Journal of Transportation Engineering, Part A: Systems

Surrogate safety measures such as traffic conflicts are gaining more and more attention for traffic safety analysis. The traffic conflict technique evaluates the frequency and severity of traffic conflicts at a location typically using various time proximity indicators such as the time-to-collision (TTC) and post-encroachment time (PET). However, growing concerns have been raised that time proximity indicators may not be effective measures for measuring conflict severity in less-organized traffic environments. In such environments, mixed road users are likely to share small spaces and take evasive action to prevent conflicts or collisions. The objective of this study was to examine and compare the time proximity (TTC) indicator and evasive action-based (yaw rate and jerk) indicators for evaluating the severity of powered two-wheeler (PTW) conflicts. PTW usage is growing in many developing countries such as China, and there has been concern about their impact on safety. Video data were collected at a middle block shared traffic street in Kunming, China. Traffic conflict analysis was conducted using automated video-based computer vision techniques. Ordered-response models were used to relate the conflict indicators to safety experts’ evaluation of conflict severity. A random effect model was developed to account for the unobserved heterogeneity that affects conflict severity. As well, a random intercept model was developed to assess the effect of incorporating the variation in each expert evaluation. The results showed that the yaw rate ratio was efficient in measuring conflict severity for electric (e)-scooters, motorcycles, and bicycles. The TTC was an efficient indicator in measuring conflict severity for e-bikes and bicycles.

Currently studies exploring the effect of real-time traffic and weather characteristics on crash severity are relatively limited, and studies focusing on powered two-wheelers (PTWs) are even fewer. The objective of this paper is to add to current knowledge through the investigation of PTW occupant injury severity on the urban motorway Attica Tollway in Athens, Greece. For that purpose, 163 crashes (23 severe/fatal and 140 slight) that occurred between 2006 and 2011 were analyzed by considering real-time traffic and weather parameters. The required crash data was extracted from the Greek crash database of the National Technical University of Athens and also Attica Tollway. Traffic data were extracted from the Traffic Management Center of Attica Tollway, and weather data were extracted from the Hydrological Observatory of Athens (HOA). To reduce data dimensionality and to overcome multicollinearity between variables, a two-step approach was followed. First, principal component analysis and random forests were used as preliminary analyses. To account for the low number of and proportion of killed and severely injured (KSI) occupants, a Firth logistic regression was then applied. Results showed that in general, traffic and speed variations caused more severe injuries whereas weather parameters seemed to have no effect. It was also found that mean speeds and variations in volumes of trucks increased the crash injury severity of PTW crashes, but it was interesting that an increased proportion of trucks led to slighter injuries. The findings of this paper are considered promising and will contribute to better understanding of occupant injury severity on urban motorways by focusing on PTWs.

On freeways, managed lanes (MLs) have emerged as an effective dynamic traffic management strategy. MLs have been successfully implemented as an important facility in improving traffic mobility and generating revenue for transportation agencies. In this study, scenarios were built and tested in microsimulation to specify the safest accessibility level and decide on the safe weaving length near access zones. The findings indicated that the conflict rate on MLs was 48% and 11% lower than that of general purpose lanes (GPLs) in the peak and the off-peak periods, respectively. A log-linear model was developed with estimation of odds multipliers for the conflict frequency analysis. The results suggested that one accessibility level was the safest option in the 14.5-km (9-mi) corridor. The length of 305 m (1,000 ft) per lane change was shown to be the safest weaving length near access zones. Additionally, a weaving length of 183 m (600 ft) per lane change was not recommended. The findings of this study represent a further step toward improving access design of MLs.

Obtaining an optimal stability condition is very important to benefit traffic flow operations for a mix of connected and automated vehicles (CAVs) and regular vehicles. In view of multiple spatial anticipations of CAV car-following models, this paper presents a stability analysis method for mixed CAV flow from the perspective of the uniform local platoon. Transfer function theory was used to derive the stability criterion of the uniform local platoon, based on which a stability chart of equilibrium speeds and CAV feedback coefficients was calculated. Numerical simulations were also performed on a segment of highway with an on-ramp using car-following models to evaluate the impacts of the stability analysis on fuel consumption and emissions [carbon monoxide (CO), hydrocarbons (HC), and nitrogen oxides (${\mathrm{NO}}_{\mathrm{X}}$)]. The stability chart indicates that by controlling CAV feedback coefficients, the optimal stability condition can be obtained, in which the uniform local platoon remains stable for all driving speeds. Moreover, the stability analysis method can reduce fuel consumption and traffic emissions.

In recent years, the customized bus (CB) has been introduced and popularized in China to improve the attraction and service level of public transportation. A key point of the CB system, the route design problem, is always formulated as a vehicle routing problem with pickup and delivery (VRPPD). However, VRPPD cannot sufficiently describe the in-vehicle passengers of multiple vehicles involved. In this paper, a mixed integer programming model is developed to formulate a multivehicle routing problem, with suggestions for bus stop locations and routes. Meanwhile, the model can determine passenger-to-vehicle assignment based on a series of constraints, like operation standard and number of stations. In solving the problem, a numerical example is used to compare a genetic algorithm (GA) and branch-and-cut algorithm. The comparison results illustrate that GA is more efficient with lower complexity. Finally, in order to apply and evaluate the proposed model, a real-world case study using smartcard data is conducted to compare the approach with the current CB route design method in Beijing.

In this paper, mixed-integer programming for the railway network design problem with a new objective function is proposed. The model considers development projects (new line construction and existing line improvement), available budget, block capacity, and origin-destination matrix demand. A standard code is used to calculate the capacity of the blocks. The objective function is to minimize the total cost of direct and indirect (external) costs. We add the effect of railway network development on the road network to the model. The proposed model is implemented for the railway network of Iran and solved by an exact method.

Actual dwell time (DT) at rail transit stations is one of the most significant and useful parameters in evaluating the scheduled timetable and making any dynamic operation adjustments as necessary. It can be influenced by many factors with intricate nonlinearities that make it difficult to build an overall model. The purpose of this paper is to develop a data analytics approach to modeling the train actual DT by combining a density-based spatial clustering of applications with noise (DBSCAN) with the support vector regression (SVR) algorithm. There are three steps in this modeling approach: (1) identifying factors that influence train actual DT; (2) integrating data collected from automatic fare collection and automatic train supervision, and then classifying actual DT data by using the DBSCAN approach; and (3) establishing a nonlinear model to estimate the actual DT based on the SVR method, which is calibrated by using grid search techniques. Using a station along Line 12 in the Shanghai metro network as an example, this innovative hybrid approach delivers high quality results in that the mean squared error of the training set is 2.87 s and the coefficient of determination is 0.66. Application results indicate that the influential factors identified and the nonlinear model developed can be used to explain and predict the train actual DT well, and that the developed model can be applied to assist decision making at both the tactical and operational levels.

Truck lane restriction is one of the most common highway lane management strategies; it ensures that heavy vehicles mostly use the outer lanes of a multilane highway. A higher percentage of heavy vehicles in the design lane increases the required load-bearing capacity of the pavement structure and therefore increases construction cost. This study considers three probable strategies to control this cost trend for flexible pavement: distribution of trucks across multiple lanes, design of heterogeneous pavement structures across lanes, and separation of car lanes from truck lanes. It was found that the construction cost of one unit length of highway is less for each strategy than the cost incurred with truck lane restriction in place. In order to take advantage of platooning, cars and trucks may not continue to use highway lanes in the traditional way. The cost advantages reported here can be exploited for selecting future highway lane management strategies in the environment of connected and autonomous vehicles.

Weather is an important contributing factor for bus operations. It is essential to develop a weather-responsive bus scheduling system to improve the level of bus service. Bus travel time is a significant factor to be considered while planning bus schedules. For example, if the next day’s bus travel time can be predicted according to the weather forecast and historical bus global positioning system (GPS) data, then it will be significant for adjusting bus scheduling in a timely manner. Harbin is the northernmost provincial capital of China with a population of about 4.5 million people. The modal share of bus transport is 40%, and the bus fleet includes 5,000 GPS-equipped buses. The minimum temperature in winter can reach $-38°\mathrm{C}$. Harbin is considered as the study area for our research because of the relative dominance of public transit usage, long winters, and cold weather. In this study, the bus operation data of Harbin in the winter of 2011 and 2012 were analyzed from a statistical perspective. The travel time series of successive buses was found to be autocorrelated, i.e., the travel times of previous two buses affect the travel time of a third bus. Thus, a fitting model for the next day’s bus travel time was proposed based on the historical GPS data and cumulative snowfall level in the weather forecast. The model was evaluated by taking Bus Lines 18, 68, and 69 in Harbin as examples. The mean absolute percentage error of each bus line was less than 9%. Taking Bus Line 18 as an example, it was found that the bus travel time increased by 0.483 min if the cumulative snowfall level increased by 1.