오늘날, 단거리 이동에 편리하고 접근성이 좋아 많은 사람들이 개인형 이동장 치(personal mobility, PM)를 이용하고 있다. 그러나, 이용자가 빠르게 증가함에 따라 PM 관련 교통사고 또한 급증하였다. 최근 5년간(2017~2021년) 전체 교통 사고는 1.6% 감소하였지만, PM 교통사고는 연평균 96.2% 증가하여 PM 안전성 에 대한 우려가 커지고 있다. 안전모 착용 의무와 ‘제2종 원동기장치자전거’ 이상 의 면허 의무 등을 포함하여 PM 안전 규정이 강화되었지만, 이 규정들이 PM에 대한 이해가 부족한 규제로 비판받고 있다. 따라서, PM 이용을 활성화하고 PM의 안전성을 높이기 위해 효율적인 교통안전 대책을 마련해야 할 필요가 있다. 본 연구에서는 PM 교통사고의 특성을 분석하고, PM 교통사고에 영향을 미치 는 요인을 파악하여 공학적으로 PM 안전대책을 마련하는 것을 목표로 수행되 었다. 최근 6년간 발생한 PM 교통사고 자료를 분석하여 통계적 시사점을 도출 하였고, 순서형 로지스틱 회귀분석을 수행하여 PM 교통사고의 심각도에 영향을 주는 요인들을 도출하였다. 최종적으로 수도권의 행정구역별 PM 교통사고 건수 를 예측하는 PM 교통사고 예측모형을 개발하였다. PM 교통사고만의 특징이 반 영하여 모형을 설계하였으며, 성능 평가 지표를 통해 예측모형의 정확도를 검증 하였다. PM 교통사고 자료의 기초통계분석과 회귀분석을 통해 PM 단독의 사고, 고령 의 PM 운전자, 측면으로 충돌한 사고가 PM 사고의 심각도를 높인다고 파악되 었다. PM 교통사고 예측모형은 면적 대비 인구수와 40세 미만 인구 비율, 자동 차 교통사고 건수를 독립변수로 하여 행정구역별 인구 천 명당 PM 사고 건수 를 예측하였다. 성능 지표 분석을 통해 제시된 예측모형의 성능이 좋다고 평가 하였다. 본 연구의 교통사고 예측모형은 새로운 이동 수단인 PM의 교통사고에 초점을 맞추어 개발되었으며, 지역별 PM 교통사고의 안전 등급 제시 등 다양하 게 PM 안전대책 마련에 활용될 것으로 판단된다.|In recent times, the convenience and accessibility of personal mobility (PM) have resulted in a significant increase in its usage for short-distance travel. However, with the rapid growth in PM users, the number of PM-related accidents has also risen. While overall traffic accidents decreased by 1.6% over the last five years (2017~2021), PM accidents have surged at about 96.2% per year, raising concerns about PM safety. Despite efforts to strengthen PM regulations, including helmet-wearing and licensure obligations, these measures have been criticized for their lack of understanding of PMs. Hence, it is crucial to establish efficient safety measures to prevent PM accidents and to activate the use of PMs._x000D_
<br>The objective of this study was to analyze the characteristics of PM accidents and accident factors, allowing for the engineering of effective PM safety strategies. By analyzing PM accident data from the past six years, statistical insights were derived, and ordinal logistic regression analysis was conducted to identify factors impacting the severity of PM accidents. Additionally, a PM accident prediction model for administrative regions in the metropolitan area was developed, considering PM-specific features, and validated for accuracy through performance metrics._x000D_
<br>Based on the basic statistical analysis and regression analysis of PM accident data, it was found that single PM accidents, elderly PM drivers, and side collisions contribute to increased severity of PM accidents. The PM accident prediction model utilized variable such as the population density, the percentage of people under the age of 40, and the count of car accidents, as independent factors to forecast the number of PM accidents per thousand inhabitants for each administrative region. The model demonstrated good performance based on the evaluation metrics. This research's accident prediction model focuses on the unique challenges posed by PM as a novel mode of transportation, providing valuable insights for formulating diverse PM safety strategies.
Alternative Abstract
In recent times, the convenience and accessibility of personal mobility (PM) have resulted in a significant increase in its usage for short-distance travel. However, with the rapid growth in PM users, the number of PM-related accidents has also risen. While overall traffic accidents decreased by 1.6% over the last five years (2017~2021), PM accidents have surged at about 96.2% per year, raising concerns about PM safety. Despite efforts to strengthen PM regulations, including helmet-wearing and licensure obligations, these measures have been criticized for their lack of understanding of PMs. Hence, it is crucial to establish efficient safety measures to prevent PM accidents and to activate the use of PMs._x000D_
<br>The objective of this study was to analyze the characteristics of PM accidents and accident factors, allowing for the engineering of effective PM safety strategies. By analyzing PM accident data from the past six years, statistical insights were derived, and ordinal logistic regression analysis was conducted to identify factors impacting the severity of PM accidents. Additionally, a PM accident prediction model for administrative regions in the metropolitan area was developed, considering PM-specific features, and validated for accuracy through performance metrics._x000D_
<br>Based on the basic statistical analysis and regression analysis of PM accident data, it was found that single PM accidents, elderly PM drivers, and side collisions contribute to increased severity of PM accidents. The PM accident prediction model utilized variable such as the population density, the percentage of people under the age of 40, and the count of car accidents, as independent factors to forecast the number of PM accidents per thousand inhabitants for each administrative region. The model demonstrated good performance based on the evaluation metrics. This research's accident prediction model focuses on the unique challenges posed by PM as a novel mode of transportation, providing valuable insights for formulating diverse PM safety strategies.
