DOI QR코드

DOI QR Code

A Design of the Vehicle Crisis Detection System(VCDS) based on vehicle internal and external data and deep learning

차량 내·외부 데이터 및 딥러닝 기반 차량 위기 감지 시스템 설계

  • Son, Su-Rak (Catholic Kwandong University, Department of Computer Engineering) ;
  • Jeong, Yi-Na (Catholic Kwandong University, Department of Software)
  • Received : 2021.03.02
  • Accepted : 2021.03.23
  • Published : 2021.04.30

Abstract

Currently, autonomous vehicle markets are commercializing a third-level autonomous vehicle, but there is a possibility that an accident may occur even during fully autonomous driving due to stability issues. In fact, autonomous vehicles have recorded 81 accidents. This is because, unlike level 3, autonomous vehicles after level 4 have to judge and respond to emergency situations by themselves. Therefore, this paper proposes a vehicle crisis detection system(VCDS) that collects and stores information outside the vehicle through CNN, and uses the stored information and vehicle sensor data to output the crisis situation of the vehicle as a number between 0 and 1. The VCDS consists of two modules. The vehicle external situation collection module collects surrounding vehicle and pedestrian data using a CNN-based neural network model. The vehicle crisis situation determination module detects a crisis situation in the vehicle by using the output of the vehicle external situation collection module and the vehicle internal sensor data. As a result of the experiment, the average operation time of VESCM was 55ms, R-CNN was 74ms, and CNN was 101ms. In particular, R-CNN shows similar computation time to VESCM when the number of pedestrians is small, but it takes more computation time than VESCM as the number of pedestrians increases. On average, VESCM had 25.68% faster computation time than R-CNN and 45.54% faster than CNN, and the accuracy of all three models did not decrease below 80% and showed high accuracy.

현재 자율주행차량 시장은 3레벨 자율주행차량을 상용화하고 있으나, 안정성의 문제로 완전 자율주행 중에도 사고가 발생할 가능성이 있다. 실제로 자율주행차량은 81건의 사고를 기록하고 있다. 3레벨과 다르게 4레벨 이후의 자율주행차량은 긴급상황을 스스로 판단하고 대처해야 하기 때문이다. 따라서 본 논문에서는 CNN을 통하여 차량 외부의 정보를 수집하여 저장하고, 저장된 정보와 차량 센서 데이터를 이용하여 차량이 처한 위기 상황을 0~1 사이의 수치로 출력하는 차량 내·외부 데이터 및 딥러닝 기반 차량 위기 감지 시스템을 제안한다. 차량 위기 감지 시스템은 CNN기반 신경망 모델을 사용하여 주변 차량과 보행자 데이터를 수집하는 차량 외부 상황 수집 모듈과 차량 외부 상황 수집 모듈의 출력과 차량 내부 센서 데이터를 이용하여 차량이 처한 위기 상황을 수치화하는 차량 위기 상황 판단 모듈로 구성된다. 실험 결과, VESCM의 평균 연산 시간은 55ms 였고, R-CNN은 74ms, CNN은 101ms였다. 특히, R-CNN은 보행자수가 적을 때 VESCM과 비슷한 연산 시간을 보이지만, 보행자 수가 많아 질수록 VESCM보다 많은 연산 시간을 소요했다. 평균적으로 VESCM는 R-CNN보다 25.68%, CNN보다 45.54% 더 빠른 연산 시간을 가졌고, 세 모델의 정확도는 모두 80% 이하로 감소하지 않으며 높은 정확도를 보였다.

Keywords

References

  1. Self-driving car accident, there were many collisions with general cars that followed, Young-joo Kim, Available online: https://news.joins.com/article/23718449
  2. Ji Hoon Lee, Dae Youb Kim, "A Study on Low-Overhead Collision Warning Scheme using Vehicle-to-Vehicle Communications", Journal of Korea Multimedia Society, Korea Multimedia Society, Vol. 15, No.10, pp. 1221-1227, 2012 https://doi.org/10.9717/kmms.2012.15.10.1221
  3. Sang Jun Park, Kwan Joong Kim, "A study of design mechanism for the alerting car accident", Journal of Korea Academy Industrial Cooperation Society, Korea Academy Industrial Cooperation Society, Vol.12, No.11, pp.5272-5277, 2011 https://doi.org/10.5762/KAIS.2011.12.11.5272
  4. Donghoon Shin, Kyongsu Yi, Yeonhwan Jeong, "Human Factor Considered Risk assessment of Automated Vehicle through Vehicular Communication", Proceedings of the 2017 Fall Conference of the Korean Society of Mechanical Engineers, The Korean Society of Mechanical Engineers, pp. 1563-1568, 2017
  5. Kim Byeong Su, Noh Jun Ho, Park So Young, "Design of Automatic Risky Situation Detection Model based on Real-Time Driving Information", Proceedings of Symposium of the Korean Institute of communications and Information Sciences, Korea Institute Of Communication Sciences, pp. 783-784, 2017
  6. Geun Hyung Min, Woo Seok Kim, Jun Sang Cho, Heung Bae Gil, "Vehicular Collision Risk Assessment on the Highway Bridges in South Korea", Journal of the Korea Institute for Structural Maintenance and Inspection, The Korea Institute for Structural Maintenance and Inspection, Vol.20, No.5, pp.009-017, 2016
  7. MiRa Jeong, Byoung Chul Ko, Jae Yeal Nam, "Detection of sudden pedestrian crossings using the thermal camera installed in the vehicle", Proceedings of the 2014 Spring Conference of the Korean Institute of Information Scientists and Engineers, The Korean Institute of Information Scientists and Engineers, pp.668-670, 2014