본문 바로가기
  • Home

Automatic Detection of the Access-restricted Sections based on Geographic Characteristics of At-grade Intersections in High Definition Maps

Lee Junhyung 1 SHIN CHI HYUN 2 Kim Youngjin 1 Kang Seungmin 3

1한국건설기술연구원
2경기대학교
3(주)유앤알텍

Accredited

ABSTRACT

This study focuses on the algorithm which can automatically detect the access restricted sections based on geographic characteristics of at-grade intersections in high definition maps. Some pre-works such as classification of intersection type, defining the geometrical boundary of intersection and measuring the threshold distance are necessary to detect access restricted section. The high definition map includes the lane, link, stop line and node information which are the base information to determine access-restricted sections. The geometrical boundary of intersection is defined as the cross, diverge and merge conflict area. The acceleration and deceleration lane is vertically separated from the mainline or channelized lane by the traffic island or painted island. The maximum length of the acceleration and deceleration lane is defined as 400m based on the merging and diverging influence area in highway capacity manual. The access-restricted section detection algorithm is designed based on the layer attribute data of a high definition map. The algorithm validation results show a 86.8% accuracy in the 61km long study site. Due to the high definition map error and complicated field conditions, there were some errors in the result. Nevertheless, the high definition map still the most suitable map to detect the access restricted section. The accuracy of this access restricted section detecting algorithm can be improved under the following conditions; Expand the coverage of high definition map, increase the accuracy of high definition map data and introduce a live high definition map.

Citation status

* References for papers published after 2023 are currently being built.