This study evaluated the short- and long-term prediction performances of a transformer-based trajectory-forecasting model for urban intersections. While a previous study focused on developing the basic structure of a transformer model for future trajectory prediction, the present study aimed to determine a practical prediction sequence length. To this end, multiple transformer models were trained with output sequence lengths ranging from 1 s to 10 s, and their performances were compared. The trajectory data used for training were generated through a microscopic traffic simulation, and the model accuracy was assessed using the metrics average displacement error (ADE) and final displacement error (FDE). The results demonstrate that the prediction accuracy decreases significantly when the output trajectory length exceeds 3 s. Specifically, straight-driving trajectories exhibit rapidly increasing errors, while turning trajectories maintained a relatively stable accuracy. In contrast, for turning-driving trajectories, prediction errors increased sharply during short-term forecasting, but the increase was more gradual in long-term forecasts. Additionally, the long-term prediction models produced higher errors even in the initial 1-second outputs, implying a tendency toward conservative inference under uncertain future scenarios. This conservative behavior is likely influenced by the model’s effort to minimize the overall loss across a broader prediction window, especially when trained with Smooth L1 loss function. This study provides practical insights into model design for edge-computing environments and contributes to the development of reliable short-term trajectory prediction systems for urban ITS applications.

미·중 기술패권 경쟁 가운데, 첨단 디지털 기술의 공급망과 데이터 수 집, 유출, 조작, 유포 등의 우려가 제기되면서 공급망안보와 데이터안보 가 교차하는 양상이 두드러지기 시작했다. 데이터안보나 공급망안보 각 각에 대한 기존 연구는 있었으나, 이 둘의 교차점에 초점을 맞춘 연구는 드물다. 이 연구에서 다루는 데이터안보 사안들은 정부뿐 아니라 개인이 나 민간 행위자들이 취급하는 데이터들이 국가안보적 사안으로 전환되는 모습을 보인다. 이러한 신흥위협이 공급망위협과 상호작용하며, 국가안보 적 우려로 비화한 사례로 화웨이의 5G, ZPMC 항만 대형 크레인, 미래 자동차의 핵심 기술 라이다, 중국산 DJI 드론 등이 있다. 본 연구는 이러 한 사례들에 대해 국가 차원에서 어떤 대응이 이루어졌는지 탐구하였다. 특히, 관련 이슈들을 안보화하고 대응책을 앞서 제시하고 있는 미국에 초점을 맞추었다. 또한 기존의 안보화 이론을 보완하여 정치경제적 측면 을 추가하였다. 행정명령, 전략서 발간, 연방 예산안 그리고 새로운 전문 기구들의 창설을 통해 신흥 위협에 대처하고 있으며, 그 과정에서 공급 망 재편, 표준, 규범, 규제의 설립을 통해 변화하는 국가의 안보 제공자 로서의 역할을 조명하였다.

긴급차량 골든타임 확보를 위해 지방자치단체에서는 ‘긴급차량 우선신호’ 시스템을 도입하고 있다. 그러나 긴급차량 우선신호로 인하 여 일반차량의 지체시간은 증가하게 되고 기존 신호주기로 바로 복귀하여 교차로 전체의 지체를 유발할 것으로 예상된다. 해당 지체 를 해결하기 위해 일반차량을 고려한 ‘회복신호 산정’ 연구가 수행되고 있다. 그러나 현재까지 교차로 유형별로 적절한 회복신호는 무 엇인지에 대한 연구는 부족한 상황이다. 본 연구에서는 교차로를 유형별로 구분하여 일반차량 지체도를 감소시키기 위한 회복신호의 필요성을 검증하고자 한다. 우선신호의 경우 긴급차량이 교차로를 통과하기까지 걸리는 시간을 계산하여 부여하였다. 회복신호의 경우 는 우선신호에서 부여한 시간만큼 일반차량에게 보상하는 신호방식을 적용하였으며, 도입 효과를 SUMO 교통 시뮬레이션을 통해 비 교 분석하였다.

PURPOSES : This study aims to investigate the reliability of the real-time estimation of intersection traffic volumes based on the integration of location intelligence data and smart intersection data. METHODS : Location intelligence data (LID) and smart intersection data were obtained at eight intersections in Inju-daero, Incheon. The two datasets were then integrated to estimate traffic volumes for intersections in the shadow section, where traffic information was not expected to be obtained. The traffic estimation accuracy was evaluated using the total traffic, approach traffic, and turning movement volumes at the intersections. The estimated traffic was compared with the actual traffic volumes in the smart intersection data to validate the reliability of traffic estimation. RESULTS : The average traffic estimation error for the total intersection volume was approximately 4.5% for the five intersections in the shadow section. The estimation errors for the approach volumes (less than 5%) were also consistently low, except from 12 pm to 1 pm. CONCLUSIONS : The findings of this study suggest that location intelligence data can be combined with smart intersection data to estimate real-time traffic for shadow sections on roadways. This could enable a cost-effective cooperative intelligent transport system (C-ITS) when the municipal budget is limited, ultimately leading to the sustainable operation of C-ITS.

PURPOSES : This study evaluates the effectiveness of traffic flow optimization when giving safety strategy guidance to a connected autonomous vehicle (CAV) based on information received through infrastructure cooperation in a V2X environment for non-signal intersection. METHODS : To evaluate the effectiveness of safety strategy guidance based on developed traffic flow control algorithm at a non-signalized intersection, it was implemented on simulation. A scenario based on the Level of Service (LOS) and the market penetration rate(MPR) of autonomous vehicles was established. The simulation results were divided into safety, operation, and environment to evaluate the effect, and the effect of optimizing traffic flow was finally derived through the integrated evaluation score. RESULTS : As a result, when safety strategy guidance was provided, the number of conflicts and CO emissions decreased by about 29% and about 15%, improving safety and environmental performance. In the case of operation, the mean of delay time was increased overall by 1%, but in the case of MPR 50 and above, the delay time was reduced by about 38%, thereby increasing operation. Finally, the aspect of traffic flow optimization, effectiveness of safety strategy guidance was derived through the integrated evaluation score, and the average integrated evaluation score improved from MPR 20 or higher. CONCLUSIONS : Providing guidance had the effect of optimizing traffic flow at a non-signal intersection. In the future, V2X communications will provide CAV with algorithm-based guidance developed in this study to control driving behavior. it will support safe and efficient driving at non-signal intersections.

PURPOSES : In the case of a turning maneuver at an at-grade intersection or changing the driving path, the trajectory of a vehicle with a long body, such as a large bus or an articulated bus, should be analyzed from the perspective of road design. In this study, an articulated bus was selected to analyze the off-tracking, swept path width, and lane encroach hment for vehicle turning. METHODS : In this study, four scenarios were developed for right- and U-turn situations. For the right-turn situation, cases were divided into radii of 15 m (Scenario 1) and 40 m (Scenario 2). For the U-turn situation, the cases were analyzed based on a U-turn after stopping at the stop line (Scenario 3) and without stopping at the stop line for the U-turn (Scenario 4). Each scenario was examined at 5° (Right-turn) and 10° (U-turn) angles to analyze the off-tracking, swept path width, and lane encroachment. In addition, four Global Positioning System (GPS) antennas were installed on top of the articulated bus to obtain the driving trajectory of the vehicle. GPS locational reference points were marked on the testing ground to improve positioning accuracy. RESULTS : As a result of the right-turn analysis at an intersection radius of 15 m (Scenario 1), the average off-tracking per angle was 1.04 m, the average swept path width was 3.89 m, and the lane encroachments occurred at an angle of 65° to 70°. For the right-turn analysis at an intersection radius of 40 m (Scenario 2), the average off-tracking per angle was 3.71 m, and the average swept path width was 3.31 m. Unlike the results for the 15-m radius, no lane encroachment was found. Furthermore, the averages of the off-tracking in the at-grade intersection U-turn situation were 2.65 m (Scenario 3) and 2.54 m (Scenario 4), and the average swept path width was 6.15 m. CONCLUSIONS : The required driving width when an articulated bus performs a turning maneuver at an at-grade intersection was analyzed, revealing the implications that must be considered for busway design.

PURPOSES : In this study, we examined the installation and the effect of the s-type approach lane marking near the stop line of a typical intersection access road. We examined the possibility of installation and standardization of this facility and its impact on vehicular speed management and carbon emission reduction. METHODS : To review the installation and standardization possibilities, the geometric size of the marking was set. The possibility of standardization was examined by applying it to lane markings. The velocity before and after the installation of the marking was compared and analyzed through the velocity estimation equation to assess the impact on speed management. Carbon emissions were estimated by comparing the emissions before and after applying the marking. RESULTS : The s-type approach lane marking can be installed near the stop line of the intersection access road. It was possible to standardize the lane marking by suggesting a formula to determine the size of the geometry. Additionally, the marking enabled vehicular speed management and improvement in the carbon yield. The marking decreased speeds by approximately 10 km/h, from the original speed of 36 km/h to 25.5 km/h after installation. The standard deviation per vehicle was reduced by approximately 5.9 km/h, from 5.8 km/h to 0.9 km/h. Additionally, carbon emissions decreased by 17%, from 14.1 g/40 m to 11.7 g/40 m. CONCLUSIONS : The geometry and size of the lane marking installation can be set near the stop line of the approach road. Standardization of this facility was also possible. The s-type lane marking, installed at the stop line of the approach road, has the potential to control the speed, reduce the acceleration or deceleration, and reduce the carbon emission. In the future, it is expected that such lane markings can be applied to multi-faceted areas

PURPOSES : The purpose of this study is to develop models of accidents occurring at circular intersections related to the time of day and night and driver gender, and to provide countermeasures for safer circular intersections. METHODS: Seventy intersections built before 2008 were surveyed for inclusion in the modeling. Traffic accident data from 2008 to 2014 were collected from the TAAS data set of the Road Traffic Authority. Sixteen variables explaining the accidents including geometry and traffic volume were selected from the literature and seven multiple linear regression models were developed using SPSS 20.0. RESULTS: First, the null hypotheses, that the number of traffic accidents are not related to driver gender or time of day, were rejected at a 5% level of significance. Second, seven statistically significant accident models with R2 value of 0.643-0.890 were developed. Third, in daytime models by gender, when the right-turn-only lane was selected as the common variable, the number of lanes, presence of driveways and speed humps, diagrammatic exit destination sign, and total entering traffic volume were evaluated as specific variables. Finally, in nighttime models by gender, when the diagrammatic exit destination sign was selected as the common variable, total entering traffic volume, presence of right-turnonly lanes, number of circulatory road way lanes, and presence of splitter islands and driveways were identified as specific variables. CONCLUSIONS: This study developed seven accident models and analyzed the common and specific variables by time of day and gender. The results suggest approaches to providing countermeasures for safer circular intersections.

PURPOSES : The purpose of this study is to present a linear programing optimization model for the design of lane-based lane-uses and signal timings for an isolated intersection. METHODS: For the optimization model, a set of constraints for lane-uses and signal settings are identified to ensure feasibility and safety of traffic flow. Three types of objective functions are introduced for optimizing lane-uses and signal operation, including 1) flow ratio minimization of a dual-ring signal control system, 2) cycle length minimization, and 3) capacity maximization. RESULTS : The three types of model were evaluated in terms of minimizing delay time. From the experimental results, the flow ratio minimization model proved to be more effective in reducing delay time than cycle length minimization and capacity maximization models and provided reasonable cycle lengths located between those of other two models. CONCLUSIONS : It was concluded that the flow ratio minimization objective function is the proper one to implement for lane-uses and signal settings optimization to reduce delay time for signalized intersections.

PURPOSES : Before-and-after studies of red light cameras were conducted with the aim of reducing the number of side right-angle collisions. Three different methods were used for the before-and-after studies, and the analysis results were compared. METHODS: This research used the naive before-and-after method, the comparison-group method, and the empirical Bayes method to study the effects of red light cameras on side-angle collisions. The results of the three before-and-after methods were compared and interpreted in terms of safety indications at signalized intersections. RESULTS: The research results showed that side right-angle collisions can be reduced by installing red light cameras at signalized intersections. All three methods guarantee safety improvements of 25~30% on average. With regard to the results of each method, the naive before-and-after method, the comparison-group method, and the empirical Bayes method showed safety improvements of 25.6%, 27.8%, and 29.7%, respectively. CONCLUSIONS: It was concluded that red light cameras are an effective countermeasure to improve intersection safety. In particular, by installing red light cameras, side right-angle collisions can be reduced by up to approximately 25~30%.

PURPOSES: Recently, there are increasing bicycle accidents along with increasing bicycles users. Bicycle accidents occurred frequently by perpendicular collision form at intersection inner. In order to improve safety of bicycle, drivers need to be aware of bicycles on the road and intersection geometric designs need to be designed to reduce risk associated with collisions between bicycles and car. This study aims to review the location of bicycle crossing in the viewpoint of bicycle safety. METHODS: Four types of bicycle crossing by curve radius and driver's check around the behavior are set to simulate the risk of collisions between bicycles and car turning right. Simulation using fortran programming are conducted on total 60 cases. RESULTS : Bicycle crossing located behind of crosswalk is lower the risk of collisions with car in all cases. In addition to the larger curve radius of pavement edge at intersection and the more pay attention to the rear by the turn head to the right is too low the risk of collisions with car. CONCLUSIONS : It is show that the location of bicycle crossing is safer behind than in front of crosswalk in the viewpoint of bicycle safety.

PURPOSES : The purpose of this study is to propose delay-minimizing operation methodology of a signalized intersection based upon optimization of lane-uses on approaching lanes for an intersection. METHODS: For the optimization model of lane-uses, a set of constraints are set up to ensure feasibility and safety of the lane-uses, traffic flow, and signal settings. Minimization of demand to saturation flow ratio of a dual-ring signal control system is introduced to the objective function for delay minimization and effective signal operation. Using the optimized lane-uses, signal timings are optimized by delay-based model of TRANSYT-7F. RESULTS : It was found that the proposed objective function is great relation with delay time for an intersection. From the experimental results, the method was approved to be effective in reducing delay time. Especially, cases for two left-turn lanes reduced greater delays than those for a left turn lane. It is noticed that the cases for different traffic volume by approach reduced greater delays than those for the same traffic volume by approach. CONCLUSIONS : It was concluded that the objective function is proper for lane-uses optimizing model and the operation method is effective in reducing delay time for signalized intersections.

PURPOSES : The present paper is to compare vehicles' CO2 emissions in roundabouts and signalized intersections. METHODS : The present paper uses the SIDRA software with variables of traffic and road conditions. RESULTS : The results of the study are as follows : First, when entering traffic volumes are more than 1600pcph, vehicle's CO2 emissions in roundabouts are lower than those of signalized intersections regardless of the left turn ratio. Second, When entering traffic volumes are more than 2800pcph, vehicles's CO2 emissions in 2-lane approaches are lower than those of 1-lane approaches in signalized intersection. Third, when entering traffic volumes are more than 1600pcph, vehicle's CO2 emissions of CASE B are lowest. (CASE B is the condition with one exclusive left-turn lane and one exclusive straight lane and one shared straight lane with right-turn.) Also, CASE A is the condition that vehicle's CO2 emissions in roundabouts are lower than those of signalized intersections between 1600pcph and 3600pcph. (CASE A is the condition with one exclusive left-turn lane and one shared straight lane with right-turn.) But, when entering traffic volumes are more than 4000pcph, vehicle's CO2 emissions in signalized intersections is lower than those of roundabouts. CONCLUSIONS : It may be concluded that vehicle's CO2 emissions on roundabouts are much lower than those of signalized intersections, especially, when entering traffics volumes are between 1600pcph and 3600pcph in 1-lane or 2-lane approaches.

PURPOSES : Plastic deformation is frequently made in intersection asphalt pavement at its early age due to deceleration and stoppage of vehicles. This study has been performed to provide a mechanistic basis for reasonable selection of paving method to minimize the plastic deformation at intersection. METHODS : Pavement layer, temperature, traffic volume of the intersections managed by the Daejeon Regional Construction and Management Administration were collected to calculate asphalt dynamic modulus with pavement depth by using a prediction equation suggested by the Korean pavement design guide. Performance of ordinary dense-graded asphalt pavement, polymer modified asphalt pavement, and fiber reinforced asphalt pavement was analyzed by finite element method and the results were used in a performance model to predict the plastic deformation. RESULTS : In aspect of performance, the three paving methods were usable under low traffic while the fiber reinforced asphalt pavement was the most suitable under heavy traffic. CONCLUSIONS : Reasonable paving method suitable for traffic characteristics in the intersection might be decided by considering economic feasibility.