International Journal of Transportation Science and Technology最新文献

Familiar and unfamiliar drivers may exhibit different behaviours in response to the road environment. Overall familiarity with the road environment is a human factor believed to play a role in road crash injury severity due to its effect on a driver's decision-making process, reaction time, etc. Hence, there is a need to separately analyse familiar and unfamiliar drivers regarding the injury severity of crashes. Using a six-year database of 30,481 rural two-vehicle crashes in Guilan province, Iran, this research first defined four categories of crashes, reflecting various levels of the involved drivers’ familiarity with the environment (72% of drivers were from the same vs. 28% from a different province). Next, the injury severity of crashes in each familiarity crash category was analysed using both non-parametric (classification and regression trees) and parametric (logistic regression) methods. When both crash parties were unfamiliar, several results are different compared to when both parties were familiar or when ignoring driver familiarity. For instance, young at-fault drivers increased the injury severity of crashes if they were unfamiliar, while they decreased the crash severity if they were familiar. Also, crashes in winter tended to be more severe when one or especially both crash parties were unfamiliar, but winter crashes were less severe when both drivers were familiar or when driver familiarity was ignored. Overall, when both drivers were familiar, 63% of crashes were injury/fatal; however, when both drivers were unfamiliar, only 31% of crashes involved an injury or fatality.

The service time and time headway of the vehicles are used to define the equivalency factor of different vehicle classes at the toll plaza. Both the service time and time headway are point measures and do not account for the system delay (time difference when a vehicle enters the queue and leaves the tollbooth) caused to the vehicle. The present study aims to quantify the system delay incurred to the vehicles at electronic toll collection (ETC) lanes under mixed traffic conditions using a microsimulation approach. Field data collected from one toll plaza located on a National Highway are used for simulation model generation. A new terminology called system delay-based toll equivalency factor (DTEF) is introduced to convert the different vehicle classes into equivalent passenger cars. The DTEF variation for different approach volumes and heavy commercial vehicle (HCV) compositions were checked at different ETC penetration levels. A total of 288 scenarios have been worked out, and simulation runs have been made for all such scenarios to obtain the DTEF values. The average DTEF value of HCV was obtained as 2.20. Further, it is found that with an increase in approach volume and HCV share in the traffic stream, the DTEF value increases. The outcome of the present study will be useful to field practitioners and engineers to determine the capacity in equivalent DTEF/hr and level of service of a toll plaza in terms of system delay.

The present study aims to investigate the impact of texting and web surfing on the driving behavior and safety of young drivers on rural roads. For this purpose, driving data were gathered through a driving simulator experiment with 37 young drivers. Additionally, a survey was conducted to collect their demographic characteristics and driving behavior preferences. During the experiment, the drivers were distracted using contemporary smartphone internet applications i.e., Facebook Messenger, Facebook and Google Maps. Regression analysis models were developed in order to identify and investigate the effect of distraction on accident probability, speed deviation, headway distance, as well as lateral distance deviation. Additionally, random forest (RF), a machine learning classification algorithm, was deployed for real-time distraction prediction. It was revealed that distraction due to web surfing and texting leads to a statistically significant increase in accident probability, headway distance and lateral distance deviation by 32%, 27% and 6%, respectively. Moreover, the driving speed deviation was reduced by 47% during distraction. Apart from the real-time prediction, the RF revealed that headway distance, lateral distance, and traffic volume were important features. The RF outcomes revealed consistency with regression analysis and drivers during the distractive task are more defensive by driving at the edge of the road near the hard shoulder and maintaining longer headways. Overall, driving behavior and safety among young drivers were both significantly affected by the investigated internet applications.

The empirical Bayes (EB) method based on parametric statistical models such as the negative binomial (NB) has been widely used for ranking sites in the road network safety screening process. In this paper a novel non-parametric EB method for modeling crash frequency data based on Conditional Generative Adversarial Networks (CGAN) is proposed and evaluated over a real-world crash data set. Unlike parametric approaches, there is no need for a pre-specified underlying relationship between dependent and independent variables in the proposed CGAN-EB and they are able to model any types of distributions. The proposed methodology is applied to real-world and simulated crash data sets. The performance of CGAN-EB in terms of model fit, predictive performance and network screening outcomes is compared with the conventional approach (NB-EB) as a benchmark. The results indicate that the proposed CGAN-EB approach outperforms NB-EB in terms of prediction power and hotspot identification tests.

Public Transport (PT) provides passenger mobility and contributes to sustainable transportation. To achieve this a PT system must provide continuous accessible service and connections for passengers. PT reliability is considered a major obstacle to growing its market share. Current solutions primarily address travel time reliability through methods like priority lanes and traffic signal priority. Dwell time reliability improvement, in turn, can be achieved by the use of smart cards which reduce the variability in boarding and alighting times. Another factor affecting reliability is in-vehicle crowdedness which causes delays and increases dwell time variability. To mitigate crowdedness, we propose a monetary approach that dynamically changes the fare based on the in-vehicle crowdedness level in a manner similar to congestion pricing. This approach would shift some passengers from boarding the over-crowded vehicle to waiting for the next, less crowded vehicle, while compensating them for the additional waiting. Passengers unwilling to wait might pay a penalty if the additional waiting time is reasonable. To assess the attitude of passengers towards a dynamic fare model, a stated preference questionnaire was developed to assess the factors that affect the choice of whether or not to board an over-crowded vehicle. Based on panel data and the fixed effect logit model it was revealed that the higher the waiting time, the lower the willingness to board the next vehicle. However, monetary schemes (penalties or discounts) increased the willingness to wait and board the next vehicle. Moreover, the willingness to wait was higher when a penalty was introduced compared to a discount, which is in line with the prospect theory. The results suggest that it is possible to construct a dynamic fare model that using data on vehicle crowdedness levels and waiting times obtained from advanced data collection systems, which is integrated within a mobile payment application. This approach could reduce crowdedness and increase reliability.

The cut-ins (one kind of lane-changing behaviors) have result in severe safety issues, especially at the entrances and exits of urban expressways. Risk prediction and characteristics analysis of cut-ins are part of the essential research for advanced in-vehicle technologies which can reduce crash occurrences. This paper makes some efforts on these purposes. In this paper, twenty-four participants were recruited to conduct the experiments of multi-driver simulation for risky driving data collection. The surrogate measures, Time Exposure Time-to-Collision (TET) and Time Integrated Time-to-collision (TIT) were employed to quantify the risk of cut-ins, then k-means clustering was applied for risk classification of 3 levels. Multiple candidate variables of two kinds were extracted including 10 behavioral variables and 7 driver trait variables. Based on these variables, three prediction models including decision tree (DT), gradient boosting decision tree (GBDT) and long short-term memory (LSTM) are used for predicting the risks of cut-ins. Results from data validity verification show that the data collected from multi-driver simulation experiments is valid compared with real-world data. From results of risk prediction models, the LSTM, with an overall accuracy of 87%, outperforms the GBDT (80.67%) and DT (76.9%). Despite this, this paper also concludes the merits of the DT over the GBDT and LSTM in variable explanation and the results of DT suggest that controlling the proper lane-changing gap and short duration of cut-ins can help reduce risks of cut-ins.

Comprehension of traffic signs is important to road safety. This review aims to study the extent to which road users in different countries comprehend traffic signs and to identify which ergonomic principles in traffic sign design can affect the levels of comprehension. We conducted an extensive literature review dealing with comprehension of public traffic signs directed at any road user. We searched Journal articles indexed by Scopus, ScienceDirect, and Web of Science. The search identified 35 articles that assessed the comprehension of 931 traffic signs in 26 countries, including six studies that tested the comprehension of new versus existing traffic signs. Various methods have been implemented to measure traffic signs’ comprehension levels and assess traffic sign design’s conformity to different ergonomic principles. Results indicate high variability in the comprehension levels of signs, e.g., signs such as “Road works” and “No U-turn” are highly comprehended (comprehension levels over 90 %), while other signs like “termination of road” are rarely comprehended by road users. Regarding the acceptable comprehension levels, 23.1 % of the assessed traffic signs achieved levels above 85 %; and 53.3 % of signs have comprehension levels lower than 67 %. On the other hand, twenty-four studies evaluated how traffic signs comply with ergonomic design principles. Incorporating ergonomic principles into the design of traffic signs can improve comprehension levels. However, apart from the familiarity, there is uncertainty about the ergonomic principles that could maximize the comprehension of traffic signs. Efforts should be made to ensure that different populations of road users sufficiently comprehend traffic signs.

The rapid growth in data collection, storage, and transformation technologies offered new approaches that can be effectively utilized to improve traffic crash prediction. Considering the probability of traffic crash occurrence vary due to the spatiotemporal heterogeneity, this study proposes a state-of-the-art deep learning-based model that incorporates spatiotemporal information for the short-term crash prediction, named as Deep Spatiotemporal Hybrid Network (DSHN). The model integrates Convolutional Neural Network (CNN), Long Short-term Memory (LSTM), and Artificial Neural Network (ANN) to incorporate the synergistic power of individual models. The study utilizes different data sources such as big traffic data collected from Paris road network sensors, weather conditions, infrastructure, holidays, and crash data. The results indicated that the proposed DSHN model outperforms the baseline models with an Area Under Curve (AUC) of about 0.800, an accuracy of 0.757, and a false alarm rate of 0.217. In addition, the importance of each data type is evaluated to investigate their impacts on the prediction performance of models. The sensitivity analysis results indicate that the road sensor data that includes average speed, vehicle kilometer traveled (VKT), and weighted average occupancy has the highest impact on the prediction accuracy.

Bus Rapid Transit (BRT) has advantages over rail-based systems as a public transportation system. The ease of implementation and low investment costs attract many cities to develop BRT systems, including Banjarmasin, Indonesia. Banjarmasin currently has eight BRT stop points that reach only two sub-districts out of five. The limited range of BRT stops within the city can affect the level of accessibility of the BRT system. The accessibility of the transit system itself can be seen from the number of daily passengers. This study aims to analyze the criteria that affect the level of accessibility of the BRT stops in the study area and then compile a model based on significant criteria. Previous literature on accessibility modeling shows varied methods and approaches. In this study, the system accessibility was measured using the composite method and modeled using Geographically Weighted Regression (GWR), which is a relatively new approach. The results show that seven criteria affect the level of accessibility of the BRT stops. The model was first built mathematically using OLS. Then, GWR analysis was accomplished on spatial variables, resulting in a higher significance model. Furthermore, the GWR produces a visual-spatial model and performs simulation and sensitivity tests to make the research purpose more informative. The spatial criteria for the accessibility of the BRT stop locations in the model include the distance of stops to the road intersection, mix-use entropy index, population density, and land value.

Traffic signs play a crucial role in managing traffic on the road, disciplining the drivers, thereby preventing injury, property damage, and fatalities. Traffic sign management with automatic detection and recognition is very much part of any Intelligent Transportation System (ITS). In this era of self-driving vehicles, calls for automatic detection and recognition of traffic signs cannot be overstated. This paper presents a deep-learning-based autonomous scheme for cognizance of traffic signs in India. The automatic traffic sign detection and recognition was conceived on a Convolutional Neural Network (CNN)- Refined Mask R-CNN (RM R-CNN)-based end-to-end learning. The proffered concept was appraised via an innovative dataset comprised of 6480 images that constituted 7056 instances of Indian traffic signs grouped into 87 categories. We present several refinements to the Mask R-CNN model both in architecture and data augmentation. We have considered highly challenging Indian traffic sign categories which are not yet reported in previous works. The dataset for training and testing of the proposed model is obtained by capturing images in real-time on Indian roads. The evaluation results indicate lower than 3% error. Furthermore, RM R-CNN’s performance was compared with the conventional deep neural network architectures such as Fast R-CNN and Mask R-CNN. Our proposed model achieved precision of 97.08% which is higher than precision obtained by Mask R-CNN and Faster R-CNN models.