隧道环境毫米波雷达目标识别与分类算法

doi:10.12305/j.issn.1001-506X.2025.05.08

系统工程与电子技术 ›› 2025, Vol. 47 ›› Issue (5): 1453-1460.doi: 10.12305/j.issn.1001-506X.2025.05.08

• 传感器与信号处理 • 上一篇

隧道环境毫米波雷达目标识别与分类算法

姜智杰¹^,², 宋恒³^,*, 胡楠³, 段兰茜³, 曹平¹^,⁴

1. 中国科学技术大学核探测与核电子学国家重点实验室, 安徽合肥 230026
2. 中国科学技术大学近代物理系, 安徽合肥 230026
3. 中铁四局集团有限公司管理与技术研究院, 安徽合肥 230041
4. 中国科学技术大学核科学技术学院, 安徽合肥 230026

收稿日期:2024-04-23 出版日期:2025-06-11 发布日期:2025-06-18
通讯作者: 宋恒
作者简介:姜智杰 (1996—), 男, 博士研究生, 主要研究方向为毫米波雷达信号处理
宋恒 (1979—), 男, 正高级工程师, 博士, 主要研究方向为智能信息处理
胡楠 (1996—), 男, 助理工程师, 硕士, 主要研究方向为人工智能
段兰茜 (1997—), 女, 助理工程师, 硕士, 主要研究方向为毫米波雷达硬件系统
曹平 (1980—), 男, 副教授, 博士, 主要研究方向为高速数据采集、读出及实时处理
基金资助:
中铁四局集团有限公司(2023-48)

Target recognition and classification algorithm of MMW radar in tunnel

Zhijie JIANG¹^,², Heng SONG³^,*, Nan HU³, Lanxi DUAN³, Ping CAO¹^,⁴

1. State Key Laboratory of Particle Detection and Electronics, University of Science and Technology of China, Hefei 230026, China
2. Department of Modern Physics, University of Science and Technology of China, Hefei 230026, China
3. Management and Technology Institute, China Railway No.4 Engineering Group CO., LTD, Hefei 230041, China
4. School of Nuclear Science and Technology, University of Science and Technology of China, Hefei 230026, China

Received:2024-04-23 Online:2025-06-11 Published:2025-06-18
Contact: Heng SONG

摘要/Abstract

摘要：

毫米波雷达具有全天候、高精度、低成本、无感式的优点, 在隧道环境下进行安全监测具有极大的优势。但由于隧道多径干扰严重, 经典信号处理算法目标识别错误率高, 三维卷积深度学习算法计算复杂度高, 实时性差, 阻碍了毫米波雷达在隧道中的应用。基于此, 提出一种高效的深度学习算法方案, 可以实现人员、车辆等目标的高精度实时定位和分类。算法采用信号处理方法对雷达中频数据进行多维度压缩编码；使用Mamba网络针对雷达时空序列数据进行特征提取；使用视场数据热图估计目标位置；使用目标位置局部区域特征估计目标类别, 避免不相干区域信号干扰, 提高目标识别精准率。算法基于二维卷积设计, 建立雷达数据到目标位置及类别的非线性映射关系, 可有效控制计算复杂度。隧道试验表明, 算法的平均交并比(mean intersection over union, mIoU)、平均精准率(average precision, AP)和速度分别为0.68, 65.26%, 22.5 ms/frame, 与三维卷积CenterRadarNet算法相比, mIoU降低3%, AP提升9%, 速度提升53.3%。证明算法具有良好性能, 在实际工程中具有应用价值。

关键词: 毫米波雷达, 目标识别, 实时, 深度学习, 隧道

Abstract:

Millimeter wave (MMW) radar exhibits all-weather capability, high precision, low cost, and non-contact sensing, rendering it highly suitable for safety monitoring in tunnels. However, due to severe multipath interference in tunnels, classical signal processing algorithms have a high error rate in target recognition. The high computational complexity and poor real-time performance of three-dimensional (3D) convolutional deep learning algorithms hinder the application of MMW radar in tunnels. To regard this, an efficient deep-learning algorithm scheme is proposed, which can achieve high-precision real-time positioning and classification of targets such as individuals, vehicles, and other targets. The algorithm utilizes a signal processing method to compress and encode the radar intermediate frequency data across multiple dimensions, uses Mamba network to extract features from radar spatio-temporal sequence data, uses heatmap of field data to estimate target location, and uses only local regional features of target location to estimate target category, avoiding incoherent regional signal interference and improving target recognition accuracy. The algorithm is designed based on two-dimensional (2D) convolution, and a nonlinear mapping relationship between radar data and target location and category is established, effectively managing computational complexity. Experiments in the tunnel show that the mean intersection over union (mIoU), average precision (AP) and speed of the algorithm are respectively 0.68, 65.26%, 22.5 ms/frame, compared with the 3D convolutional CenterRadarNet algorithm, mIoU is reduced by 3%, AP is increased by 9%, and speed is increased by 53.3%, which proves that the algorithm has good performance and has application value in actual application.

Key words: millimeter wave (MMW) radar, target recognition, real time, deep learning, tunnel

中图分类号:

TN958.94

姜智杰, 宋恒, 胡楠, 段兰茜, 曹平. 隧道环境毫米波雷达目标识别与分类算法[J]. 系统工程与电子技术, 2025, 47(5): 1453-1460.

Zhijie JIANG, Heng SONG, Nan HU, Lanxi DUAN, Ping CAO. Target recognition and classification algorithm of MMW radar in tunnel[J]. Systems Engineering and Electronics, 2025, 47(5): 1453-1460.

图/表 9

图1

图2

图3

图4

表1

图5

图6

图7

表2

参考文献 32

1	郑艾辰, 黄锋, 林志, 等. 2008年至2016年我国隧道工程施工安全事故统计与分析[J]. 施工技术, 2017, 46 (S1): 833- 836.
	ZHENG A C , HUANG F , LIN Z , et al. Statistics and analysis of tunnel construction safety accidents in China from 2008 to 2016[J]. Construction Technology, 2017, 46 (S1): 833- 836.
2	LUKE R A . Diamonds with dispatch: computerized LHD dispatching at DeBeers Finsch-a progress report[J]. World Mining Equipment, 1992, 16 (6): 24-25, 27-28.
3	宋磊. ZigBee技术在物联网中的应用展望[J]. 现代制造技术与装备, 2017 (2): 43- 44.
	SONG L . Application prospect of ZigBee technology in internet of things[J]. Modern Manufacturing Technology and Equipment, 2017 (2): 43- 44.
4	AHMED N , GHARGHAN S K , MUTLAG A H , et al. Children tracking system based on zigbee wireless network and neural network[J]. Journal of Techniques, 2023, 5 (1): 103- 113. doi: 10.51173/jt.v5i1.838
5	徐潇潇, 谢林柏, 彭力. 基于WiFi信号强度特征的室内定位系统设计[J]. 计算机工程, 2015, 41 (4): 87- 91.
	XU X X , XIE L B , PENG L . Design of indoor positioning system based on WiFi signal strength characteristics[J]. Computer Engineering, 2015, 41 (4): 87- 91.
6	LEIRA F S , HELGESEN H H , JOHANSEN T A , et al. Object detection, recognition, and tracking from UAVs using a thermal camera[J]. Journal of Field Robotics, 2021, 38 (2): 242- 267. doi: 10.1002/rob.21985
7	RAVINDRAN R , SANTORA M J , JAMALI M M . Multi-object detection and tracking, based on DNN, for autonomous vehicles: a review[J]. IEEE Sensors Journal, 2020, 21 (5): 5668- 5677.
8	李伟杰, 戴亚文, 唐雨, 等. 基于UWB的地铁隧道定位系统设计[J]. 计算机测量与控制, 2019, 27 (8): 198-201, 206.
	LI W J , DAI Y W , TANG Y , et al. Design of subway tunnel positioning system based on UWB[J]. Computer Measurement and Control, 2019, 27 (8): 198-201, 206.
9	ANISH S , NEHARIKA V , ENVER B , et al. A review of millimeter wave device-based localization and device-free sensing technologies and applications[J]. IEEE Communications Surveys & Tutorials, 2022, 24 (3): 1708- 1749.
10	魏军, 韩君龙, 王巍, 等. 基于毫米波雷达的铁路隧道人员定位技术研究[J]. 隧道建设(中英文), 2023, 43 (S2): 223- 229.
	WEI J , HAN J L , WANG W , et al. Research on personnel location technology of railway tunnel based on millimeter wave radar[J]. Tunnel Construction, 2023, 43 (S2): 223- 229.
11	XING Y C, KANHERE O, JU S H, et al. Indoor wireless channel properties at millimeter wave and sub-terahertz frequencies[C]//Proc. of the IEEE Global Communications Conference, 2019.
12	WANG Y Z , JIANG Z Y , LI Y D , et al. RODNet: a real-time radar object detection network cross-supervised by camera-radar fused object 3D localization[J]. IEEE Journal of Selected Topics in Signal Processing, 2021, 15 (4): 954- 967. doi: 10.1109/JSTSP.2021.3058895
13	CHENG J H, KUAN S Y, LATAPIE H, et al. CenterRadarNet: joint 3D object detection and tracking framework using 4D FMCW radar[EB/OL]. [2024-03-24]. https://arxiv.org/abs/2311.01423.
14	GU A, DAO T. Mamba: linear-time sequence modeling with selective state spaces[EB/OL]. [2024-03-24]. https://arxiv.org/abs/2312.00752.
15	PEARCE A , ZHANG J A , XU R , et al. Multi-object tracking with MMwave radar: a review[J]. Electronics, 2023, 12 (2): 308. doi: 10.3390/electronics12020308
16	张晨, 叶舟, 吕宇宙, 等. 基于ResNet的智能恒虚警目标检测方法[J]. 上海航天(中英文), 2023, 39 (5): 71- 78.
	ZHANG C , YE Z , LV Y Z , et al. Intelligent constant false alarm detection method based on ResNet[J]. Aerospace ShangHai (Chinese & English), 2023, 39 (5): 71- 78.
17	CHENG Y W , SU J R , JIANG M X , et al. A novel radar point cloud generation method for robot environment perception[J]. IEEE Trans. on Robotics, 2022, 38 (6): 3754- 3773. doi: 10.1109/TRO.2022.3185831
18	SUN Y, HUANG Z, ZHANG H Z, et al. 3DRIMR: 3D reconstruction and imaging via mmWave radar based on deep learning[C]//Proc. of the IEEE International Performance, Computing, and Communications Conference, 2021.
19	周驰, 李智, 徐灿. 基于深度神经网络的空间目标结构识别[J]. 中国空间科学技术, 2019, 39 (1): 32- 39.
	ZHOU C , LI Z , XU C . Spatial target structure recognition based on deep neural network[J]. China Space Science and Technology, 2019, 3, 9.
20	孟继成, 杨万麟. 基于核函数的雷达一维距离像目标识别[J]. 电子与信息学报, 2005, 27 (3): 462- 466.
	MENG J C , YANG W L . Radar one-dimensional range image target recognition based on kernel function[J]. Journal of Electronics and Information, 2005, 27 (3): 462- 466.
21	JITHESH V, SAGAYARAJ M J, SRINIVASA K G, et al. LSTM recurrent neural networks for high resolution range profile based radar target classification[C]//Proc. of the 3rd International Conference on Computational Intelligence & Communication Technology, 2017.
22	GAO C, LIU H W, ZHOU S H, et al. Maneuvering target tracking with recurrent neural networks for radar application[C]// Proc. of the International Conference on Radar, 2018.
23	VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Proc. of the International Conference on Neural Information Processing Systems, 2017: 6000-6010.
24	DUAN K W, BAI S, XIE L X, et al. CenterNet: keypoint triplets for object detection[C]//Proc. of the IEEE/CVF International Conference on Computer Vision, 2019: 6568-6577.
25	HUSSAIN M . YOLOv1 to v8: unveiling each variant-a comprehensive review of YOLO[J]. IEEE Access, 2024, 12, 42816- 42833. doi: 10.1109/ACCESS.2024.3378568
26	JIANG P Y , ERGU D J , LIU F Y , et al. A review of Yolo algorithm developments[J]. Procedia Computer Science, 2022, 199, 1066- 1073. doi: 10.1016/j.procs.2022.01.135
27	LEE Y H , KIM Y . Comparison of CNN and YOLO for object detection[J]. Journal of the Semiconductor & Display Technology, 2020, 19 (1): 85- 92.
28	GIRSHICK R. Fast R-CNN[C]//Proc. of the IEEE International Conference on Computer Vision, 2015: 1440-1448.
29	LI W Z . Analysis of object detection performance based on Faster R-CNN[J]. Journal of Physics: Conference Series, 2021, 1827 (1): 012085. doi: 10.1088/1742-6596/1827/1/012085
30	WEI L, DRAGOMIR A, DUMITRU E, et al. SSD: single shot MultiBox detector[C]//Proc. of the European Conference on Computer Vision, 2016: 21-37.
31	ZHAI S P , SHANG D R , WANG S H , et al. DF-SSD: an improved SSD object detection algorithm based on DenseNet and feature fusion[J]. IEEE Access, 2020, 8, 24344- 24357. doi: 10.1109/ACCESS.2020.2971026
32	LU X C , JI J , XING Z Q , et al. Attention and feature fusion SSD for remote sensing object detection[J]. IEEE Trans. on Instrumentation and Measurement, 2021, 70, 5501309.

参数名	参数值
起始频率/GHz	60
调频斜率/(MHz/μs)	79
扫频带宽/GHz	3.16
采样率/MSPS	12.5
采样点数	64
每帧chirp数	32
帧周期/ms	20
通道数	3个发射天线, 4个接收天线

评价指标	本文方法		CenterRadarNet
评价指标	车库	隧道	车库	隧道
mIoU	0.69	0.68	0.72	0.70
AP/%	67.04	65.26	61.15	59.84
速度/(ms/frame)	20.6	22.5	46.9	48.2

[1]	郭泽坤, 杨洪飞, 刘峥, 谢荣, 冉磊, 李嘉楠. 基于MGMD空间的窄带雷达空中目标识别方法[J]. 系统工程与电子技术, 2025, 47(4): 1136-1145.
[2]	聂千祁, 沙明辉, 朱应申, 王崇宇, 崔念强. 基于盲源分离结合奇异谱分析的雷达多分量信号识别方法[J]. 系统工程与电子技术, 2025, 47(4): 1168-1175.
[3]	袁磊, 杨艳娟, 郭毅, 戴鹏. 相关噪声下基于深度学习的LDPC码码率半盲识别算法[J]. 系统工程与电子技术, 2025, 47(4): 1335-1345.
[4]	王奇, 王子瑶, 郑峻峰. 考虑多源噪声及信号传输的雷达系统仿真模型[J]. 系统工程与电子技术, 2025, 47(3): 768-778.
[5]	李家宽, 冯博, 刘红亮, 叶春茂, 余继周. 基于角度引导注意力的气动目标宽带PD识别方法[J]. 系统工程与电子技术, 2025, 47(3): 807-816.
[6]	黄冬, 刘钢, 江雪云, 刘立志, 郭金林. 基于滚动优化结构的行动过程实时决策方法[J]. 系统工程与电子技术, 2025, 47(3): 904-911.
[7]	沈夏闰, 李若楠, 张昊田. 基于CVAE-LSTM的服务器KPI异常检测[J]. 系统工程与电子技术, 2025, 47(3): 1019-1027.
[8]	李雪健, 陈永强, 马宏, 刘杨, 王育欣, 焦义文. 基于并行计算的PCAL信号相位实时提取系统设计[J]. 系统工程与电子技术, 2025, 47(2): 376-389.
[9]	连红飞, 李东升, 蒋彦雯, 范红旗, 肖怀铁, 王国嫣. 一种伯努利粒子滤波器的FPGA实现[J]. 系统工程与电子技术, 2025, 47(2): 398-405.
[10]	刘晓琳, 郭梦娇, 李卓. Dueling DQN优化下的航班延误自适应图卷积循环网络预测方法[J]. 系统工程与电子技术, 2025, 47(2): 568-579.
[11]	付卫红, 张鑫钰, 刘乃安. 基于多尺度融合神经网络的同频同调制单通道盲源分离算法[J]. 系统工程与电子技术, 2025, 47(2): 641-649.
[12]	迟俊, 孙溶辰, 孙志国, 易振宇. 基于传播图论的6 GHz隧道信道建模[J]. 系统工程与电子技术, 2025, 47(1): 316-323.
[13]	蔡伟, 王鑫, 蒋昕昊, 杨志勇, 陈栋. 基于解耦的小样本目标检测方法研究[J]. 系统工程与电子技术, 2024, 46(9): 2941-2950.
[14]	尹建国, 盛文, 蒋伟. 基于深度残差收缩网络的雷达空中目标识别[J]. 系统工程与电子技术, 2024, 46(9): 3012-3018.
[15]	张明龙, 吴雨林, 魏文强, 沈园杰, 郭世盛, 崔国龙. 基于稀疏矩阵填充的级联毫米波雷达高分辨测角方法[J]. 系统工程与电子技术, 2024, 46(8): 2629-2640.

隧道环境毫米波雷达目标识别与分类算法

Target recognition and classification algorithm of MMW radar in tunnel

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 32

相关文章 15

编辑推荐

Metrics

本文评价