Journal of Transportation Engineering, Part A: Systems

The primary goal with which any transportation system is designed is to make efficient use of the available infrastructure to achieve better level of service (LoS). However, LoS is observed to be deteriorating on urban roads, especially near signalized intersections, primarily due to the suboptimal operation of traffic signals. To achieve optimal performance of traffic signals, knowledge about the traffic states prevalent in the vicinity of the intersection is essential. Traffic states in general can be estimated at both macroscopic and microscopic level by employing various mathematical and data-driven approaches. However, obtaining these variables near the intersection is difficult and challenging under varying traffic conditions. This paper presents a systematic review of the state of the art in traffic state estimation (TSE) near signalized intersections both under homogeneous, lane-based, and heterogeneous less lane disciplined (HLLD) traffic conditions. This is expected to be a guide to traffic engineers, decision makers, and researchers aiming to gain pertinent knowledge about the sensors that can be used, data that needs to be collected, estimation methods that are suitable, and the intersection performance measures that need to be evaluated. The gaps in the current state of the art and future research directions are highlighted. In addition, insights on ways to address challenges pertaining to TSE near intersections under HLLD traffic conditions are also discussed.

Abstract

A safety treatment of shoulder rumble strips, shoulder widening [from 0 to 0.6 m (2 ft)] and pavement resurfacing (RSR) was deployed to over 644 centerline km (400 mi) of rural two-lane highways in Missouri. The effectiveness of the RSR treatment was evaluated with the empirical Bayesian before/after study method based on the roadway characteristics, traffic and crash data collected. Crash modification factors (CMF) for 20 categories of crash severity and/or collision type were calculated. Generally, the RSR treatment is effective for improving the safety performance of rural two-lane roads, especially for fatal and injury crashes. The RSR treatment is specifically effective for run-off road and head-on crashes. A cumulative residual (CURE) analysis reveals that fixed-value CMFs for all annual average daily traffics (AADT) range are likely to result in biased estimation on the RSR’s safety effectiveness. The crash modification function (CMFunction) was introduced in this study and a sample size of 32 was selected to develop the CMFunction based on the sensitivity analysis results. CMFunctions were developed for two collision types of different crash severities. The CURE plot indicates that crashes predicted with CMFunctions are more accurate than those with CMFs and therefore CMFunctions are more reliable than fixed-value CMFs for evaluating the effectiveness of safety treatments.

Practical Applications

The roadway departure crashes, in which the vehicles left the roadways and hit trees or other fixed objects on roadway sides, can cause fatalities or serious injuries and is a big safety concern for rural highways. Many state and local agencies deployed various safety treatments to eliminate the roadway departure crashes or alleviate the crash severities. The Missouri Department of Transportation installed shoulder rumble strips together with shoulder widening and pavement resurfacing on rural two-lane highways to reduce the roadway departure crashes. Results from statistical analysis indicate that totally it will reduce 9% of roadway departure crashes; however, the treatment can be more effective for high-volume roadways than for low-volume roadways (specifically, roadways with AADT less than 2,000). To provide more reliable results, the authors developed a series of function to estimate how many crashes can be reduced with this safety treatments. State and local agencies can use the tool to prioritize roadway segments for safety improvements and quantify their safety benefits.

Bus exterior advertising provides a powerful way to establish brand awareness because it can reach a mass of audiences with a high frequency. For a certain advertisement category, advertising effectiveness is largely dependent upon its exposure times to the target audience who takes interest in advertisement, which is termed targeted advertising. Given that the distribution of target audiences over a city varies among different advertisement categories, a practical way of enhancing overall advertising effectiveness is to deploy a bus with certain advertisement category to the bus line that best fits its target area. This gives rise to a decision-making problem of targeted bus exterior advertising and bus scheduling. In this paper, the problem is formulated as a biobjective optimization model with objectives of maximizing the quantified advertising effectiveness and minimizing the number of bus fleet size to cover all trips. The advertising effectiveness is quantified using audience demographic data. The deadheading of buses is also enabled in the scheduling process to facilitate both objectives. The Non-dominated Sorting Genetic Algorithm-II-Large Neighborhood Search (NSGA-II-LNS) algorithm is developed to solve the biobjective problem with the incorporation of large neighborhood search operators into the framework of the NSGA-II to improve solution quality. Various experiments were set up to verify the proposed model and solution algorithm.

Abstract

Sudden infectious diseases and other malignant events cause excessive costs in the supply chain, particularly in the transportation sector. This issue, along with the uncertainty of the development of global epidemics and the frequency of extreme natural disaster events, continues to provoke discussion and reflection. However, transport systems involve interactions between different modes, which are further complicated by the reliable coupling of multiple modes. Therefore, for the vital subsystem of the supply chain-multimodal transport, in this paper, a heuristic algorithm considering node topology and transport characteristics in a multimodal transport network (MTN): the Reliability Oriented Routing Algorithm (RORA), is proposed based on the super-network and improved k-shell (IKS) algorithm. An empirical case based on the Yangtze River Delta region of China demonstrates that RORA enables a 16% reduction in the boundary value for route failure and a reduction of about 60.58% in the route cost increase compared to the typical cost-optimal algorithm, which means that RORA results in a more reliable routing solution. The analysis of network reliability also shows that the IKS values of the nodes are positively correlated with the reliability of the MTN, and nodes with different modes may have different transport reliabilities (highest for highways and lowest for inland waterways). These findings inform a reliability-based scheme and network design for multimodal transportation.

Practical Applications

Recently, the COVID-19 epidemic and the frequency of natural disasters such as floods have prompted scholars to consider transport reliability. Therefore, efficient and reliable cargo transportation solutions are crucial for the sustainable development of multimodal transport in a country or region. In this paper, a new algorithm is designed to obtain a reliability-oriented optimal routing scheme for multimodal transport. Using actual data from the Yangtze River Delta region of China as an example for experimental analysis, we obtain that: (1) the proposed algorithm is superior in terms of efficiency, accuracy, and route reliability, which means that the new algorithm can quickly find more reliable routing solutions in the event of urban transport infrastructure failures; and (2) highway hubs have the greatest transport reliability. Conversely, inland waterway hubs are the least reliable. The influence of national highways and railways on the multimodal transport system is unbalanced. These findings provide decision support to transport policymakers on reliability. For example, transport investments should be focused on building large infrastructure and increasing transport capacity, strengthening the connectivity of inland waterway hubs to hubs with higher transport advantages, and leveraging the role of large hubs.

Understanding how drivers exit freeways and interact with deceleration speed-change lanes (SCL) and ramp controlling features is important for adequate design of exit ramp terminals that meet drivers’ expectations and needs. Therefore, this paper investigates drivers’ diverging behavior along exit ramp terminal segments, including freeway right lane (FRL), SCL, and ramps based on video-based trajectory data collected using unmanned aerial vehicles (UAVs). Trajectories of 3,259 vehicles were collected at six sites as the vehicles moved on the FRL or SCL and off-ramp. Diverging behavior measures, including diverging speed, diverging location, deceleration rate on SCL, SCL utilization, and speeds at ramp gore nose and end of the SCL, were then extracted and used for the qualitative and quantitative analysis. The qualitative analysis showed that drivers exiting a freeway at taper-type SCLs tended to start deceleration on the FRL, which impacted the speeds of nonexiting vehicles. On the other hand, this behavior was not evident at parallel-type SCLs. Additionally, drivers were found to adopt a single overall deceleration rate from the diverge point to the end of deceleration. Finally, observations of data confirmed the importance of accounting for the effects of ramp controlling features on the diverging behavior and vehicle deceleration needs at freeway exit ramp terminals. The paper suggests that the design of deceleration SCLs should take into consideration the effect of ramp controlling features in the design of deceleration SCLs.

How to strike a balance between detection speed and recognition accuracy has become a major challenge in real-time object detection. In this research, the YOLOv5 (You Only Look Once version 5) model was lightweight and optimized to improve the detection speed and accuracy of the network. To prune the backbone and neck are to simplify the network structure and reduce the parameters. The lightweight structure of the C3 module was designed and incorporated into the attention mechanism to improve the feature extraction capability of the network. For the public traffic sign dataset, the label assignment strategy and the loss function of YOLOv5 were refined to alleviate the imbalance between positive and negative samples and to better compute the loss, resulting in more stable and efficient training. Compared with other mainstream single-stage models, it achieves a better trade-off between speed and accuracy. With only 0.85 M parameters, 91.9% of mAP (mean average precision) and 360 FPS (frames per second) were achieved, which were 16.26% mAP and 26.67 FPS higher than the conventional YOLOv5n, respectively. The performance of our lightweight model in traffic sign detection far exceeds the most advanced achievements.

This study is the first in the literature to investigate the joint equilibrium of departure time and parking location choices in the station-based one-way carsharing system. Time-varying road tolls are optimized to eliminate queueing deadweight loss, minimize schedule delay cost, and induce an economically efficient parking pattern. The round-trip commuting pattern under the fully privately owned vehicle environment was also studied, and the aggregate social cost of the two travel modes are compared with each other to verify the potential benefit gained from implementing the one-way carsharing service. We then investigated demand elasticity with respect to the user’s daily travel cost in a one-way carsharing-only system, and an externality-based charging mechanism is proposed to achieve the optimal demand-supply level. In the one-way carsharing-only travel mode, the following results were obtained: (1) the evening commute pattern is not a simple mirror image of the morning one when spatial parking is considered in both scenarios with and without pricing; (2) when taking account of elastic demand, in optimal time-varying pricing regimes, the charging scheme automatically leads to the socially optimal demand level, but in free-pricing regime, additional user anonymous tax must be charged to achieve the optimal demand level; and (3) the station-based one-way carsharing service benefits the public in aggregate social cost saving compared with the mode of driving a self-owned car.

Road networks are the backbones of urban communities and can be vulnerable to many disruptions caused by natural hazards, crashes, and rush hours, etc. Numerous efforts have been made in recent years to building more resilient urban road traffic systems against various disruptions through an improved understanding of resilience performance and more effective traffic intervention. A new resilience modeling methodology has been developed for disrupted traffic systems under natural hazards in terms of both local traffic performance and percolation-based robustness at the network scale. Hybrid data enhancement is conducted by integrating limited real-world data and microscopic traffic simulation to address the common challenge of data scarcity under hazards. The study further investigates the feasibility of applying early active traffic intervention through traffic signal design optimization only at some strategic intersections to improve both the traffic performance and resilience at local and network scales. A case study of the City of Fort Collins during snowstorms is made to demonstrate the methodology. The contributions of this study are reflected by shedding light on the following critical, yet unanswered questions: (1) how to deal with the common data shortage of disrupted traffic systems under hazards, (2) how to provide comprehensive time-progressive resilience assessment of a disrupted traffic system at different spatial scales, and (3) is it feasible to apply traffic intervention, such as smart intersection traffic control, only at limited intersections during hazards to improve both system resilience and traffic performance.

The resurgence of demand-responsive transit (DRT) has been pushed toward a more sharable and user-friendly mobility service to reshape the urban mobility ecosystem. Nevertheless, to fulfill the swiftly growing diversified and personalized demand, how to depict the cancellation behavior of passengers remains an often overlooked but extremely significant challenge. This paper attempts to discuss the cancellation behavior in the two-phase optimization process. With regard to the unconfirmed reservation cancellation, bounded rationality is incorporated to portray the decision-making process of passengers. Bernoulli random variable is utilized to discuss the stochastic possibility of order cancellation. Simultaneously, a two-phase biobjective model considering the cancellation behavior is constructed to balance the inherent trade-off between passengers and operators. In addition, the algorithm based on multidirectional local search is constructed to achieve the Pareto front for the proposed model. Small-scale experiments in the Nguyen-Dupuis network are illustrated to validate the effectiveness of the algorithm. Subsequently, the Beijing case is further exemplified to evaluate the performance of DRT service considering the cancellation behavior. A reasonable penalty mechanism contributes to ensuring operating profit while substantially diminishing the penalty cost. Therefore, this paper facilitates not only capturing the cancellation behavior of passengers but also developing a more flexible and cost-effective DRT service.

This paper develops a mixed platoon control strategy incorporating vehicle dynamics and time-varying vehicle-to-everything (V2X) delays to guarantee the consensus of a mixed platoon and reduce the impacts caused by the inconstant driving behavior of human-driven vehicles (HVs). In particular, a system control framework is established for a mixed vehicle platoon, which bears HVs and connected automated vehicles (CAVs). More precisely, this system control framework considers directly controlling CAVs and indirectly guiding the HVs to improve the consensus of the whole platoon with respect to velocity error and headway. Furthermore, based on the third-order closed-loop dynamic model, the consideration of vehicle dynamics and time-varying delays are taken into account. Then, theoretical analysis employs Lyapunov–Krasovskii theory to derive the delay boundary that determines the asymptotic stability and local string stability. Finally, a performance comparison with the existing algorithm is carried out to further demonstrate the advantages of the proposed strategy.

It has been well-recognized that driving behaviors significantly impact the fuel consumption of vehicles. To explore how well deep learning methods can predict fuel consumption precisely and efficiently and then guide drivers to go in an energy-saving way, we propose a fuel consumption prediction model, namely FuelNet, based on long short-term memory (LSTM) neural networks in this study. First, we develop the proposed FuelNet model with numerous vehicle kinematics data and corresponding fuel consumption data collected in the test field and real-world scenarios. And we analyze the relationship between the prediction accuracy and different combinations of input variables, training set size, and the sampling interval of the raw data. Second, we conduct intensive field tests to demonstrate the applicability of our model to fuel consumption prediction for different speed conditions and vehicle types. Furthermore, the superior prediction performance of FuelNet is shown by comparing it with five other types of models, such as the physical model, statistical and regression model, conventional neural networks model, and other deep learning models. Finally, we apply it to three real case studies, which verify that FuelNet can precisely predict fuel consumption for different driving trajectories in many scenarios such as signalized intersection (average value of RE is 0.049), campus environments (RE is 0.030), urban roads (RE is 0.077), and highways (RE is 0.097), as well as can contribute to detecting abnormal fuel consumption.

Traffic sign retroreflectivity is critical for nighttime visibility, an important factor in driver safety. Current methods of sign retroreflectivity assessment are expensive, time-consuming, dangerous, or subjective. There is an urgent need to explore an alternative method that is cost-effective, safe, objective, and can be operated during daytime or nighttime. One such method is mobile lidar. However, a methodology utilizing lidar cloud data for practical retroreflectivity condition assessment is still lacking because of the inability to numerically correlate lidar retro-intensity readings to the retroreflectivity standard set by the Manual on Uniform Traffic Control Devices (MUTCD). In addition, there is also a need to explore sign deterioration behavior using real-world lidar data. In this study, we (1) propose a practical, categorical traffic sign condition assessment using lidar data; (2) establish a preliminary correlation between the retro-intensity and retroreflectivity readings to determine the minimum retro-intensity thresholds for condition assessment of different sheeting types and colors; (3) validate the accuracy of the assessment by comparing it with standard nighttime visual inspection outcomes; (4) demonstrate the practical implementation through a feasibility study at Georgia Interstate 285; and (5) reveal the retro-intensity deterioration trends using historical lidar cloud data. The results show that the proposed methodology can reliably yield results comparable to manual measurements, potentially reducing sign retroreflectivity condition assessment effort, increasing the transportation agency’s productivity, and filling gaps where manual assessment is not possible. Additionally, the retro-intensity deterioration trends can help transportation agencies to understand the long-term behavior of sign retro-intensity and predict the optimal timing for sign replacement.

Short-term origin-destination (OD) demand predicting plays an indispensable role in intelligent transportation systems and ride-hailing service operations. However, most studies are carried out in trip ends travel demand prediction, i.e., trip generation/attraction, while paying less attention to predicting OD flows in citywide traffic management. To this end, a novel generalized framework is proposed, named two-stage fusion framework ($TFF$). $TFF$ consists of two stages, where the first stage utilizes an attention-based spatio-temporal graph convolutional network (AST-GCN) to predict the trip generation/attraction, and the second stage develops a modified Kalman filter (KF) to predict OD flow which is converted into a coefficient matrix. Lastly, the final predicted OD can be obtained by integrating the trip generation/attraction and the coefficient matrix. In AST-GCN, a gated fusion mechanism and dynamical zone proximity matrix are applied to improve the capacity of capturing the spatio-temporal interdependence among traffic analysis zones. In KF, we redefine the state vector and observed vector and their covariance matrixes, and then introduce the Box-Cox technique to standardize the deviation matrix to adapt to KF. The proposed model is tested on a large-scale real-world data set from Chongqing, China, and the results indicate that $TFF$ can achieve a satisfactory performance of short-term OD prediction. Moreover, the proposed framework has the convenient scalability, because either of the two stages is replaceable by other predictors.

Methods of optimizing offsets in coordinated traffic signal systems can be grouped into two categories: bandwidth based and flow based (e.g., minimizing delay). To our knowledge, no studies have directly compared methods from both categories. This study fills this gap with quantitative comparison between flow- and bandwidth-based methods. This paper evaluates the performance of 11 methods of offset optimization, including several different objectives that make use of arrival profiles (e.g., maximizing arrivals on green, minimizing delay, minimizing the number of stops) and several different methods of bandwidth maximization. These are compared against a default scenario with all offsets set to zero, and a manual method for determining offsets. An eight-intersection simulation network is used for comparison. The outcomes are compared in terms of the effective bandwidth, delay and stops, arrivals on green, and travel time. The results show that flow-based methods that incorporate models of delay and/or stops tend to yield more directionally balanced travel delays as well as lower total arterial delay and stops. Bandwidth-based methods achieved higher bandwidth and arrivals on green. Some additional comparisons are made among the 11 tested methods.

Abstract

The overtaking process on two-lane rural roads is a critical task for drivers and directly impacts safety. This paper investigates the vehicle’s passing path from the road-engineering point of view, consisting of a series of successive parametric spiral arcs, based on a real driving experiment with respect to the roadway’s posted speed. Passing maneuvers were recorded utilizing global navigation satellite system (GNSS) receivers. The curved paths were determined for two different and mostly typical posted speed values (90 and $110\mathrm{km}/\mathrm{h}$), where the impeding (passed) vehicle was assumed to travel under steady-state conditions ($20\mathrm{km}/\mathrm{h}$ below the respective posted speed values). The authors, by examining these curved paths, intend to quantify their trajectories and respective critical parameters, such as overtaking distance, headway distance, and acceleration performance. Four predictive lognormal models for each speed scenario were developed to estimate the radius of each overtaking curved section with high accuracy. Headway distance observed in this study is very close, while total overtaking distance is greater than those reported in the published literature. The present research is a source for road authorities to assess overtaking sight distance with critical overtaking parameters such as posted speed and vehicle speed differences as well as an opening paradigm of how the passing process can be standardized and therefore deployed in future advanced driver assistance systems.

Practical Applications

The provision of overtaking maneuvers on two-lane rural roads is regarded as a key safety priority during their geometric and operational design. The objective of the paper is to standardize the passing vehicles’ trajectories during the overtaking process. Besides quantifying the geometry of the overtaking successive curves, the authors aim to understand the potential degree of influence of other involved parameters, such as vehicle speed, acceleration, and longitudinal displacement of vehicles as well as the total overtaking length. The results revealed that the utilization of a certain mathematical curve (spiral), which is also applied in road design, delivers very accurate results. The impact of the other examined parameters also delivered interesting conclusions. The findings of the present analysis can be further analyzed by highway design agencies for adopting additional criteria to determine more accurate overtaking sight distances. In terms of vehicle automation, the paper mainly addresses cases where only the passing vehicle can be automated (the impeding vehicle conventional). For the examined speed difference value between the passing and the impeding vehicles, the research is an opening paradigm of how the passing process can be standardized and therefore deployed in existing advanced driver assistance systems (ADAS).