The Journal of The Korea Institute of Intelligent Transport Systems (한국ITS학회 논문지)

The Korea Institute of Intelligent Transport Systems (한국ITS학회)
권호 표지
DOI QR코드

DOI QR Code

A Study on the Development of Traffic Volume Estimation Model Based on Mobile Communication Data Using Machine Learning

머신러닝을 이용한 이동통신 데이터 기반 교통량 추정 모형 개발

  • Dong-seob Oh (Industry Promotion Office, Center for Supporting Innovative Business, ITS Korea) ;
  • So-sig Yoon (Traffic Bureau of Korea Police Agency) ;
  • Choul-ki Lee (Dept. of Transportation Systems Eng., Univ. of Ajou) ;
  • Yong-Sung CHO (Planning and Strategy Office, ITS Korea)
  • 오동섭 (한국지능형교통체계협회 산업진흥본부 기술지원센터) ;
  • 윤소식 (경찰청 교통국 ) ;
  • 이철기 (아주대학교 교통시스템공학과 ) ;
  • 조용성 (한국지능형교통체계협회 기획조정본부 )
  • Received : 2023.04.25
  • Accepted : 2023.07.04
  • Published : 2023.08.31
Download PDF
⟨ Previous Next ⟩

Abstract

This study develops an optimal mobile-communication-based National Highway traffic volume estimation model using an ensemble-based machine learning algorithm. Based on information such as mobile communication data and VDS data, the LightGBM model was selected as the optimal model for estimating traffic volume. As a result of evaluating traffic volume estimation performance from 96 points where VDS was installed, MAPE was 8.49 (accuracy 91.51%). On the roads where VDS was not installed, traffic estimation accuracy was 92.6%.

본 연구는 이동통신 로그 데이터를 통해 산출된 교통량 정보를 활용하여 기존 검지기에 준하는 교통량 정보를 추정하기 위해, 머신러닝의 앙상블 기법을 기반으로 하는 최적의 이동통신 기반 교통량 추정 모형을 개발하는 것이다. 이동통신 데이터를 통해 계측된 교통량 등의 정보와 VDS 실측 데이터를 활용하여 머신러닝 모형들을 통해 비교·분석한 결과, LightGBM 모형이 교통량 추정의 최적모형으로 선정되었다. 국도 1, 3, 6호선 검지영역 96개소를 대상으로 교통량 추정 모형의 성능을 평가한 결과, 전체 검지영역의 경우 MAPE 8.49로 교통량 추정 정확도가 91.51%로 분석되었다. VDS가 설치되지 않은 구간의 경우 교통량 추정 정확도는 92.6%로, VDS 설치가 어려운 구간에서도 LightGBM 교통량 추정 모형이 적용 가능하였다.

Keywords

Acknowledgement

본 논문은 오동섭의 박사학위로 아주대학교에 제출되었던 논문을 수정·보완하여 작성하였습니다.

References

  1. Caceres, N., Romero, Luis, M., Benitez, F. G. and Castillo, J. M.(2012), "Traffic Flow Estimation Models Using Cellular Phone Data", IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, vol. 13, no. 3, pp.1430-1441. https://doi.org/10.1109/TITS.2012.2189006
  2. Cheon, M. J., Choi, H. J., Park, J. W., Choi, H. Y., Lee, D. H. and Lee, O.(2021), "A study on the Traffic Prediction throught Catboost Algorithm", Journal of the Korea Academia-Industrial Cooperation Society, vol. 22, no. 3, pp.58-64. https://doi.org/10.5762/KAIS.2021.22.10.58
  3. Chung, Y. S., Nam, S. G. and Song, T. J.(2018), "A Travel Time Budget Estimation Using a Mobile Phone Signaling Data", Journal of he Korean Society of Civil Engineers, vol. 38, no. 3, pp.457-465.
  4. Jang, J. W., Park, C. S., Baek, N. C. and Lee, M. Y.(2011), "Analysis on Video Image Detection System Performance by Vehicle Speed", Journal of Korea Society of Transportation, vol. 23, no. 5, pp.105-112.
  5. Ji, B. S. and Hong, E. J.(2019), "Deep Learning Based Real Time Road Traffic Prediction Using LongTerm Evolution Access Data", Sensors, vol. 19, 5327, pp.1-16. https://doi.org/10.1109/JSEN.2018.2879233
  6. Kim, D. H. and Kim, S. G.(2002), "A study on Testing the Ability of Vehicle Detection System in ATMS", Journal of Korea Society of Transportation, vol. 20, no. 5, pp.231-244.
  7. Lee, E. J.(2020), A study on road speed of Seoul Gangnam-gu and Jung-gu predictions based on Machine Learning Methods, Ewha Womans University.
  8. Lee, H. M., Jeon, G. S. and Jang, J. A.(2019), "Predicting of the Severity of Car Traffic Accidents on a Highway Using Light Gradient Boosting Model", Journal of the Korea Institute of Electronic Communication Sciences, vol. 15, no. 6, pp.1123-1130.
  9. Lee, S. B.(2019), Development of Dynamic Traffic Volume Estimation Model Using DSRC Probe Data, Seoul National University.
  10. Lee, Y. H. and Hong, S. Y.(2019), "A Machine Learning Approach to the Prediction of Individual Travel Mode Choices", Journal of the Korea Data Information Science Society, vol. 30, no. 5, pp.1011-1015. https://doi.org/10.7465/jkdi.2019.30.5.1011
  11. Nam, M. W., Park, D. S., Jang, Y. J. and Lee, H. C.(2021), "Prediction of Traffic Accident Casualties Using Machine Learning", Korea Society of Computer Information, vol. 29, no. 1, pp.27-30.
  12. Oh, J. S.(2001), "Development of a Daily Traffic Volume Estimation Model Using Regression Analysis", Journal of the Korean Official Statistics, vol. 6, no. 2, pp.161-179.
  13. Snowdon, J., Gkountouna, O., Zufle, A. and Pfoser, D.(2018), Spatiotemporal Traffic Volume Estimation Model Based on GPS Samples, Association for Computing Machinery.
  14. Texas Transportation Institute(2002), VEHICLE DETECTOR EVALUATION, pp.1-77.
  15. Yaghoubi, F., Catovic, Armin., Gusmao, A., Pieczkowski, J. and Boros, P.(2021), Traffic Flow Estimation using LTE Radio Frequency Counters and Machine Learning, Cornell University, pp.1-9.
  16. Yoon, H. S.(2016), AADT estimation of connection road section in highway using large-scale GPS trip data, Seoul National University.