All Issue

2024 Vol.42, Issue 5 Preview Page

Article

Drone Vision-based Lane Change Prediction in Urban Vehicle Driving with Transformer

드론 영상 인식과 Transformer를 활용한 도심 주행 차량의 차로 변경 예측 모델 연구

Seokjun HONG1
,
Jaegyun IM1
,
Howoo LEE1
,
Dongho KA2
,
Changhee LEE3
,
Byeongjoon NOH4*
홍 석준1
,
임 재균1
,
이 호우1
,
가 동호2
,
이 창희3
,
노 병준4*
1Student, Department of AI and Big Data, SoonChunHyang University, Asan 31538, Korea
2Senior Researcher, ITS Team, Nota AI, Seoul 06164, Korea
3General Manager, ITS Team, Nota AI, Seoul 06164, Korea
4Assistant Professor, Department of AI and Big Data, SoonChunHyang University, Asan 31538, Korea
1순천향대학교 AI·빅데이터학과 학부과정
2Nota AI ITS팀 수석연구원
3Nota AI ITS팀 부장
4순천향대학교 AI·빅데이터학과 조교수
*Corresponding Author
31 October 2024. pp. 551-564
PDF XML Citation
Abstract

Lane changes pose a risk of accidents due to lane interference between vehicles, affecting traffic flow. Therefore, research on predicting lane changes is necessary to reduce accident risks and improve the traffic system. This study proposes a method for extracting vehicle information from top-view drone footage and predicting lane changes. We used the deep learning-based object detection model YOLOv8 for vehicle detection in the footage and extracted Traffic Context from the detected objects. Using this extracted information, we performed lane change prediction utilizing a Transformer model. For our experiments, we utilized drone footage of road conditions in Guseong-dong, Yuseong-gu, Daejeon. The experimental results demonstrated that YOLOv8 achieved a high object detection performance with an mAP50 of 0.99 and an mAP50-95 of 0.75, indicating a high degree of agreement between detected and actual objects. In terms of lane change prediction, the Transformer model outperformed other models, achieving an accuracy of 0.977, which is 0.02 higher, and the lowest misclassification rate in the confusion matrix. This study is expected to contribute to improving traffic systems by serving as a valuable resource for efficient traffic flow management in the future.

Keywords
drone-vision
lane change prediction
traffic context
transformer
yolov8

차로 변경은 차량 간의 차선 간섭으로 인해 사고의 위험이 존재하며, 이에 따라 교통 흐름에 영향을 미친다. 따라서, 사고의 위험성을 줄이고 교통 시스템을 원활하게 하기 위해 차로 변경 예측에 관한 연구는 필요하다. 본 연구에서는 드론으로 수집된 차량의 탑 뷰(top-view) 영상 정보를 활용하여 차량에 대한 정보를 추출하고, 차로 변경을 예측하는 방법에 대하여 제시한다. 먼저, 영상 내 차량 객체 탐지를 위해 딥러닝 기반 객체 탐지 모델인 YOLOv8을 사용하였으며, 객체 정보를 바탕으로 Traffic Context를 추출한다. 추출된 정보를 바탕으로 Transformer 모델을 활용하여 차로 변경 예측을 수행한다. 본 연구 실험에서는 드론으로 촬영한 대전 유성구 구성동의 도로 상황 영상을 활용하였다. 실험 결과 YOLOv8의 객체 탐지 성능으로는 mAP50에서 0.99, mAP50-95에서는 0.75로 검출 객체와 실체 객체 간의 높은 일치도를 확인하였다. 차로 변경 예측 모델에서는 Transformer가 다른 모델들에 비해 0.02 높은 0.977의 Accuracy를 보였으며, 혼동 행렬에서 가장 낮은 오분류를 확인하였다. 본 연구가 향후 효율적인 교통 흐름 관리를 위한 자료로 활용되어 교통 시스템 개선에 기여할 수 있을 것으로 기대된다.

키워드
드론 비전
차로 변경 예측
traffic context
transformer
yolov8
References
1

Chen T., Shi X., Wong Y. D., Yu X. (2020), Predicting Lane-changing Risk Level Based on Vehicles' Space-series Features: A Pre-emptive Learning Approach, Transp. Res. Part C: Emerg. Technol., 116, Elsevier, 102646.

10.1016/j.trc.2020.102646
2

Jeong D. H., Kim M. J., Kim H. K., Chung Y.S. (2021), Development and Evaluation of Centralized Autonomous Vehicle Mobility Service Using ADAS Data, J. Korean Soc. Transp., 39(5), Korean Society of Transportation, 631-642.

10.7470/jkst.2021.39.5.631
3

Jiang C., Ren H., Ye X., Zhu J., Zeng H., Nan Y., Huo H. (2022), Object Detection from UAV Thermal Infrared Images and Videos Using YOLO Models, Int. J. Appl. Earth Obs. Geoinf., 112, Elsevier, 102912.

10.1016/j.jag.2022.102912
4

Jokhio S., Dürr M., Bärgman J., Baumann M. (2024), Influence of Surrounding Traffic on Lane Change Dynamics: Insights from a Video-based Laboratory Study, Transp. Res. Part F: Traffic Psychol. Behav., 105, Elsevier, 87-98.

10.1016/j.trf.2024.06.025
5

Ko E. J., Kim S. H., Kim H. J. (2021), Microscopic Traffic Analysis of Freeway Based on Vehicle Trajectory Data Using Drone Images, J. Korean Soc. Intell. Transp. Syst., 20(6), The Korean Society of Intelligent Transport Systems, 66-83.

10.12815/kits.2021.20.6.66
6

Kim J. H., Han D. S. (2020), A Path Prediction Algorithm of Surrounding Vehicles Based on Sensor Fusion for Safe Lane Change, J. Korean Inst. Commun. Inf. Sci., 45(5), The Korean Institute of Communications and Information Sciences, 828-836.

10.7840/kics.2020.45.5.828
7

LabelImg, https://github.com/tzutalin/labelImg, 2024.2.15.

8

Lee J. H., Hwang Y. H., Kwon H. J., Choi J. W., Lee J. T. (2023), Deep Learning-Based Roundabout Traffic Analysis System Using Unmanned Aerial Vehicle Videos, IEMEK J. Embed. Syst. Appl., 18(3), The Institute of Embedded Engineering of Korea, 125-132.

9

Lee S. H., Kim D. K. (2023), Development of Rerouting Algorithm of Autonomous Vehicles for Mitigating Traffic Congestion in Mixed Traffic Environments, The 89th Conf. of KST, Korean Society of Transportation, 333-334.

10

Lertworawanich P. (2006), Safe-following Distances Based on the Car-following Model, In PIARC International Seminar on Intelligent Transport System (ITS) in Road Network Operations, Kuala Lumpur, Malaysia.

11

Liu Q., Zhang Y., Yang G. (2023), Small Unopened Cotton Boll Counting by Detection with MRF-YOLO in the Wild, Comput. Electron. Agric., 204, Elsevier, 107576.

10.1016/j.compag.2022.107576
12

Lin T. Y., Maire M., Belongie S., Hays J., Perona P., Ramanan D., et al. (2014), Microsoft Coco: Common Objects in Context, In Comput. Vis. - ECCV 2014: 13th Eur. Conf., Part V 13, Springer, 740-755.

10.1007/978-3-319-10602-1_48
13

Luo Y., Xiang Y., Cao K., Li K. (2016), A Dynamic Automated Lane Change Maneuver Based on Vehicle-to-Vehicle Communication, Transp. Res. Part C: Emerg. Technol., 62, Elsevier, 87-102.

10.1016/j.trc.2015.11.011
14

Mozaffari S., Arnold E., Dianati M., Fallah S. (2022), Early Lane Change Prediction for Automated Driving Systems Using Multi-task Attention-based Convolutional Neural Networks, IEEE Trans. Intell. Veh., 7(3), IEEE, 758-770.

10.1109/TIV.2022.3161785
15

Patel S., Griffin B., Kusano K., Corso J. J. (2018), Predicting Future Lane Changes of Other Highway Vehicles Using RNN-based Deep Models, arXiv preprint arXiv:1801.04340.

16

Redmon J., Divvala S., Girshick R., Farhadi A. (2016), You Only Look Once: Unified, Real-time Object Detection, In Proc. IEEE Conf. Comput. Vis. Pattern Recognit., IEEE, 779-788.

10.1109/CVPR.2016.91
17

Shawky M. (2020), Factors Affecting Lane Change Crashes, IATSS Res., 44(2), Elsevier, 155-161.

10.1016/j.iatssr.2019.12.002
18

Song R., Li B. (2021), Surrounding Vehicles' Lane Change Maneuver Prediction and Detection for Intelligent Vehicles: A Comprehensive Review, IEEE Trans. Intell. Transp. Syst., 23(7), IEEE, 6046-6062.

10.1109/TITS.2021.3076164
19

Vaswani A., Shazeer N., Parmar N., Uszkoreit J., Jones L., Gomez A. N., et al. (2017), Attention Is All You Need, Adv. Neural Inf. Process. Syst., 30, 5998-6008.

20

Wissing C., Nattermann T., Glander K. H., Hass C., Bertram T. (2017), Lane Change Prediction by Combining Movement and Situation-based Probabilities, IFAC-PapersOnLine, 50(1), Elsevier, 3554-3559.

10.1016/j.ifacol.2017.08.960
21

Xue Q., Xing Y., Lu J. (2022), An Integrated Lane Change Prediction Model Incorporating Traffic Context Based on Trajectory Data, Transp. Res. Part C: Emerg. Technol., 141, Elsevier, 103738.

10.1016/j.trc.2022.103738
22

YOLOv8, https://github.com/ultralytics/ultralytics, 2024.3.13.

23

Yang L., Li X., Guan W., Zhang H. M., Fan L. (2018), Effect of Traffic Density on Drivers' Lane Change and Overtaking Maneuvers in Freeway Situations: A Driving Simulator-based Study, Traffic Inj. Prev., 19(6), Taylor & Francis, 594-600.

10.1080/15389588.2018.147147029757689
24

Yuan R., Abdel-Aty M., Gu X., Zheng O., Xiang Q. (2023), A Unified Modeling Framework for Lane Change Intention Recognition and Vehicle Status Prediction, Physica A: Stat. Mech. Appl., 632, Elsevier, 129332.

10.1016/j.physa.2023.129332
25

Zhang Y., Sun P., Jiang Y., Yu D., Weng F., Yuan Z., Wang X. (2022), Bytetrack: Multi-object Tracking by Associating Every Detection Box, In Eur. Conf. Comput. Vis., Springer Nature, 1-21.

10.1007/978-3-031-20047-2_1
26

Zhao D., Lam H., Peng H., Bao S., LeBlanc D. J., Nobukawa K., et al. (2016), Accelerated Evaluation of Automated Vehicles Safety in Lane-change Scenarios Based on Importance Sampling Techniques, IEEE Trans. Intell. Transp. Syst., 18(3), IEEE, 595-607.

10.1109/TITS.2016.258220827840592PMC5103645
27

Zheng O., Abdel-Aty M., Yue L., Abdelraouf A., Wang Z., Mahmoud N. (2022), CitySim: A Drone-based Vehicle Trajectory Dataset for Safety-Oriented Research and Digital Twins, arXiv preprint arXiv:2208.11036.

10.1177/03611981231185768
Information
  • Publisher :Korean Society of Transportation
  • Publisher(Ko) :대한교통학회
  • Journal Title :Journal of Korean Society of Transportation
  • Journal Title(Ko) :대한교통학회지
  • Volume : 42
  • No :5
  • Pages :551-564
  • Received Date : 2024-05-23
  • Revised Date : 2024-06-17
  • Accepted Date : 2024-08-02
  • DOI :https://doi.org/10.7470/jkst.2024.42.5.551
Journal Informaiton Journal of Korean Society of Transportation Journal of Korean Society of Transportation
Archives
: 2015 Vol.34~
  • NRF
  • KOFST
  • crossref cited-by
  • crosscheck
  • orcid
  • open access
  • ccl
Journal Informaiton Journal Informaiton - close