] Zhao J X, Xu H, Liu H C, et al. Detection and tracking of pedestrians and vehicles using roadside LiDAR sensors [J]. Transportation Research Part C-Emerging Technologies, 2019, 100: 68-87.

[12] Solomitckii D, Barneto C B, Turunen M, et al. Millimeter-Wave Radar Scheme With Passive Reflector for Uncontrolled Blind Urban Intersection [J]. IEEE Transactions on Vehicular Technology, 2021, 70(8): 7335-7346.

[13] Li X, Tao X W, Zhu B, et al. Research on a Simulation Method of the Millimeter Wave Radar Virtual Test Environment for Intelligent Driving [J]. Sensors, 2020, 20(7).

[14] Isiker H, Unal I, Tekbas M, et al. An auto-classification procedure for concealed weapon detection in millimeter-wave radiometric imaging systems [J]. Microwave and Optical Technology Letters, 2018, 60(3): 583-594.

[15] Balal Y, Balal N, Richter Y, et al. Time-Frequency Spectral Signature of Limb Movements and Height Estimation Using Micro-Doppler Millimeter-Wave Radar [J]. Sensors, 2020, 20(17).

[16] Lin Z P, Lv T J, Mathiopoulos P T. 3-D Indoor Positioning for Millimeter-Wave Massive MIMO Systems [J]. IEEE Transactions on Communications, 2018, 66(6): 2472-2486.

[17] Passafiume M, Rojhani N, Collodi G, et al. Modeling Small UAV Micro-Doppler Signature Using Millimeter-Wave FMCW Radar [J]. Electronics, 2021, 10(6).

[18] Peng Y, Abdel-Aty M, Shi Q, et al. Assessing the impact of reduced visibility on traffic crash risk using microscopic data and surrogate safety measures [J]. Transportation Research Part C-Emerging Technologies, 2017, 74: 295-305.

[19] Lim H S, Park H M, Lee J E, et al. Lane-by-Lane Traffic Monitoring Using 24.1 GHz FMCW Radar System [J]. IEEE Access, 2021, 9: 14677-14687.

[20] Lim H S, Lee J E, Park H M, et al. Stationary Target Identification in a Traffic Monitoring Radar System [J]. Applied Sciences-Basel, 2020, 10(17).

[21] Nimac P, Krpic A, Batagelj B, et al. Pedestrian Traffic Light Control with Crosswalk FMCW Radar and Group Tracking Algorithm [J]. Sensors, 2022, 22(5).

[22] Ho T-J, Chung M-J. Information-Aided Smart Schemes for Vehicle Flow Detection Enhancements of Traffic Microwave Radar Detectors [J]. Applied Sciences-Basel, 2016, 6(7).

[23] Liu H Q, Li N, Guan D Y, et al. Data Feature Analysis of Non-Scanning Multi Target Millimeter-Wave Radar in Traffic Flow Detection Applications [J]. Sensors, 2018, 18(9).

[24] Zhang M F, Fu R, Guo Y S, et al. Cyclist detection and tracking based on multi-layer laser scanner [J]. Human-Centric Computing and Information Sciences, 2020, 10(1).

[25] Wu J Q, Zhang Y S, Tian Y, et al. Automatic Vehicle Tracking with LiDAR-Enhanced Roadside Infrastructure [J]. Journal of Testing and Evaluation, 2021, 49(1): 121-133.

[26] Farag W. Road-objects tracking for autonomous driving using lidar and radar fusion [J]. Journal of Electrical Engineering-Elektrotechnicky Casopis, 2020, 71(3): 138-149.

[27] Malzer C, Baum M. Constraint-Based Hierarchical Cluster Selection in Automotive Radar Data [J]. Sensors, 2021, 21(10).

[28] Jung Y, Seo S W, Kim S W. Curb Detection and Tracking in Low-Resolution 3D Point Clouds Based on Optimization Framework [J]. IEEE Transactions on Intelligent Transportation Systems, 2020, 21(9): 3893-3908.

[29] Cao L, Zhang X Y, Wang T, et al. An Adaptive Ellipse Distance Density Peak Fuzzy Clustering Algorithm Based on the Multi-target Traffic Radar [J]. Sensors, 2020, 20(17).

[30] 金元华, 叶春茂, 鲁耀兵, 等 一种基于CFAR检测和密度聚类的ISAR图像预处理方法 [J]. 太赫兹科学与电子信息学报, 2020, 18(02): 278-283+90.

[31] 刘全周, 贾鹏飞, 李占旗, 等. 改进的交互式卡尔曼滤波对雷达数据处理技术研究 [J]. 机械科学与技术, 2020, 39(08): 1248-1255.

[32] 何静. 利用卡尔曼滤波预测船舶航行轨迹异常行为 [J]. 舰船科学技术, 2017, 39(02): 16-18.

[33] Wei Y, Hong T, Kadoch M. Improved Kalman Filter Variants for UAV Tracking with Radar Motion Models [J]. Electronics, 2020, 9(5).

[34] Kim T, Park T-H. Extended Kalman Filter (EKF) Design for Vehicle Position Tracking Using Reliability Function of Radar and Lidar [J]. Sensors, 2020, 20(15).

[35] 冯辉, 叶海洋, 徐海祥, 等. Alpha-beta滤波在船舶动力定位状态估计中的应用 [J]. 舰船科学技术, 2018, 40(03): 64-67.

[36] 刘常娟, 王运锋, 杨玲. 不同Alpha-Beta滤波算法的精度分析及改进 [J]. 四川大学学报(自然科学版), 2018, 55(02): 265-270.

[37] Song S, Wu J, Zhang S, et al. Research on Target Tracking Algorithm Using Millimeter-Wave Radar on Curved Road [J]. Mathematical Problems in Engineering, 2020, 2020.

[38] Rakai L, Song H, Sun S, et al. Data association in multiple object tracking: A survey of recent techniques [J]. Expert Systems with Applications, 2022, 192.

[39] 韩蕾蕾. 一种基于D-CJPDA的多目标数据互联算法 [J]. 电光与控制, 2020, 27(06): 6-10.

[40] 潘泉, 叶西宁, 张洪才. 广义概率数据关联算法 [J]. 电子学报, 2005, (03): 467-472.

[41] 张强, 于俊朋, 谢苏道. GM-PHD雷达密集多目标跟踪应用研究 [J]. 现代雷达, 2019, 41(08): 41-44+69.

[42] 侯淋, 杨顺华. 光电雷达多目标跟踪技术发展分析 [J]. 电光与控制, 2021, 28(08): 65-70.

[43] Patole S, Torlak M, Wang D, et al. Automotive Radars A review of signal processing techniques [J]. IEEE Signal Processing Magazine, 2017, 34(2): 22-35.

[44] Bhattacharjee P, Mitra P. A survey of density based clustering algorithms [J]. Frontiers of Computer Science, 2021, 15(1).

[45] 龚轩, 乐孜纯, 王慧, 等. 多目标跟踪中的数据关联技术综述 [J]. 计算机科学, 2020, 47(10): 136-144.

[46] 张琤, 赵菡, 林家骏. 基于关联性能评估的多目标跟踪关联门算法 [J]. 华东理工大学学报(自然科学版), 2019, 45(02): 336-343.

[47] Kumar M, Mondal S. Recent developments on target tracking problems: A review [J]. Ocean Engineering, 2021, 236.

[48] 彭刚, 袁晓, 刘闻. 雷达辐射源信号聚类分选算法综述 [J]. 雷达科学与技术, 2019, 17(05): 485-492.

[49] 黄辰. 几种典型聚类方法在雷达信号分选中的应用浅析 [J]. 电子信息对抗技术, 2017, 32(05): 1-4+62.

[50] van de Velden M, D'Enza A I, Markos A. Distance-based clustering of mixed data [J]. Wiley Interdisciplinary Reviews-Computational Statistics, 2019, 11(3).

[51] Schubert E, Sander J, Ester M, et al. DBSCAN Revisited, Revisited: Why and How You Should (Still) Use DBSCAN [J]. Acm Transactions on Database Systems, 2017, 42(3).

[52] Yu T-Y, Wu C-M. Application of Adaptive alpha-beta-gamma and alpha-beta-gamma-delta Filters to Tracking Systems [J]. Sensors and Materials, 2017, 29(4): 419-427.

[53] 李刚, 赵德阳, 解瑞春, 等. 基于改进的Sage-Husa自适应扩展卡尔曼滤波的车辆状态估计 [J]. 汽车工程, 2015, 37(12): 1426-1432.

[54] 孙旭, 张教镭, 张耀升, 等. 一种改进型Sage-Husa自适应跟踪滤波器 [J]. 电子信息对抗技术, 2020, 35(06): 37-41.