• International Journal of Internet, Broadcasting and Communication
• /
• v.16 no.1
• /
• pp.63-71
• /
• 2024

Deep reinforcement learning (RL) significantly influences autonomous vehicle development by optimizing decision-making and adaptation to complex driving environments through simulation-based training. In deep RL, an activation function is used, and various activation functions have been proposed, but their performance varies greatly depending on the application environment. Therefore, finding the optimal activation function according to the environment is important for effective learning. In this paper, we analyzed nine commonly used activation functions for RL to compare and evaluate which activation function is most effective when using deep RL for autonomous vehicles to learn highway merging. To do this, we built a performance evaluation environment and compared the average reward of each activation function. The results showed that the highest reward was achieved using Mish, and the lowest using SELU. The difference in reward between the two activation functions was 10.3%.

• International Journal of Highway Engineering
• /
• v.10 no.2
• /
• pp.221-227
• /
• 2008

This study identifies the relationship between traffic data quality obtained from loop detectors and their location. Traffic data basically shows traffic flow conditions and thus, these information can be used as inputs for various transportation management strategies. Out study presents how to determine optimal downstream detector location on merging area in order to enhance the effects of ramp metering strategies. Microscopic simulation model, PARAMICS, is used as the main analytical tool. Assuming that detector location relies heavily on traffic flow characteristics in each roadway segment, we perform statistical analysis to identify homogeneous traffic conditions on merging area.

• International Journal of Highway Engineering
• /
• v.16 no.1
• /
• pp.99-109
• /
• 2014

PURPOSES : Climbing lanes on expressways managed by the Korea Expressway Corporation (KEC) have been hot potatoes due to conflicts between slow-moving vehicles such as trucks and other vehicles at the merging section as well as the less popularity with the slow-moving vehicles. In order to resolve such problems, KEC has altered existing climbing lanes to passing-type climbing lanes in 1999. The new type of climbing lanes showed an apparent improvement in mobility. For example, the speeds of vehicles using both climbing lane and other lanes improved a lot. However, there has been no clear evidence about improved safety. METHODS : This research effort was initiated to evaluate the safety of the new passing-type climbing lanes using the comparison-group(CG) method based on three-year-long traffic accident data sets before and after the change, respectively. RESULTS : The passing-type climbing lanes showed twice increased traffic accidents even though the traffic accidents on old type climbing lanes increased 1.1% during the same periods. In addition, in-depth study, the merging area of the passing-type climbing lanes was found out to be the weakest section where 43.8% traffic accidents out of total traffic accidents happened. It is noted that the merging area of the old type climbing showed only 25.0% traffic accidents. CONCLUSIONS : The new passing-type climbing lanes were found to be weak in terms of safety when compared with the old type climbing lanes. Especially, the merging area should be improved to reduce the risk of traffic conflicts between slow-moving vehicles and other vehicles.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.29 no.2D
• /
• pp.167-173
• /
• 2009

In general, an influenced area of merging section is defined as 500 m including 100 m upstream and 400 m downstream. However, from an observation of the actual traffic flow, it is found that merging effect influences more on downstream than upstream. In this study, an influenced area of merging section on freeway is analyzed by using turbulence which is defined as conflicts between vehicles. In order to overcome the limits of existing traffic flow detection system established with intervals of about 500 m, this study uses raw data collected from the detectors which are established in entrance ramps with similar road conditions. To divide data of each point into similar road conditions, the data of total 72 hours is sorted by Level of Service. An influenced area analyzed by standard deviation of speed is 700 m section of highway, including 300 m upstream and 400 m downstream, for both right and left ramps. It is the result including upstream 200 m more than previous studies.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.41 no.7
• /
• pp.575-583
• /
• 2017

This paper describes the design and evaluation of a driving mode decision algorithm for automated driving for merge situations on highways. For the development of a highly automated driving control algorithm for merge situations, the driving mode decision is crucial for merging appropriately. There are two driving modes: lane keeping and lane changing (merging). The merge mode decision is determined based on the state of the surrounding vehicles and the remaining length of the merge lane. In the merge mode decision algorithm, merge possibility and the desired merge position are decided to change the lane safely and quickly. A safety driving envelope is defined based on the desired driving mode using the information on the surrounding vehicles' behaviors. To obtain the desired steering angle and longitudinal acceleration for maintaining the subject vehicle in the safe driving envelope, a motion planning controller is designed using model predictive control (MPC), with constraints that are decided considering the vehicle dynamics, safe driving envelope, and actuator limit. The proposed control algorithm has been evaluated via computer simulation studies.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.38 no.6
• /
• pp.879-885
• /
• 2018

The highway weaving section is a point where there is a conflict caused by entry and exit vehicles. In order to minimize the impact on the main line, an accumulation line should be installed. The design speed of the collecting part is 50 km/h, but the actual driver does not have proper deceleration according to the design speed. In this study, considering the driving behaviors, the design specification, and the driver 's viewing angle, the appropriate separation distance for safe entry and exit of collecting, connecting, and collecting roads was examined. As a result of the analysis, it is found that a distance of 60m is required from the point where the merging starts. The results of this study are expected to contribute to the improvement of driver safety due to the conflict between entry and exit vehicles when applied to the weaving section where collecting roads are installed.

• International Journal of Highway Engineering
• /
• v.11 no.2
• /
• pp.217-228
• /
• 2009

Sections under the influence of merging in an uninterrupted facility create irregular interaction between vehicles, such as lane change, speed acceleration and deceleration because of the merging of ramp traffic flows which have traffic characteristics different from those of the main line. This causes a confused traffic flow phenomenon(turbulence), which is considered an unstable traffic characteristic between various continuous points in consideration of v conditions. In this study, in merge influence sections, detectors by lane-point were installed to create time and space-series -traffic data. The least significant difference(LSD), as the criteria for discriminating a significant speed change between points, was calculated to examine the turbulence. As a result, turbulence in merge influence section was found to change the zones of such occurrence and the seriousness levels according to traffic condition. Thus, the maximum merge influence section due to the turbulence was created in the traffic condition before congestion when traffic increases. According to characteristics of changes in speed, merge influence section was divided into upstream 100m$\sim$downstream 100m(a section of speed reduction), and downstream 100m$\sim$downstream 400m(a section of reduced speed maintenance and acceleration).

• Journal of Korean Society of Transportation
• /
• v.29 no.3
• /
• pp.123-137
• /
• 2011

In general, flow propagation has been explained using the shock wave theory which is expressed as a function of variations in volume and density. However, the theory has certain limitation in portraying heterogeneous flow, e.g., flow propagation between lanes. Motivated by this fact, this study seeks a new measure for analyzing the propagation characteristics of traffic flow at three sections of highway (i.e., merging area, weaving section, and basic section) from temporal and spatial perspectives, and then develops a model for estimating the measure for the flow propagation. The "shock wave speed" which is the measure widely adopted in literature, was first applied to describe the propagation characteristics, but it was hard to find distinct characteristics in the propagation. This finding inspires to develop a new measure named "Impulse Volume". It is shown that the measure better explains the propagation characteristics at the three study sections of highway. In addition, several models are also developed by performing multi-regression analyses to explain the flow propagation between lanes. The models proposed in this paper can be distinguished in three sections and the lane placement.

• International Journal of Highway Engineering
• /
• v.20 no.4
• /
• pp.59-72
• /
• 2018

PURPOSES : This study was conducted to develop expressway safety treatments based on the analysis results of older driver behaviors through literature review, surveys, and driving simulator experiments. METHODS : In this study, three analyses were conducted: surveys of 700 older drivers to find the risk segments they recognized, driving simulator experiments with older and younger drivers to investigate driver behaviors, and expert surveys to find the priority of expressway safety treatments for older drivers. RESULTS : Through survey results it was found that merging areas and tunnels were identified as the most dangerous areas, and more dangerous older driver behaviors were observed on those expressway segments in the driving simulator experiments. In addition, the priorities of safety treatments for each segment of expressways were decided based on expert surveys. CONCLUSIONS : It was concluded that choice and concentration strategies of expressway safety treatments for older drivers should be applied as perceptions regarding dangerous spots and older driver behaviors, including geometric designs, safety facilities, regulation, and institutes to improve expressway safety.

• International Journal of Highway Engineering
• /
• v.12 no.2
• /
• pp.91-97
• /
• 2010

This study started from the problem recognition of climbing lane installed in Korea roads. Because design standards of climbing lane installed don't match traffic pattern of korean drivers, coefficient of utilization of climbing lane is low and merging section between climbing lane and main lane has traffic accident possibilities. For this, brand-new design standards developed from the present lane design criterion, taper lenghs, and traffic signs, then field adoption test was carried out to prove the effectiveness. As a result, coefficient of utilization of climbing lane and average traffic velocity in climbing section are improved and the economic analysis also shows that brand-new standards has high feasibility for low cost. In case of broad application to not only expressway but national and local road based on the study, it could be a significant contribution to traffic flow improvement.