Journal of Transportation Engineering, Part A: Systems

Although fully implemented autonomous vehicles (AVs) seem to be on the cusp of reality, standard evaluation and testing procedures still are lacking. This study conducted a preliminary evaluation of the technical feasibility, safety, and reliability of using AV technology, in particular a low-speed, self-driving shuttle known as Olli. The study designed a set of 12 testing scenarios and performed experiments to evaluate the operational capabilities, safety, and reliability of the self-driving shuttle on the University at Buffalo’s Connected and Automated Vehicles (CAVs) proving grounds. The scenarios were designed to evaluate the vehicle’s performance while simulating the operational scenarios that the shuttle would encounter when deployed in the real world at a medical and educational campus in downtown Buffalo, New York. Preliminary results provide insight into the operational characteristics of the self-driving shuttle; its stopping distance behavior; its ability to detect and safely react to obstacles, conflicts, and other hazards on the road; its car-following behavior; and the impact of inclement weather conditions on performance.

Single-railcar freight shipments that move in multiple freight trains and are sorted at several classification yards between origin and destination remain an important source of traffic and revenue for North American railways. Despite the role of both mainlines and yards in freight rail transportation performance, little attention is devoted to investigations of classification yard performance, and few yard capacity models and tools are available. To address this need, this paper seeks to quantify the relative influence of different factors on several yard performance metrics. Simulation experiments conducted with YardSYM, a discrete-event simulation model developed specifically to analyze hump classification yards, examine varying the volume, number of blocks, block size distribution, number of outbound trains, train departure distribution, train arrival time variability, and arriving block volume variability. In addition to the expected sensitivity of yard performance to railcar throughput volume, the most influential factors change depending on the particular yard performance metric, emphasizing the criticality of matching performance metrics to railroad business objectives. The results of this research serve to screen variables for the future development of multivariable yard performance models and facilitate more informed business decisions regarding yard operating plans that make more efficient and economical use of existing yard capacity.

String stability analysis is one of key factors to understanding traffic flow dynamics. Cooperative adaptive cruise control (CACC) vehicles are desired to enhance stability of vehicular flow by monitoring multiple vehicles. From the perspective of mixed traffic, this study proposes a generalized analysis method for CACC feedback gains to achieve stable mixed flow. To deal with this, a random mixed flow is divided into general small platoons, in which one tail CACC monitors multiple vehicles ahead. These general platoons with different lengths are defined as connected vehicular systems. Based on the generalized car-following models, transfer function theory is employed to derive string stability criteria that can keep these connected vehicular systems stable for all possible vehicle speeds. A case study is also carried out to validate usability of the proposed generalized work. The proposed generalized method is applicable to various car-following models and large connected vehicular systems, thereby helping achieving the string stability enhancement of mixed traffic flow.

Speed prediction models are commonly developed using maximum and minimum operating speeds measured at or within specific locations on the tangents and horizontal curves, respectively. However, the actual distribution of the maximum and minimum operating speed positions on the entire lengths of the geometric elements (tangents and curves) have not been rigorously studied and, therefore, present opportunities to refine further and develop robust speed prediction models. This paper presents the probability distributions (normal, lognormal, gamma, and Weibull) for speed positions on the entire tangent and curve lengths using continuous speed data recorded on two-lane rural highways in India. The findings showed that the data could be best approximated for a large number of horizontal curves and tangents using the normal and Weibull distributions. Also, the results showed that the operating speeds measured at the midpoint of the horizontal curves overestimate the actual minimum operating speeds measured over the entire curve length of horizontal curves between 1.2 and $1.9\mathrm{km}/\mathrm{h}$. Similarly, maximum operating speeds measured at or within a 200-m length from the point of curvature into the approach tangents underestimate the maximum operating speeds measured over the entire length of tangents between 1.33 and $1.77\mathrm{km}/\mathrm{h}$. The results of this study highlight the importance of considering the entire length of geometric elements in developing accurate speed prediction models for use in evaluating highway design consistency.

The aim of this study was to evaluate and identify the factors influencing operating speed considering context classification. The study focused on three context classifications: C3R–Suburban Residential, C3C–Suburban Commercial, and C4–Urban General. Tobit models were proposed and developed using big data, including traffic and roadway characteristics, land-use attributes, and sociodemographic information. Three years of INRIX speed data were obtained to calculate the 85th-percentile speed. The study proposed an approach to adjust the 85th-percentile speed from INRIX data, given that traffic flow on arterials could be disrupted by signalized intersections. Afterward, empirical analysis was conducted by developing three Tobit models: Generic, C3C/C3R, and C4 models using the adjusted 85th-percentile speed. For the three developed models, several variables [e.g., inside shoulder width, speed limit, and number of signalized intersections per mile (1.609 km)] were found to have significant influence on the 85th-percentile speed. The analysis also revealed potential speed management countermeasures that have significant impact on the 85th-percentile speed which, when implemented, could reduce speed-related crashes and enhance the safety of vulnerable road users.

In recent years, many machine learning-based methods have emerged to detect faulty bearings. However, most of these methods may not be practical due to the need to collect a large number of fault samples for training. This paper developed a novel few-shot learning framework for the fault diagnosis of freight train rolling bearings. The proposed method has the capability to transfer the learning outcome from one bearing fault diagnosis model to another different but related task for which very limited training data are available. The authors established a single-wheelset platform to collect acceleration signals of different types of bearing faults. The authors preprocessed the data through data segmentation and frequency domain transformation, and divided the data into training and test sets according to a certain ratio. A one-dimensional convolutional neural network (1D-CNN) was established to automatically extract the features of the bearing vibration signals and classify the fault types. The authors implemented two few-shot learning methods through parameter fine-tuning and a conditional Wasserstein generative adversarial network (C-WGAN). A case study demonstrated the classification performance of the proposed models. The results showed that the diagnosis capability of the 1D-CNN in the frequency domain is significantly superior to that in the time domain. However, when the amount of data is small, the 1D-CNN model does not work. In contrast, the few-shot learning of bearing faults works well for both the fine-tuned CNN and C-WGAN models. Furthermore, the classification performance of the C-WGAN is better than that of the fine-tuned CNN when the training data are extremely limited.

Recent advances in vehicle-to-vehicle (V2V) communication technologies enable autonomous vehicle platoons to better utilize the information of other members in the platoon to achieve the desired longitudinal intervehicular spacing specifications. However, most existing studies adopt a fixed time headway $h_w$, and do not take into account the effects of V2V communication delay on the desired design specifications for the platoons. Moreover, for stability analysis, the commonly adopted Routh-Hurwitz approach usually needs a lot of computation time, which will significantly reduce the efficiency and flexibility of the platoons in real-time applications. In order to overcome the aforementioned issues with the existing results on autonomous vehicle platoons, this paper investigates the robust control and optimization problem for a platoon of autonomous and connected vehicles subject to V2V communication delay and uncertain parasitic actuation lag under a constant time headway policy (CTHP). First, by employing the Padé approximation approach, the spacing error propagation transfer function, from the preceding vehicle to the following vehicle, under immediate predecessor’s acceleration, velocity, and position feedback control mechanism is derived. In addition, in order to improve the computation efficiency, a signature method is utilized for the Hurwitz stability analysis. Then the string stability analysis for attenuating the spacing propagation error along the platoon is performed. Both lower and upper bounds for the time headway $h_w$ as well as the admissible sets for controller gains are achieved for Hurwitz stability and string stability. Comparative simulation results are finally provided to demonstrate the effectiveness of the proposed design scheme.

With the rapid increase of vehicle ownership in China, toll collection stations on expressways have become some of the most congested areas. Compared with a manual toll collection system (MTC), the electronic toll collection (ETC) system has advantages of rapidness and convenience. This paper comprehensively explored characteristics of the traffic flow of an ETC lane of the study area located on the Guizhou Expressway, China. The short-term traffic flows of ETC lanes are divided into low, moderate, and high volumes. In case of the low volume, this paper found that it is more reasonable to use Poisson distribution to predict the probability of ETC arrivals than to predict the ETC traffic flow. In case of the high volume, vehicles queue to pass the ETC lane, and ETC throughputs turn into a uniform distribution. This paper mainly focused on the moderate volume, discussed distribution of the ETC traffic flow, and predicted the ETC short-term traffic flow. Firstly, this paper applied the multiple Gaussians to fit the ETC traffic flow, and then used a Gaussian mixture model (GMM) to calculate the probability distribution of different Gaussian components. To verify the rationality of the distribution, Gaussian mixture regression (GMR) was utilized to predict short-term ETC traffic flow. The theoretical and experimental analyses demonstrated that when the ETC traffic volume is low, the prediction error of GMR is large, and with the ETC traffic volume increasing, GMM effectively can fit the distribution of the ETC short-term traffic flow and GMR can make accurate predictions. With the same parameters, GMM and GMR are capable of predicting about 74% of expressway toll stations’ ETC traffic flows in the study area. Moreover, the experimental results indicated that when the ETC traffic volume is greater than 15 passenger car units $(\mathrm{pcu})/5\min$, the GMR achieves a better forecasting performance than that of long short-term memory (LSTM) and the autoregressive integrated moving average (ARIMA). This in turn verified that the ETC traffic volumes will satisfy a GMM distribution, and GMR will be a better choice for forecasting short-term traffic flow.

Applying the electronic route map is a cornerstone for improving the management and operational efficiency of transportation systems. However, the practice of electronic route map designed specifically for the transit system is relatively scarce, because buses are still running and navigating in a complex environment. One major reason is that the online electronic map often fails to match the real-world bus network due to delayed updating of database. To address this issue, this paper proposes a method to fit the bus route trajectory curve based on global positioning system (GPS) data. A complete data processing methodology is established, including GPS trajectory similarity comparison, similar trajectories merging, and simplification of merged bus route trajectories. This paper innovatively proposes a merging algorithm named dynamic time warping–median-based fusion (DTW-MF), which aims to minimize the bias caused by outliers in GPS fusion. Additionally, the Douglas-Peucker algorithm is used for trajectory simplification, which smoothens the bus route trajectories and makes it easier to understand. A case study is conducted in a real-world bus system with 93 bus routes in Hangzhou, China. The results show that the proposed DTW-MF method outperforms the other two commonly used methods, namely DTW barycenter averaging (DBA) and prioritized shape averaging (PSA) algorithms, both in merging performance and computational efficiency.

Pedestrian crossing activity at unmarked crosswalks on urban roads in developing countries like India is a common phenomenon. Pedestrians used to cross the road at their place of interest to access residential and commercial land use. Such crossing operations cause a twofold effect: first the pedestrian put themselves at high risk of collision with the moving traffic, and they also affect the normal traffic operations, which causes excessive delay to vehicular traffic. Available pedestrian crossing warrants (PCWs) based on ${\mathrm{PV}}^2$ (where P is the volume of crossing pedestrians per hour and V is per-hourly traffic volume) focus on the safe movement of pedestrians, but traffic operation is least considered in such warrants. The present study deals with the development of pedestrian crossing facility warrants based on pedestrian safe movement along with vehicular delay, which includes traffic efficiency as well. The traffic data collected from four different locations through the videographic survey were used for evaluating the delay for the individual vehicle due to pedestrian crossing activity. The delay thus evaluated was further validated using the probe vehicle data collected during the same observation period. The vehicle was fitted with the GPS-based performance box instrument that provides the vehicle trajectory at any instance. A significant number of sample data were collected with the probe vehicle and delay was determined using trajectory data. The finding of the delay obtained from the field data was validated incorporating a simulation-based analysis to resemble the interaction of pedestrian flow dynamics with vehicular flows. The crossing behavior of pedestrians was simulated using microscopic simulation software VISSIM. The microsimulation model was first calibrated and validated for field traffic conditions. The total delay was assessed for the traffic stream, and it was found to follow a second-degree polynomial relationship with ${\mathrm{PV}}^2$ values. The clustering analysis was carried out using K-means clustering to define a threshold value for pedestrian crossing facilities. These warrants were also identified based on ${\mathrm{PV}}^2$ utilizing the relationship between total delay and ${\mathrm{PV}}^2$ values. The results of the present study will be useful for practicing engineers for designing pedestrian crossing facilities considering pedestrian safety and vehicular operation as well.