Journal of Transportation Engineering, Part A: Systems

Transportation projects have significant impacts on the transportation system’s resilience as well as people’s welfare. In addition, governments bear a financial burden associated with developing such projects due to the significant resources needed both during and after the founding period. Hence, the process of evaluating transportation projects is essential. This work presents a complete guideline for decision makers to evaluate transportation projects or choose from alternatives. The evaluation process of transportation projects consists of two parts: travel demand modeling and the assessment phase. The four-step modeling approach is described in this paper as an illustration of travel demand models, which is considered the initial stage in the evaluation process. After that, evaluation criteria affecting the judgment of decision makers are presented. The selection of these criteria forms the framework for the project evaluation process. Next, each evaluation criterion’s key performance indicators (KPIs) are discussed. Decision-making techniques such as benefit–cost analysis (BCA) and multiple criteria decision-making (MCDM) tools are then carefully explained. The forenamed techniques are used throughout the phase of assessment and prioritization.

Traffic flow prediction plays an important role in intelligent transportation systems (ITS) on freeways. However, incomplete traffic information tends to be collected by traffic detectors, which is a major constraint for existing methods to get precise traffic predictions. To overcome this limitation, this study aims to propose and evaluate a new advanced model, named transfer learning–based long short-term memory (LSTM) model for traffic flow forecasting with incomplete traffic information, that adopts traffic information from similar locations for the target location to increase the data quality. More specifically, dynamic time warping (DTW) is used to evaluate the similarity between the source and target domains and then transfer the most similar data to the target domain to generate a hybrid complete training sample for LSTM to improve the prediction performance. To evaluate the effectiveness of the transfer learning–based LSTM, this study implements empirical studies with a real-world data set collected from a stretch of I-15 freeway in Utah. Experimental study results indicate that the transfer learning–based LSTM network could effectively predict the traffic flow conditions with a training sample with missing values.

Varying implementations and availability of vehicle automation result in vehicles using significantly different driving behaviors. Adaptive cruise control implementations vary by manufacturer and also differ from human driving. Due to the prevalence of partial automation on new vehicles, researchers and practitioners interested in forecasting future traffic conditions are considering mixed or multiclass flow scenarios. Prior work on automated vehicles proposed a multiclass kinematic wave theory where the flow-density relationship changes in space and time in response to local variations in class proportions, They solved this theory using a Godunov approximation (cell transmission model). Capacity and congested wave speed are assumed to vary with respect to class proportions, and a triangular flow-density relationship is used. This problem is challenging because the flow-density relationship varies endogenously with the movement of vehicles, resulting in two unknowns that must be solved together: vehicle movements (cumulative counts) and boundaries defining changes in the flow-density relationship. This paper derives a multiclass link transmission model to solve the multiclass kinematic wave theory for any finite number of classes. In the process, we derive a multiclass Newell’s method to find exact solutions as a linear program. A simplified iterative algorithm is obtained to more quickly solve the multiclass link transmission model to be useful for large networks.

Fog has a critical impact on traffic safety on freeways. Variable speed limit (VSL) control can effectively harmonize vehicle speed and improve safety. However, most existing weather-related VSL controllers are limited in their ability to adapt to the dynamic traffic environment. This study developed an adaptive VSL control strategy under fog conditions with consideration of macroscopic factors that affect traffic safety risks. The optimal factors of VSL control were obtained by solving an optimization problem that coordinated safety and mobility using a genetic algorithm. The results showed that the proposed VSL control under fog conditions effectively improved safety without significantly decreasing mobility; risks were reduced by as much as 37.15% with only a 0.48% increase of total travel time. The VSL control strategy in this study can help transportation managers design VSL control systems to enhance freeway safety under fog conditions.

Traffic state information, road network structure characteristics, and built environment characteristics are factors influencing traffic safety, which will alleviate or aggravate traffic safety problems. This paper analyzes the relationship between these factors and traffic accidents involving either property damage only (PDO) crashes or killed and severe injury (KSI) crashes. The spatiotemporal distribution of the two types of accidents was analyzed. Abundant explanatory variables were extracted from accident data, license plate recognition (LPR) data, OpenStreetMap (OSM) data, and point of interest (POI) data based on complex network methods and information entropy theories. Geographical and temporal weighted regression (GTWR), geographically weighted regression (GWR), and ordinary least squares (OLS) models were used to analyze the influencing factors of traffic safety, respectively. The results demonstrate that the GTWR model performs best in modeling spatiotemporal heterogeneity data. Traffic state information, road network structure, and built environment factors all have significant effects on traffic accidents, and traffic state information have the highest correlation with traffic accidents among all factors. The greater the traffic volume, the more likely are traffic accidents. The strongest correlation is between PDO crashes and traffic state in the morning peak, in the evening peak, and at night. For a road network divided into grids, the more important the intersections in the grid, the greater is the street circuity, and the more chaotic the street direction, the more likely PDO crashes are to occur in the grid. Furthermore, the diversity of land use is positively correlated with traffic accidents in urban areas, whereas the correlation is negative in suburban areas, which reflects the spatial heterogeneity.

As the main road pattern, multilane roads are prevalent on high-grade highways. In a connected vehicle environment, drivers can perceive the full view of traffic information on each lane, which provides more opportunities for flexibly changing lanes on multilane highways. In addition, drivers can predict the traffic status on short notice and regulate the vehicle’s operating state in advance. This study determined the predictive effect of drivers in a multilane scenario using the lattice hydrodynamic model. The stability criteria of the proposed model were deduced via the reductive perturbation method; when the stability conditions do not hold, the modified Korteweg–de Vries (mKdV) equation can be deduced. By solving this above equation, we derived the kink–antikink soliton wave solution, which can be used to analyze and explain the formation and evolution process of traffic jams. The results show that the number of lanes and the prediction time of drivers considerably affect the stability of traffic flow. Simulation examples verified that when the predictive time is fixed and the number of lanes increases from 1 to 4, the fluctuation amplitude of traffic density decreases from 0.2 to 0.08 even with exogenous initial disturbance; when the number of lanes is fixed, the density fluctuation amplitude decreases as the predictive time increases; and when the predictive time increases to 0.7, the fluctuation amplitude of traffic density approximates to 0 and is a uniform flow.

As a key node in the urban road network, the operational efficiency of urban signal intersections directly affects the general operational efficiency of road network traffic. To improve the general efficiency of the intersection considering road dimensions, site red lines, and other constraints, some signalized intersections often use a straight-right lane or a dedicated right-turn lane to right-turn vehicles in the far right lane of the entrance lane. A straight-right lane will improve the efficiency of straight-through traffic during a green light; however, straight-through vehicle will prevent right-turn vehicles from crossing the intersection. A dedicated right-turn lane will reduce delays for right-turn vehicles; however, it will not improve the efficiency of straight-through traffic. To address these issues, this paper took the basic idea of spatial and temporal separation of straight-right and right-turn vehicles on the straight-right lane as the basis for designing a dynamic straight-right lane (DSRL) organization form, and for researching key design parameters and supporting control strategies. Using hydrodynamic analogy theory, the delays of the DSRL design were compared with those of the static straight-right lane design and the dedicated right-turn lane design, the relationship between each key design parameter and vehicle delays was analyzed, and sensitivity analysis under different traffic environments was conducted. This paper presents the basic concept of DSRL design, discusses and researches its important components, and developed new ideas and methods for the design with supporting control of straight-right lanes at urban intersections.

Accurately predicting freeway incident clearance time and analyzing influential factors are essential for developing effective traffic incident management strategies. Existing approaches for analyzing traffic incident clearance time include statistical models and machine learning models. Whereas the statistical approach is able to quantify the impact of influential factors on incident clearance time, it often yields unsatisfactory levels of the prediction accuracy. Conversely, the machine learning approach lacks model interpretability but can generate accurate predictions. To combine the advantages of both approaches, a survival analysis model based on deep neural network (DeepSurv) is applied to predict the traffic incident clearance time. We used the SHapley Additive exPlanations (SHAP) method to interpret the modeling results of the DeepSurv model and analyze the impact of influential factors on traffic incident clearance time. Results show that the DeepSurv model outperforms statistical models (i.e., Cox proportional hazard, accelerated failure time and quantile regressions) and traditional machine learning models (i.e., support vector machine, random forest, and extreme gradient boosting algorithm) in terms of prediction performance. The analysis results indicate that response time, incident type (collision), lane closure type (all travel lanes blocked, total closure), involvement of fire and traffic control are significant influential factors affecting traffic incident clearance time. Our findings indicate that the proposed DeepSurv model is a more effective approach to predict traffic incident clearance time.

The postpandemic period has seen a significant increase in traffic volume on freeways, necessitating the implementation of advanced traffic management systems, such as lane-level freeway tolling systems, to predict traffic patterns and alleviate congestion. Although deep learning models have proven effective in predicting traffic states, little research has focused on lane-level traffic prediction, which is crucial for emerging intelligent transportation applications. To address this gap, this study develops a lane-level road segment graph and proposes a lane-based road network traffic volume prediction model, GCN-LSTM, that combines graph convolution network (GCN) and long short-term memory (LSTM). The proposed model employs different graph Laplacian matrices, and the performance of these corresponding derived models is compared with that of existing traffic prediction models. The proposed model is evaluated using traffic volume data collected from inductive loop detectors installed on freeways in the Seattle area, including both high-occupancy toll lanes and general-purpose lanes. The results demonstrate that the GCN-LSTM model with the combinatorial Laplacian matrix outperforms other models. Additionally, the model’s prediction performance remains consistent when using input data with various temporal ranges. Furthermore, excluding high-occupancy toll lane data from the dataset improves the prediction accuracy, highlighting the importance of developing specialized models for lane-level traffic prediction tasks.

Abstract

The left-turn waiting area (LTWA) has been commonly used in many signalized intersections to increase the capacity of left-turn movements. However, LTWAs may negatively affect the performance of arterial signal coordination because the lag-lag left-turn phase sequence turns out to be the only feasible left-turn treatment to accommodate the presence of LTWAs. To investigate the effects of LTWAs on arterial signal coordination, this paper proposes an analytical model based on the classic MAXBAND algorithm to quantify the gaps between the actual bandwidth and the maximum possible bandwidth. The model can further identify the applicable signal timing condition of LTWAs in an arterial signal system. A case study of a signalized arterial with LTWAs in Wuhan City was conducted to verify the effectiveness of the proposed method. Results showed that the signal timing parameters obtained from the proposed model resulted in a shorter queue length, a lower delay, and a smaller number of stops when the lag-lag phase sequence was adopted in comparison with the other phase sequences.

Practical Applications

The left-turn waiting area (LTWA) has been adopted in many cities to improve operational efficiency for turning traffic, especially for large intersections. However, a lag-lag phase sequence has to be adopted at signalized intersections with LTWA, which would affect the performance of arterial coordination due to the limitations of the phase sequence. The model proposed in this paper could assist field engineers with exploring the appropriate signal timing condition for an arterial coordination system with LTWA intersections, where the lag-lag left-turn phasing scheme could produce wider bandwidth than the other phase sequences. In addition, the rationality of LTWAs in an arterial system could be evaluated further according to the real traffic situation.