Smart Transportation Systems 2024

Urban logistics, merging the collaborative dynamics of public transit systems and Unmanned Aerial Vehicles (UAVs), heralds a new era in door-to-door delivery services. This innovative collaboration leverages the widespread reach of public transit and the flexibility of UAVs to increase delivery efficiency and reliability. UAVs adeptly handle the first/last mile, circumventing urban congestion and drastically cutting down delivery times. This cooperation not only drives a reduction in logistics operations’ carbon footprint but also marks a major step toward in sustainable urban development. This study delves into the first/last mile logistics challenge, proposing a framework where the regularity of public transit systems is harmoniously intertwined with UAVs’ flexibility. The numerical simulations indicate that this cooperation significantly bolsters delivery efficiency while concurrently reducing dependency on conventional urban delivery vehicles, achieving substantial energy and cost savings.

The integration of autonomous vehicles into traffic necessitates a comprehensive evaluation of elements such as infrastructure and traffic dynamics. This paper presents a structured framework designed to assess the suitability of autonomous driving by examining a range of environmental factors that form the operational design domain (ODD). It employs a combination of a qualitative survey and the Analytical Hierarchy Process to determine the significance of various indicators, culminating in the creation of a weighted sum model to evaluate drivability. Key factors influencing drivability include dynamic variables, such as pedestrian traffic and activities around schools. In a case study conducted in Uppsala, diverse data sources were amalgamated to create drivability maps. These maps provide preliminary insights into the feasibility of autonomous driving and contribute to the development of a specialized information layer for autonomous vehicles.

The paper aims to develop a set of key variables that can be used to model accessibility for informed and sustainable regional planning for sparsely populated remote areas. Based on a thematic analysis of empirical data from a Swedish case, the study extracts regional peculiarities related to the concept of regional accessibility. These peculiarities formulate a set of key variables that can be used to develop new accessibility models or expand existing ones to support decision-making at the policy design stage through a knowledge-based approach. The findings of the paper pinpoint characteristics of destinations that unpack the regional accessibility concept. The paper contributes with a fresh perspective on regional planning, addressing the often-overlooked distinct features of sparsely populated remote areas. Based on these insights, the suggested set of key variables reflects the regional peculiarities without neglecting the uniqueness of these areas. The extracted variables may be incorporated in modeling to simulate accessibility dynamics to offer contextually informed perspectives for policymakers. As a result, the findings from this paper strengthen informed decision-making concerning accessibility in planning for sustainable regional development.

In order to reduce the operating risks of commercial buses (CBs) and improve the safety supervision level of passenger transportation enterprises, this paper proposes a risk assessment method for CBs driving risk based on extension cloud mode. Firstly, based on historical GPS data of CBs and on-site surveys, we establish a road section driving risk evaluation system covering 10 indicators. Secondly, we refer to the relevant norms to classify the risk level into 5 levels and determine the indicator weights based on the game theory idea using the portfolio assignment method. Then, we construct an extension cloud model to determine the road traffic risk level of the road section with the principle of maximum affiliation. Finally, we verify the validity and accuracy of the model by taking an actual passenger route as an example. Comparison results show that the method proposed in this paper is scientific and reasonable, and all the results are highly credible, which can effectively identify the high-risk road sections of CBs operation, and provide a theoretical basis for the refined management of road passenger transport operation safety.

To investigate the passenger flow characteristics of electric bus routes and address the balance of demand problem between company and passengers, this paper proposes an integrated approach to operation period dividing and timetabling for a bus route. First, based on historical travel time data, an ordered clustering algorithm is used to divide the operating hours into time-of-day intervals. This approach incorporates constraints on travel time and passenger flow demand due to departure intervals, while considering the maximum number of vehicles permitted on a given route. With the objective of minimizing operation costs and passenger waiting time costs, we formulate a dispatching interval optimization model to determine the optimal dispatching intervals for the bus route. Finally, we utilize a one-dimensional search algorithm to solve the model. Case study is conducted to evaluate the performance of proposed method. Results indicate that compared to the current scheme, the optimized approach reduces the required number of buses by 14, with an increase in the full load rate on the highest section ranging between 0.58 and 9.59%.

In-service buses are susceptible to stochastic factors, specifically bus bunching, which can introduce unreliability. To address the ongoing bus bunching problem, we propose a comprehensive control strategy that incorporates speed guidance, intersection signal adjustments, and cooperative overtaking. A dynamic dispatching model based on the strategy is developed, with the objective function considering headway irregularity, the total travel time of passengers, and delay of the private vehicle. The model is solved using a genetic algorithm. In this paper, we conduct numerical experiments and compare the performance of the proposed strategy with the uncontrolled strategy. The results demonstrate that the proposed strategy can not only ensure the stability, but also improve the efficiency of the bus, thereby validating the performance of this method.

Mobility as a Service (MaaS) is a service that is designed to provide users with seamless mobility while also bringing about positive impacts on the economy, society, transportation, and the environment. In this study, a tourist spot is selected from a pre-determined group of destinations, and a complex domain optimization is performed using both quantitative and qualitative data. An on-demand bus route optimization system is also used, which considers the needs of the residents. The system then proposes a sightseeing route that is close to both the customer and the management, as well as a sightseeing route after the complex area optimization. This will lead to the development of a MaaS system for Level 4, which will be beneficial to all three parties involved: residents, customers, and the system management and operation side.

Large-scale urban traffic state estimation is essential in intelligent transportation systems (ITSs), particularly in applications like smart navigation and travel mode recommendations, where the precision of trajectory generation is of utmost importance. In this context, a generated trajectory refers to the macro-level path selection between an origin and a destination, tailored to incorporate real-time, personalized routing preferences that accommodate individual user needs and current traffic conditions. Nevertheless, existing studies frequently fail to account for the continuity of the generated trajectories, leading to an accumulation of errors, and often do not cater to personalized user requirements. This paper presents a framework based on Artificial Intelligence Generated Content (AIGC) to facilitate the generation of personalized, continuous trajectories that accurately mirror real-world conditions and user preferences, thereby avoiding the pitfalls of error accumulation. Departing from conventional grid-based spatial–temporal methods, our framework aligns generated trajectories directly with the actual road network and takes into account surrounding Points of Interest (POIs) that could influence travel decisions. Our approach offers a solution to users unsure about waypoint inclusion in their travel plans, greatly enhancing their experience by providing a range of flexible and personalized options. This represents a substantial advancement in the domain of personalized travel recommendations, signifying a transformative step in the evolution of ITSs.

The present-day technological advancements in the domain of autonomous vehicles (AVs) claim diversified benefits in terms of safety, reliability, and reduced rate of accidents. Providing safety guarantees for all AVs without exception in all possible situations is considered to be of paramount importance for the successful deployment of AVs on the road. The application of formal verification methods is suggested in this context to guarantee the safety of AVs. Therefore, these methods are computationally incompliant especially in the case of complex systems due to multitudinous interactive components specifically. Keeping this context in mind we have proposed an Assume/Guarantee (A/G)-based contract methodology for guaranteeing the safety of AVs. Our preeminent objective is to avoid collision between AVs during overtaking maneuvers by providing rigorous safety guarantees using Contract-Based Design (CBD). Our approach comprises the following steps: 1. formal specification of basic and updated, A/G contracts based on collision-specified scenarios, 2. creation of collision-based scenarios in CARLA scenic for testing the validity of the proposed approach, 3. implementation of scenario-oriented controller implementing A/G contracts referred as scenario oriented contracts, 4. testing of controller implementing basic and updated A/G contracts with the help of simulation performed in CARLA Scenic while considering the defined collision oriented scenarios. The first step followed in our proposed methodology for the safe overtaking maneuver of AVs is the implementation of basic A/G contracts covering existing speed limits: the maximum speed limit for AV is 60km/h. These basic A/G contracts are replaced by updated A/G-based contracts covering relaxed speed limits: the maximum speed limit for AV is 130km/h when the AV has a risk of collision. The proposed methodology leverages the A/G contracts and shows adequate results by proving that the CBD can provide a propitious road map for guaranteeing the safety of AVs during overtaking maneuvers.

Amidst rapid urbanization and evolving transport needs, electric scooters (e-scooters) have been reshaping short-distance urban trips. This study offers a systematic framework for spatial analysis of shared micro-mobility in Gothenburg, Sweden—using Geographically Weighted Regression (GWR) and Multiscale-GWR (MGWR) models. The research aims to decipher the city’s shared e-scooter demand for various factors such as transit proximity, land use patterns, road infrastructure, demographics, and weather conditions. Investigation comparatively deciphers GWR and MGWR models, which outperform global regression models in terms of fitness, and interpretability for the spatial heterogeneity in shared e-scooter demand. However, MGWR’s complexity sometimes leads to overfitting, with its results lacking clear interpretation. The study identifies significant spatial variations in shared e-scooter demand, linked with specific urban characteristics, providing a deeper understanding of how different factors contribute to shared e-scooter usage across various city zones. These findings are crucial for shared e-scooter ventures, urban planners and policymakers, offering a nuanced framework for integrating e-scooters into urban transport systems. The research underscores the effectiveness of spatial econometric approaches in urban mobility management, highlighting the importance of efficient spatial models for shared e-scooter demand analysis in urban contexts.

Predicting the collision risk is crucial for active traffic safety, as accurate collision risk prediction can help to take appropriate evasive behavior to effectively prevent rear-end collisions. Most existing studies have rarely analyzed the comprehensive impact of driving behavior and the driving environment on collision risk under different types of evasive behavior. To address this issue, this study extracts evasive events from vehicle trajectory data on highways and combines surrogate measures to quantify collision risk, aiming to establish a collision risk prediction model using CatBoost (Categorical Boosting) and SHAP (SHapley Additive exPlanation) methods. The results show that the proposed prediction model performed well, with accuracies of 90.36% and 91.77%, respectively. Due to the advantages of SHAP, the results of feature relative importance and specific impact analysis indicate that the influencing factors for collision risk vary among different types of evasive behavior, and the impact of road environment features cannot be ignored. The proposed model has the potential for implementation in ADASs to enhance their active safety capabilities.

The increasing adoption of battery-electric buses (BEBs) necessitates effective methods for managing their energy use, particularly in urban areas. This paper aims to accurately predict the State of Charge (SOC) of BEBs in Guangzhou, China, an essential factor for efficient energy management. We employed supervised machine learning techniques, specifically Random Forest and eXtreme Gradient Boosting (XGBoost), to develop predictive models for SOC. The study involved selecting eleven key features based on their inter-correlations to construct these models. The performance of the models was evaluated using the root mean square error (RMSE) metric. Results indicated an RMSE of 2.31 for Random Forest and 8.59 for XGBoost, demonstrating the effectiveness of these methods in predicting SOC with considerable accuracy. This research contributes to optimizing the operation of electric buses by providing reliable tools for SOC estimation, crucial for planning and managing urban public transportation systems.

As the dynamics of travel demands continue to shift, the accurate prediction of traffic conditions has become increasingly critical. This paper comprehensively charts the evolution of traffic flow prediction methodologies through four distinct phases. Initially, the focus was on assumptions and statistical methods. The second stage advanced to data-driven approaches with in-depth analysis of traffic data. Subsequently, machine and deep learning techniques were introduced, utilizing historical data for future predictions. The current stage explores the potential of Artificial Intelligence Generative Content (AIGC) approaches, including reinforcement learning and generative models for more precise strategies. This paper provides a structured reference for the field, outlining significant literature and advancements in traffic flow prediction.

To evaluate the robustness of emerging mixed traffic flow under snowy conditions from the perspective of crash risk, this study proposes a TTC (Time-to-Collision) dynamic threshold determination method based on extreme value theory to quantify the crash risk. This study simulates traffic flow during peak hours under continuous snowfall conditions in the SUMO software, and use eta-squared to evaluate the robustness of traffic flow under the various Market Penetration Rates (MPR). The results indicate that at different extreme quantiles, as MPR increases, eta-squared shows a downward trend. Especially under high MPR (80%), eta-squared can decrease by more than 45%. This result indicates that emerging mixed traffic flow under high MPRs can effectively improve the robustness under snowy conditions and reduce crash risk compared to traditional traffic flow.

Traffic signs on highways play a key role in providing road information and guiding drivers to drive safely. However, traditional reflective film signs have limited visibility at night, which can easily cause traffic accidents. In order to compare the visibility of traditional reflective film signs and LED active luminous signs, this study firstly collected the driver’s vehicle operating speed data and the driver’s visual recognition of the signs at the time point data through the real-vehicle experiments. Then, one-way ANOVA and independent samples Kruskal–Wallis test were used to investigate the differences and significance of the visual recognition distances between the two types of signs at night under different sign types, different speeds and different lane positions. The results show that LED active luminous signs have better and more stable visual recognition effects. When the vehicle speed is higher, the driver's visual recognition distance of LED active luminous signs is longer. Moreover, there is no significant difference in the visual recognition effect of LED active illuminated signs when they are located in different lane positions.

To quantitatively analyze the driving forgiveness of highways, this study incorporates the varying psychological impacts of different road alignments on drivers. Heart rate increment is adopted as the evaluation index, and physiological data of drivers is collected during road experiments. The experimental data are processed and analyzed using MATLAB to investigate the relationship between driving speed and heart rate increment. It is determined that a nonlinear strong correlation exists between the two variables. Accordingly, a polynomial regression model is constructed to represent this relationship. Furthermore, by integrating heart rate inflection points and other relevant information, evaluation criteria for roadway forgiveness are established with heart rate increment as the key index.

Shared e-scooters (SES), as a crucial part of shared micro-mobility services, are increasingly recognized as an eco-friendly transportation option. Understanding the factors influencing SES usage and how it replaces traditional transport modes is key to sustainable mobility planning. We conducted a stated preference experiment to explore travel choices for SES, considering different alternatives, contexts, and user characteristics. Using a multinomial logit model on survey data from Gothenburg, Sweden, we found that SES could potentially decrease reliance on private cars. The results also indicate a lower preference for SES among private car users, compared to public transport and active mode users. We discuss operational recommendations for micro-mobility planners and operators to mitigate any unintended substitution impacts.