All Issue

2025 Vol.43, Issue 2 Preview Page

Article

Analysis of Take-Over Cluster Points and Influencing Factors Using Automated Driving Operation Data

자율주행 운행데이터를 활용한 제어권전환 다발구간 및 영향 요인 분석

Sunghyun KIM1*
,
Kyeong Pyo KANG2
,
Seungjae LEE3
김 성현1*
,
강 경표2
,
이 승재3
1Researcher, Department of Mobility Transformation, The Korea Transport Institute, Sejong 30147, Korea
2Senior Researcher Fellow, Department of Mobility Transformation, The Korea Transport Institute, Sejong 30147, Korea
3Professor, Department of Transportation Engineering, University of Seoul, Seoul 02504, Korea
1한국교통연구원 모빌리티전환연구본부 연구원
2한국교통연구원 모빌리티전환연구본부 선임연구위원
3서울시립대학교 교통공학과 교수
*Corresponding Author
30 April 2025. pp. 145-160
PDF XML Citation
Abstract

As automated driving technology advances, pilot operations of automated vehicles have become increasingly active in South Korea. Currently, domestically operated automated vehicles are classified as Level 3 according to the Society of Automotive Engineers (SAE) standards. In situations where smooth operation is challenging, the safety driver seated in the vehicle assumes control, transitioning from automated to manual driving. This process, defined as a “take-over,” is a critical factor for ensuring the safety of automated driving operations. However, research on the factors contributing to take-over occurrences remains insufficient. This study investigates zones where take-overs frequently occur and identifies the static factors associated with these zones. By proposing methods to reduce take-overs, the study aims to enhance the operational efficiency of automated driving services. Domestic automated driving data were used to define analysis zones, and clustering techniques were applied to identify take-over hotspots. Statistical comparisons between take-over hotspots and non-hotspot regions were conducted to determine the factors influencing take-overs. The analysis identified 28 zones within the operational area where take-overs frequently occurred. Take-overs were found to be more prevalent in areas with intersections, a high number of lanes, or zones with three or more pedestrian crossings. Based on these findings, the study proposes services to provide preemptive information on anticipated take-over zones and suggests alternative routes to avoid take-overs. These contributions are expected to reduce take-overs and improve the efficiency and safety of automated driving services.

Keywords
automated driving
clustering
dbscan
take-over
t-test

자율주행 기술이 발전되면서 우리나라에서도 자율주행 시범운행이 활발히 진행되고 있다. 현재 국내에서 운행 중인 자율주행차는 미국 자동차공학회(Society of Automotive Engineers) 기준 Level 3로, 자율주행 중 원활한 주행이 어려운 상황이 발생하면 운전석에 탑승한 Safety Driver가 통제권을 이양받아 수동주행을 하게 된다. 이렇게 자율주행에서 수동주행으로 전환하는 과정을 제어권전환(take-over)으로 정의한다. 제어권 전환은 현재 자율주행 운행의 안전성에 중요한 요소이지만, 제어권 전환의 발생 요인에 대한 연구는 충분하지 않은 상황이다. 본 연구는 제어권전환이 빈번하게 발생하는 구간을 조사하고 해당 구간들이 가지는 정적 요소를 식별한다. 이후 제어권 전환을 줄이는 방안을 제시하여 자율주행 서비스의 운영 효율성을 높이고자 한다. 이를 위해 국내 자율주행 운행데이터를 기반으로 분석 구간을 설정하고 제어권전환 지점을 군집화하여 제어권전환 다발구간을 선정하였다. 또한 제어권 전환 다발 구간과 비다발 구간 간의 정적 요인을 통계적으로 비교하여 제어권 전환에 영향을 미치는 요인을 도출하였다. 분석 결과, 제어권전환이 자주 발생하는 구간은 운행 지역 내 28개 존이었으며, 교차로거나 왕복 차로 수가 많은 존 및 횡단보도가 세 개 이상 포함되어 있는 존에서 제어권전환이 자주 발생하는 것으로 나타났다. 이러한 분석 결과를 바탕으로, 선정된 제어권전환 다발구간과 제어권전환의 정적 요인을 고려하여 제어권 전환 예상 구간에 대한 사전 준비 정보를 제공하는 서비스와 제어권 전환을 피할 수 있는 우회 경로를 제안함으로써 제어권 전환을 줄이는 데 기여할 수 있을 것으로 기대한다.

키워드
자율주행
군집분석
밀도 기반 공간 군집화
제어권전환
t-검정
References
1

Han H. S., Jeong S. H., Yang J. H. (2019), Comparison of Non-Driving Behaviors When Responding Take-Over Requests in Automated Driving, Trans. Korean Soc. Automot. Eng., 27(9), Korean Society of Automotive Engineers, 677-686.​

10.7467/KSAE.2019.27.9.677
2

Hong D., Kim S., Kim K. (2022), The Preliminary Study on Driver's Brain Activation During Take-Over Request of Conditional Autonomous Vehicle, J. Korea Comput. Graph. Soc., 28(3), Korea Computer Graphics Society, 101-111.​

10.15701/kcgs.2022.28.3.101
3

Hwang J., Kim A. (2023), Effect of Visual and Auditory Secondary Task on Takeover Performance in Level 2 Automated Driving, J. HCI Soc. Korea, 18(3), HCI Society of Korea, 19-28.​

10.17210/jhsk.2023.09.18.3.19
4

Jang M. J., La H. B., Park S. E., Jung A. R., Lee G. R., Oh K. S. (2023), Development of a Longitudinal Control Algorithm Considering the Driver Status and Driving Characteristics for the Safe Takeover of Autonomous Vehicles, Trans. Korean Soc. Mech. Eng. A., 47(7), Korean Society of Mechanical Engineers, 603-614.​

10.3795/KSME-A.2023.47.7.603
5

Kim H. S., Kim W. J., Kim J. S., Lee S. J., Kwon O. C., Yoon D. S. (2020), A Study on the Driver's Recognition and Behavioral Response of Control Authority Transition in Level 3 Automated Driving Vehicles, J. Ergon. Soc. Korea, 39(6), Ergonomics Society of Korea, 585-596.​

10.5143/JESK.2020.39.6.585
6

Kim J. W. (2020), A Study on the User Perception/Response of Multimodality Alerts in Take-Over of Automated Vehicles, [Master's thesis, Kookmin University], 1-75.​

7

Kim N., Yang M. Y., Lee J., Kim J. (2018), A Study on the Effect of Information Types on Drivers in Takeover Period of Automated Vehicles, J. Digit. Content Soc., 19(1), Digital Content Society, 113-122.​

8

Kim T. Y., Maeng J. Y., Kang K. P., Bae S. H. (2023), Spatial Factors' Analysis Affecting Automated Driving Safety Using Spatial Information Analysis Based on Level 4 ODD Elements, J. Korea Inst. Intell. Transp. Syst., 22(5), Korea Institute of Intelligent Transportation Systems, 182-199.​

10.12815/kits.2023.22.5.182
9

Korea Transportation Safety Authority (2011), Traffic Accident Information: High-Risk Roads for Accidents, Frequent Accident Spots, Total Accidents, 1-100.​

10

Kwon Y. D. (2022), A Comparative Study of Single-Auditory and Multi-Auditory Alarm in the Autonomous Driving System Deactivation Situation Due to Operational Design Domain Exit, [Master's thesis, Kookmin University], 1-80.​

11

Lee E. S., Shin C. H. (2024), Analysis of Factors Affecting Disengagement Using Actual Driving Data in Level 3 Autonomous Bus, J. Korean Soc. Intell. Transp. Syst., 23(5), Korean Society of Intelligent Transportation Systems, 308-321.​

10.12815/kits.2024.23.5.308
12

Lee Y. J., Lim D. H., Seo Y. H., Kim H. J. (2023), Development of Monitoring Indicators for Safe Autonomous Vehicle Operation: Focused on Pangyo Zero City, Traffic Saf. Res., 42(2), The Korean Society of Transportation, 49-76.​

13

Lim D. H., Seo Y. H., Kim J. H., So J. H., Kim H. J. (2021), Monitoring Indicators for Driving Safety of Autonomous Vehicle on Arterial Road: Focused on Pangyo Zero City, Int. J. Highway Eng., 23(3), Korean Society of Highway Engineers, 119-130.​

10.7855/IJHE.2021.23.3.119
14

Namkung K., Yoo C. M., Jeon H. M., Pan Y. H. (2020), Auditory User Interface Design Guidelines for Control Transitions in the Conditional Self-Driving Phase, J. Next-Gener. Converg. Inf. Serv. Technol., 9(1), Next Generation Convergence Information Services Technology Association, 49-59.​

10.29056/jncist.2020.03.05
15

Oh H. S. (2022), Effect of Non-Driving Related Task Situations on Takeover in Level 3 Automated Driving, [Master's thesis, Korea University], 1-90.

Information
  • Publisher :Korean Society of Transportation
  • Publisher(Ko) :대한교통학회
  • Journal Title :Journal of Korean Society of Transportation
  • Journal Title(Ko) :대한교통학회지
  • Volume : 43
  • No :2
  • Pages :145-160
  • Received Date : 2024-09-20
  • Revised Date : 2024-11-12
  • Accepted Date : 2025-03-04
  • DOI :https://doi.org/10.7470/jkst.2025.43.2.145
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