Smart Transportation Systems 2022

The complexity of the car-following dynamic traffic means the driver’s workload brought by the main driving task during the car-following process. The status information between the preceding and following vehicles on the Internet of Vehicles environment provides real-time quantification of the complex dynamic traffic environment. In order to quantify the complexity of the traffic environment in real time, three dynamic traffic environment complexity metrics based on the car-following behavior spectrum are used, which are Inversed Modified Margin to Collision, Transverse Oscillation Coefficient and Velocity Instability Coefficient. And we use the driving simulator to collect driver’s behavior data under different driving tasks during car-following, calculate the specific values of the three metrics, and use the entropy method to take a weighted sum of the three metrics. At last, we also use NASA-TLS subjective load scale to obtain the driver’s subject load, and this can verify the quantification model of the car-following dynamic traffic environment. The results show that the complexity quantification model of the car-following traffic environment can accurately quantify the driver’s workload brought by the main driving task to the driver during the car-following process. The study will provide a scientific basis for the control switching of vehicles between driver and machine and supply a more efficient driving behavior.

With the rapid development of voice technology, Voice recognition technology has been bought into a large number of vehicle information systems, and different information from HMI displays during voice interaction will affect the driving status of drivers. In order to improve the interaction between the driver and the HMI (Human machine interaction) display to reduce driving distraction and to optimize the HMI information display, this study compared the difficulty of using the three common voice interaction interfaces based on voice interaction background and quantified the secondary task load generated by the voice interaction interface on the driver during driving. In this study, the indicators of head movement and the operation parameters of the vehicle during voice interaction are proposed to quantify the driving task load based on the hierarchical analysis method. To achieve the goal, 10 drivers are selected for driving simulation tests by using the UC-win/road driving simulator and head data acquisition software. During the test, every tester performs three-voice interaction tasks: map navigation, phone calls, and switching music. The results show that the proportion of driving load generated by the map navigation phone calling and switching music are 0.4898, 0.1992, and 0.311. Therefore, the HMI information display interface of map navigation and music switching needs to be simplified designed to reduce the presentation of redundant information. The study will provide a scientific basis for the voice interaction function of the HMI system and the information display design of the HMI interface.

Various advanced driving assistance systems (ADAS) are designed to help drivers make reasonable evasive decisions in emergency conditions. However, different drivers have different perceptions of driving risks, resulting in differences in the choice of evasive behaviors (e.g., car-following, overtaking, etc.). To explorer the impact of ADAS on vehicle interaction behaviors under dangerous conditions, this study objectively assessed how ADAS affected drivers’ choice of evasive behaviors in near-crash events (NCEs). An on-road experiment was conducted on public highways in real traffic. The dangerous driving data of NCEs in different working states of ADAS (i.e., activated or not) were collected and combined to classify NCEs’ risk levels: lower, medium, and higher based on the K-means clustering results. Drivers’ evasive behaviors during ADAS engagement (i.e., during the use of the forward collision warning system) were then found closely associated with NCEs’ risk levels. Differences of probabilities for evasive behaviors were further compared within and across the three risk groups. Results showed that ADAS engagement has a positive impact on reducing the occurrence of NCEs and reduces the probability that drivers in medium and high-risk groups choose overtaking behavior. ADAS engagement can help drivers maintain a greater time headway (THW) when taking evasive action. Findings in this study could optimize ADAS intervention strategies and further enhance the ability of safe driving assistance.

In order to investigate the effect factors of route deviation when drives use navigation application on freeways, we conducted a route deviation questionnaire on 219 drivers. The driver cognitive patterns were summarized from the questionnaire by exploratory factor analysis, and effective explanatory variables were screened by chi-square test. Meanwhile, related social factors were screened by chi-square test and Tukey’s Honestly Significant Difference (HSD) test. Finally, a binary logistic regression model was established to reveal the influence weight of related factors on deviation from the route. The results show that navigation application design, navigation usage habits, traffic environment and education can influence drivers’ use of navigation, with navigation application design having the greatest degree of influence, followed by navigation usage habits, and then traffic environment interference and education. Specifically, navigation application that is consistent with most drivers’ habits can reduce the probability of route deviation; those with good driving habits have less route deviation; complex traffic environments can increase the probability of route deviation; the probability of route deviation decreases as education increases. The research contributes to the optimization of freeway navigation application.

As people’s demand for travel increases, sharing travel, especially carsharing, is receiving more and more attention and anticipation. In order to fully study the characteristics of customizable car sharing products and the demand tendency of consumers, on the basis of determining the advantages of customizable car sharing, this paper intends to explore and analyze the personalized needs of users of customizable car sharing products through a series of qualitative and quantitative research methods. A great deal of research has been carried out by scholars at home and abroad on the analysis of consumer needs and satisfaction, creating a good theoretical basis and reference for the research work in this paper. From the data obtained from the questionnaire survey, this paper summarized the customizable attribute items of product and their customization priorities and proposed an analysis method of personalized demand items based on the Kano model. Firstly, the Kano model combined with the fuzzy clustering method is used to identify and screen the personalized demand items of 24 initial car sharing products obtained from the survey, and the hierarchy of demand model of the product is constructed. Then, according to the personalized demand items screened by the Kano model, the questionnaire is designed for the importance survey. Based on the data obtained from the importance survey, the initial weights of the underlying personalized demand items of the hierarchy of demand model are calculated, and the entropy method is used to adjust the weights. After that, the importance ranking of 12 personalized demand items of car sharing products is determined. The results show that the security attribute has the highest overall importance among the 12 customized functions. This fully shows that the safety customization function of car sharing is most valued by consumers, and its demand and expectation are strong. The research results can provide certain guiding significance for customized production of car sharing companies.

To efficiently deter the repeated speeders who are frequently fined but continue to commit the violation, this study attempts to investigate the characteristics of the repeated speeding behaviors and propose an increasing block fine modeling approach. Based on the off-site law enforcement data collected from the Deyang City, three speeding ranges (low (≤20%), medium (20%–50%), and high (≥50%)) were considered, and then the characteristics of repeated speeding behaviors were extracted. After that, the cost-benefit theory was introduced to develop an increasing block fine model by taking into account the speeding range and frequency. Considering the minimum total number of speeding as the goal and the economic cost as the constraint, an optimization model of increasing block fine was developed. The pattern search algorithm was used to solve the developed optimization model to determine the best number of blocks and corresponding fine within each speeding range. Finally, a case study was conducted to validate the performance of the developed model. The results show that for the repeated speeders, the percent of the low speeding behavior is the highest, and the speeding behaviors largely occur during the daytime. Furthermore, within the increasing block fine mechanism, the reduction rates of medium and high speeding behaviors are clearly higher than that of the low speeding behavior. The findings of this study offer a fresh viewpoint on the repeated speeding intervention and enable the speeding enforcement more equitable.

To accurately estimate the capacity of multi-lane drop-off exclusive roadway, a calculation approach based on effective berths is proposed. First, the study specifically analyzes different layout types of parking and through lanes and introduces the concept of effective berths of drop-off exclusive roadway. A calculation approach of the capacity of multi-lane drop-off exclusive roadway at airport terminal based on effective berth and spatiotemporal trajectory theory is proposed. The effective berth under a certain level of delay is determined via VISSIM simulation and the drop-off exclusive roadway in Jinan Yaoqiang Airport-China is taken as an example for case analysis. The results show that the accuracy of the proposed method is more than 95%. The proposed approach is more in line with the actual unload process of the drop-off vehicles than traditional methods and lays a theoretical foundation for service level evaluation and the management and guidance of drop-off vehicles.

In recent years, many cities in the world are committed to promoting the electrification of public transportation. For bus companies, how to select the right charging facilities accurately and quickly has become an urgent problem to be solved. In this paper, we propose a simulation method based on Anylogic to describe the operation of electric buses under wired charging and wireless charging conditions. We provide decision-making suggestions for bus companies by analyzing the impact of wireless charging and wired charging on operation cost and passenger waiting time. According to the simulation results, we found that the waiting time of passengers under wired charging conditions is about 8.63% higher than that under wireless charging in the same operating conditions. The use of wireless charging facilities can effectively reduce the waiting time of passengers.

Autonomous Transport System (ATS) architectures enable a wide range of new applications and bring significant benefits to transport systems. However, during the design stage, errors of the architecture can have an impact on the smooth implementation of the ATS, which will endanger the normal operation of the transport systems. To ensure a high autonomy of the ATS architecture, i.e., “functionally evolvable, logically reconfigurable and physically configurable”, the detection of ATS architecture design errors is essential. This paper aims to fill the research gap in the existing research on diagnosing or evaluating ATS architectures. Inspired by word embedding models in natural language processing communities, we propose a data-driven approach to diagnose ATS architectures without prior knowledge or rules. We use an architecture embedding model to generate vector representations of ATS architectures, then train the model through negative sampling of the training dataset to identify the features of abnormal ATS architecture. Finally, we employ the trained model to classify structural errors of the test dataset generated from the ATS architecture. The experimental results show that the proposed method gains a relatively good effect of classifying with an average accuracy of 79.3%, demonstrating the effectiveness of the method.

Electric buses have the advantages of small noise, zero emission and simple control, which can effectively reduce urban pollutant emissions and energy consumption. Therefore, vigorously promoting the development of electric buses is of great significance to accelerate the low-carbon development of the city and realizes the goal of “carbon peak” and “carbon neutrality”. The operating condition of electric buses is an important factor affecting their energy consumption. Ensure buses under reasonable working conditions can improve the operation efficiency, reduce the operation energy consumption and the operation cost of bus enterprises. For the bus route with dedicated bus lane, we divide road sections based on road characteristics and analyze the operation state of buses on different units. Considering the constraints such as the bus travel punctuality rate between stations and the intersection delay rate, we taking the minimum operation energy consumption between stations as the optimization objective. Taking the traveling speed of the bus on the road section unit between stations and the starting moment of green light on the intersection as the optimization variables, the dynamic control model of the electric buses is established. Finally, the simulated annealing algorithm is used to solve the built model. Comparing the optimization scheme with the original scheme target value and analyzing the model optimization effect, the numerical results show that the total operating energy consumption can be saved up to 8.29%, which proves the optimal travel can meet the needs of passengers while reducing operating energy consumption.

Transport emission has become an increasingly serious problem, and it is an urgent issue in sustainable transport. In this study, by constructing traffic emission models for different vehicle types and operating conditions, the changes in CO, HC, and NOx emissions of light-duty and heavy-duty vehicles before and after signal control optimization were quantified based on VISSIM simulation. The OBEAS-3000 vehicle emission testing device was used to collect data on the micro-operational characteristics of different vehicles under different operating conditions as well as traffic emission data. Based on the data collected, the VSP (Vehicle Specific Power) model combined with the VISSIM traffic simulation platform was used to calculate the emissions of light and heavy vehicles in the mixed traffic flow before and after intersection signal optimization. It is known from the study that signal control optimization has a greater impact on heavy vehicles than on light vehicles. Emissions of CO, HC, and NOx from heavy vehicles and light vehicles are all reduced, but NOx emissions from light vehicles remain largely unchanged. The research results reveal the emission patterns of light and heavy vehicles in different micro-operating conditions and establish a traffic emission model. It provides a theoretical basis for accurate traffic emission analysis and traffic flow optimization, as well as a scientific basis for the formulation of traffic management measures and emission reduction in large city transport systems.

Hard-braking constitutes a critical surrogate measure of traffic safety on urban road networks. Efforts aiming to unveil the effects of contributing factors on the occurrence of hard-braking are inadequate. This study extracted the hard-braking event (HBE) and ordinary-braking event (OBE) by GPS trajectories from float cars. The effect of several factors on HBE was examined, including the factors of time, pre-braking behaviors, and road characteristics. The possibility of HBE was compared with that of OBE through binary logistic models (BLM). To further disclose the influence of factors, we considered the interaction between variables (BLM-VI) in modeling. The analysis results indicate that the BLM-VI is superior to the classical BLM in goodness-of-fit and factor interpretation. For the factors, peak hours on weekdays and daytime on weekends are positively linked to HBE, while driving at night is not. HBEs can be triggered by pre-braking behaviors such as speeding and approaching an intersection, but it is not likely to occur after changing lanes. Roads with work zones or intensive accesses can decrease the possibility of HBE. The factors of on-road parking, median divider, and one-way control have mixed effects on HBE when they interact with the factor of speed limits.

Swedish cities are embracing shared micro-mobility systems (SMMS) such as e-scooters sharing systems to promote sustainable travel behavior in urban contexts with corresponding infrastructure planning. SMMS is associated with various social, environmental, and economic benefits, as well as providing solutions for the first- and the last-mile problem of using public transit. This study analyzes the usage patterns of e-scooter systems, based on the scooter operation data of VOI company in Gothenburg, Sweden. The used data cover the transaction data of two and half months during the summer and include over five hundred thousand valid trip records. The result shows that most trips travel a distance between 0.5–1.8 km while the duration lasts between 4–7 min. Fridays and Saturdays are the most popular days while Sunday is the least popular day. The number of trips on Sundays decreases by about 60% compared to Fridays and Saturdays. Moreover, the e-scooters are used to varying degrees in the different areas of Gothenburg. The e-scooters are used at a much higher extent in central Gothenburg compared to areas outside the city center. This can be due to several different factors such as location, land use, and accessibility. Lastly, the results show that the e-scooters are not primarily used for commuting but rather for leisure, which can be seen in the average distance and duration of the entire zone as well as the temporal distribution.

Intelligent recognition of traffic road damage is essential for realizing smart vehicles and intelligent transportation systems. The classification of road material types before recognition is a challenge for traffic road damage recognition due to differences in features such as concrete and asphalt. In addition, the widely distributed roads make environmental factors a critical factor affecting the classification. In this paper, we propose a deep learning-based road material classification method that introduces an attention mechanism to deal with the influence of different environments on road material recognition. We acquired tens of thousands of road surface images for training and testing and performed practical validation in real roads. The experiments show that our method has high accuracy and recall in road material classification.

The efficiency of the roundabout is severely constrained by the presence of multiple conflicting traffic flows in the weaving section. An accurate description of the traffic flow in the weaving section of roundabouts is the key to intelligent traffic control. In this paper, the approach, the circulation section, the weaving section and the exit of roundabouts are taken as a unit to analyze, and a model of traffic flow for entering the weaving section is established. The impact of entering flow, circulation flow and leaving flow on the condition of the weaving section is analyzed and a model of traffic flow for leaving the weaving section is established. A new traffic flow model of the weaving section is also established considering an iterative-algorithm-based parameter calibration method. Finally, this paper builds a VISSIM-based simulation platform and collects field data to verify the accuracy of the traffic flow model. The results show that the average relative error of the simulation outputs and the values of theoretical model calculation is within 10%, and with the increase of passing vehicles, the calculation error decreases. Therefore, the model established in this paper can accurately calculate the traffic flow characteristics of the weaving section in roundabouts, which can provide theoretical support for the microscopic simulation of roundabouts and the formulation of traffic management strategies.

In recent years, vehicle-to-infrastructure (V2I) and vehicle-to-vehicle (V2V) communication are the leading technologies in intelligent transportation systems. With V2I and V2V, new strategies are enabled to better facilitate emergency rescue and traffic emergency management. Since signalized intersection is the major source of delay and congestion, how to prioritize emergency vehicles (EVs) at intersections becomes an indispensable research topic and has drawn extensive studies. Albeit that emergency vehicles should be granted absolute priority, the influence of EVs on the intersection was rarely studied, especially when aggressive preemption strategies are adopted. To this end, this paper is devoted to evaluating the impacts of EVs on the control of signalized intersections, as an effort to better emergency management. Mixed-integer linear programming models are developed to optimize the intersection control before and after the pass of EVs, respectively. Case studies indicate that, by assuming the EV will take exclusively one lane at the intersection, the average delay could increase by 11.28% along with a capacity decrease of 6.63%.

Exploring the spatial and temporal distribution of traffic violations is vital to road safety management. This study investigated the traffic violations of illegal parking and disobeying the guide lane, which are the most observed in a middle-sized city Luzhou. The temporal distributions between the traffic violations are compared in time of day, day of week, and month of year. The underlying spatial dependency and cluster of the violations are investigated by global Moran’s I, local Moran’s I, and kernel density. Results show that the frequency of illegal parking is remarkably higher in the period ranging from September to December, weekdays, morning, and afternoon. However, the violations of disobeying the guide lane are more frequently observed in the first-half year and daytime. Both two types of violation are positively correlated in space. Moreover, the density of illegal parking is higher in the upper area of the city where commercial and residential zone are common, while disobeying the guide lane is mostly found in several intersections which are close to the freeway exit. The possible explanations of the spatial and temporal distributions are discussed.

The development and advances in the domain of Autonomous Vehicles (AVs) provide disparate benefits including improved safety, enhanced reliability, and reduction in accidents (which saves thousands of human lives). However, the guarantee of safety is crucial for the successful deployment of AVs on the road. The safety-related issues of AVs are disparate in nature and are not solvable with a single technological solution, but require interaction and distributed responsibility across numerous interactive components, rendering the evaluation of safety concerns a challenge. We have presented an Assume/Guarantee (A/G) based contract for solving the safety-related concerns of AVs. Our preeminent focus is on the collision avoidance of AVs. Our approach is based on 1. the formal specification of A/G based contracts 2. creation of collision-based scenarios, 3. implementation of scenario-oriented controller implementing A/G contracts, 4. testing of controller implementing A/G contracts (allows the verification of contracts in terms of minimizing collision risks) based on the simulation performed in CARLA Scenic while considering the different (collision oriented) scenarios. Each contract is based on the extended assumption considering a specific scenario, referred as scenario-oriented contracts. The proposed methodology shows adequate results and proves that the Contract-Based Design (CBD) can provide a propitious road map for solving the safety-related concerns of AV.

The road network is becoming increasingly equipped with a multitude of sensors, monitoring a wide range of operating and contextual parameters. The availability of real-time sensor data enables the realisation of diverse data-driven applications, e.g., anomaly detection, identification of insightful patterns, monitoring the evolution of relevant trends in time and delivery of actionable decision support. However, such streaming data might contain vast amounts of missing values depending on the application. This makes it very challenging, if not impossible, to fully exploit the potential of data analysis and machine learning for these data sources, and in particular real-time analysis is not feasible. We propose in this paper an imputation methodology dedicated to multi-source streaming data, with a focus on the mobility domain. The proposed approach is based on spatio-temporal profiling of the streaming behaviour derived from historical data via non-negative matrix factorisation. The profiling method takes advantage of an adaptive segmentation strategy splitting the data into rolling time windows (chunks) allowing to use the limited non-missing data as optimally as possible. The identified profiles allow to devise a dynamic and scalable imputation strategy, which is able to reliably estimate incoming missing values in streaming data as soon as they arrive.