• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2018.06a
• /
• pp.332-333
• /
• 2018

본 논문에서는 딥러닝 기반의 주행 차로 인식 기법을 활용한 차선 변경 검출 기술을 제안한다. 제안한 방법은 주행 차로, 좌우 차로, 차량 등 3 종의 이미지 데이터를 학습, 검증, 실험 데이터로 나눠 활용하였다. 주행 차로 및 차선 변경 인식을 위하여 변형된 AlexNet 모델을 개발하였다. 실험 결과 주행 차로 69.45%, 좌우 차로 66.9%, 차량 76.4%의 인식률 결과를 보여 기존 패턴인식 방법과 비교하여 우수한 결과를 보였다.

• Journal of Korean Society of Transportation
• /
• v.35 no.4
• /
• pp.278-291
• /
• 2017

Effective installation and operations of traffic safety signs in the advance warning area is of considerable interest for safer traffic management at freeway work zones. This study evaluated the feasibility of traffic safety sign installation based on a driving simulation experiment to take drivers' responsive characteristics into consideration. Both the compliance rate for speed limit and a lane-changing safety index were used for evaluating a set of driving simulation scenarios. It was identified that providing speed limit signs to reduce speed in advance and afterward guiding lane-changing could lead to safer maneuvering of vehicles entering work zones. The outcomes of this study are expected to be useful for revising freeway work zone management policies toward crash prevention.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.22 no.1
• /
• pp.276-290
• /
• 2023

Several efforts in Korea are currently underway with the goal of commercializing autonomous vehicles. Hence, various studies are emerging on autonomous vehicles that drive safely and quickly according to operating guidelines. The current study examines the path search of an autonomous vehicle from a microscopic viewpoint and tries to prove the efficiency required by learning the lane change of an autonomous vehicle through Deep Q-Learning. A SUMO was used to achieve this purpose. The scenario was set to start with a random lane at the starting point and make a right turn through a lane change to the third lane at the destination. As a result of the study, the analysis was divided into simulation-based lane change and simulation-based lane change applied with Deep Q-Learning. The average traffic speed was improved by about 40% in the case of simulation with Deep Q-Learning applied, compared to the case without application, and the average waiting time was reduced by about 2 seconds and the average queue length by about 2.3 vehicles.

• Journal of Korean Society of Transportation
• /
• v.35 no.4
• /
• pp.332-347
• /
• 2017

Analysis of the interaction between the automated vehicles and manual vehicles is very important in analyzing the performance of automated cooperative driving environments. In particular, the automated vehicle platooning can affect the driving behavior of adjacent manual vehicles. The purpose of this study is to analyze the lane change behavior of the manual vehicles in automated vehicle platonning environment and to conduct the experiment and questionnaire surveys in three stages. In the first stage, a video questionnaire survey was conducted, and responsive behaviors of manual vehicles were investigated. In second stage, the driving simulator experiments were conducted to investigate the lane change behaviors of in automated vehicle platonning environments. To analyze the lane change behavior of the manual vehicles, lane change durations and acceleration noise, which are indicators of traffic flow stability, were used. The driving behavior of manual vehicles were compared across different market penetration rates (MPR) of automated vehicles and human factors. Lastly, NASA-TLX (NASA Task Load Index) was used to evaluate the workload of the manual vehicle drivers. As a result of the analysis, it was identified that manual vehicle drivers had psychological burdens while driving in automated vehicle platonning environments. Lane change durations were longer when the MPR of the automated vehicles increased, and acceleration noise were increased in the case of 30-40 years old or female drivers. The results from this study can be used as a fundamental for more realistic traffic simulations reflecting the interaction between the automated vehicles and manual vehicles. It is also expected to effectively support the establishment of valuable transportation management strategy in automated vehicle environments.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.16 no.5
• /
• pp.1-11
• /
• 2017

Expressway network which plays an important role on land transportation system, have been developing quantitatively and qualitatively with $7{\times}9$ structure. To cope with complex geometric condition, guidance lane marking has been installed to induce safer lane-changing maneuver. However, there is no standard on guidance lane marking and its effectiveness is also verified with limited scope. The major purpose of this research is to clarify its effectiveness in terms of driving safety aspect using virtual driving simulator and to suggest standard on the proper length. To carry this out, preference data from subjects was collected and lane-changing pattern within virtual driving environment was investigated. In addition, in order to quantify the level of comfort, Electroencephalogram data was collected and validated using statistical test. Finally, it is expected that this research can be used to establish standard on guidance lane marking.

• Proceedings of the KOR-KST Conference
• /
• 2007.11a
• /
• pp.557-564
• /
• 2007

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.39 no.2
• /
• pp.273-279
• /
• 2019

Development of mountain area is increasing due to the demand for improvement of traffic convenience and development of underdeveloped area. Therefore, there frequently are sections where tunnels and interchanges are located close to each other. These sections do not only affect tunnel planning, types and length of interchanges, but also affect more on route selection. In Korea, several design criteria present each reference value but these values are very similar. In the situation, the minimum value among them is usually applied when planning roads and it could cause traffic safety problems in different site conditions. In this study, the problems of design speed, illuminance adaptation distance, and lane change intervals are analyzed by simulating the cases that the problem could occur when calculating the separation distance between tunnel and interchange. The results obtained from this study can be summarized as following: the driving speed should be applied in case that the site has a big gap between design speed and driving speed because the uniform application of the design speed is not safe; the illuminance adaptation distance should include the influence distance in the section affected by the direct light; in addition, the lane change distance should include the time to perceive the situation of the next lane after the lane change in the section required for successive lane change.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.19 no.3
• /
• pp.38-51
• /
• 2020

A bottleneck and congestion occur when a freeway is closed due to maintenance and construction activities on the freeway. Although various traffic managements have been developed to improve the traffic efficiency at freeway work zones, such as merge control, there is a limit to those controls with human drivers. On the other hand, the wireless communication of connected and automated vehicles (CAVs) enables the operation of advanced traffic management. This study developed a traffic control strategy that integrates Dynamic Merge Control (DMC) and Lane Change Control (LCC) in a CAV environment. DMC operates as an either early or late merge based on the occupancy rate of upstream of the work zone. The LCC algorithm determines the number of vehicles that need to change their lane to balance the traffic volume on open lanes. The simulation results showed that integrated control improves the cumulative vehicle count, average speed upstream, and average network travel time.

• Journal of Korean Society of Transportation
• /
• v.20 no.2
• /
• pp.93-104
• /
• 2002

The purpose of this study was to develop a improved high-order continuum model among macroscopic traffic flow models. This study was mainly performed for uninterrupted flow. In the first step, the proposed model described traffic flow at dropped lane. (no exits) It was possible to describe the traffic flow during short-term considering lane change. The proposed model was based on Payne's model. Our model was newly applied to uninterrupted traffic flow in consideration of geometry condition and driver behavior. It is possible to establish efficient control strategies, simulation and assess the effects of geometric improvements using this model. This model was simulated with field data for the actual adaption. The results of the model tests, traffic volume and density is suitably represented. we think that the results in the article can be led to predicting the situation in the near future.

• Journal of Korean Society of Transportation
• /
• v.28 no.4
• /
• pp.177-186
• /
• 2010

The principal objective of this paper is to develop road design guidelines, especially for managed lane access spacing between the expressway on-ramp (or off-ramp) and managed lane access point. Managed lanes are typically located in the expressway median and are accessed by weaving across the mainlines. The high level of lane-changing activity present in weaving areas affects capacity significantly. One promising tool for the analysis of lane-changing activity is "gap acceptance theory." This paper estimates the capacity of weaving areas based on the estimated degree of traffic turbulence using gap acceptance theory. The degree of traffic turbulence is represented by a function of the probability that lane-changing vehicles can complete their maneuvers successfully in a given weaving distance. In developing road design guidelines based on the developed gap acceptance model, the minimum managed lane access spacing is determined where the capacity with respect to the managed lane access spacing becomes stable.