Search Results (3,335)

To enhance road-testing safety for autonomous driving robotic vehicles (ADRVs), collision avoidance with sudden obstacles is essential during testing processes. This paper proposes an upper-level collision avoidance strategy integrating model predictive control (MPC) and improved artificial potential field (APF). The kinematic model of the driving robot is established, and a vehicle dynamics model considering road curvature is used as the foundation for vehicle control. The improved APF constraints are constructed. The boundary constraint uses a three-circle vehicle shape suitable for roads with arbitrary curvatures. A unified obstacle potential field constraint is designed for static/dynamic obstacles to generate collision-free trajectories. An auxiliary attractive potential field is designed to ensure stable trajectory recovery after obstacle avoidance completion. A multi-objective MPC framework coupled with artificial potential fields is designed to achieve obstacle avoidance and trajectory tracking while ensuring accuracy, comfort, and environmental constraints. Results from Carsim-Simulink and semi-physical experiments validate that the proposed strategy effectively avoids various obstacles under different road conditions while maintaining reference trajectory tracking. Full article

To reveal the influence mechanisms of seasonal climatic factors (wind speed, wind direction, temperature) and leakage direction on hydrogen dispersion and explosion behavior from single-source leaks at typical risk locations (hydrogen storage tanks, compressors, dispensers) in hydrogen refueling stations (HRSs), this work established a full-scale 1:1 three-dimensional numerical model using the FLACS v22.2 software based on the actual layout of an HRS in Xichang, Sichuan Province. Through systematic simulations of 72 leakage scenarios (3 equipment types × 4 seasons × 6 leakage directions), the coupled effects of climatic conditions, equipment layout, and leakage direction on hydrogen dispersion patterns and explosion risks were quantitatively analyzed. The key findings indicate the following: (1) Downward leaks (−Z direction) from storage tanks tend to form large-area ground-hugging hydrogen clouds, representing the highest explosion risk (overpressure peak: 0.25 barg; flame temperature: >2500 K). Leakage from compressors (±X/−Z directions) readily affects adjacent equipment. Dispenser leaks pose relatively lower risks, but specific directions (−Y direction) coupled with wind fields may drive significant hydrogen dispersion toward station buildings. (2) Southeast/south winds during spring/summer promote outward migration of hydrogen clouds, reducing overall station risk but causing localized accumulation near storage tanks. Conversely, north/northwest winds in autumn/winter intensify hydrogen concentrations in compressor and station building areas. (3) An empirical formula integrating climatic parameters, leakage conditions, and spatial coordinates was proposed to predict hydrogen concentration (error < 20%). This model provides theoretical and data support for optimizing sensor placement, dynamically adjusting ventilation strategies, and enhancing safety design in HRSs. Full article

Accurately identifying driver fatigue in complex driving environments plays a crucial role in road traffic safety. To address the challenge of reduced fatigue detection accuracy in complex cabin environments caused by lighting variations, we propose YOLO-FDCL, a novel algorithm specifically designed for driver fatigue detection under complex lighting conditions. This algorithm introduces MobileNetV4 into the backbone network to enhance the model’s ability to extract fatigue-related features in complex driving environments while reducing the model’s parameter size. Additionally, by incorporating the concept of structural re-parameterization, RepFPN is introduced into the neck section of the algorithm to strengthen the network’s multi-scale feature fusion capabilities, further improving the model’s detection performance. Experimental results show that on the YAWDD dataset, compared to the baseline YOLOv8-S, precision increased from 97.4% to 98.8%, recall improved from 96.3% to 97.5%, mAP@0.5 increased from 98.0% to 98.8%, and mAP@0.5:0.95 increased from 92.4% to 94.2%. This algorithm has made significant progress in the task of fatigue detection under complex lighting conditions. At the same time, this model shows outstanding performance on our self-developed Complex Lighting Driving Fatigue Dataset (CLDFD), with precision and recall improving by 2.8% and 2.2%, respectively, and improvements of 3.1% and 3.6% in mAP@0.5 and mAP@0.5:0.95 compared to the baseline model, respectively. Full article

The safe and reliable operation of autonomous vehicles is contingent on comprehensive testing. However, the operational scenarios are inexhaustible. Corner cases, which critically influence autonomous vehicle safety, occur at an extremely low probability and follow a long-tail distribution. Corner cases can be defined as combinations of driving task and scenario elements. These scenarios are characterized by low probability, high risk, and a tendency to reveal functional limitations inherent to autonomous driving systems, triggering anomalous behavior. This study constructs a novel corner case dataset using naturalistic driving data, specifically tailored for autonomous vehicle testing. A scenario marginality quantification method is designed to analyze multi-source naturalistic driving data, enabling efficient extraction of corner cases. Heterogeneous scenarios are systematically transformed, resulting in a dataset characterized by diverse interaction behaviors and standardized formatting. The results indicate that the scenario marginality of the dataset constructed in this study is 2.78 times that of mainstream naturalistic driving datasets, and the scenarios exhibit considerable diversity. The trajectory and velocity fluctuations, quantified at 0.013 m and 0.021 m/s, respectively, are consistent with the kinematic characteristics of real-world driving scenarios. These results collectively demonstrate the dataset’s high marginality, diversity, and applicability. Full article

Light detection and ranging (LiDAR) provides three-dimensional environmental information that is critical for maintaining the safety and reliability of autonomous driving systems. However, dust accumulation on the LiDAR window can cause detection errors and degrade performance. This study determined the optimal spray conditions for accumulating dust to evaluate LiDAR sensor cleaning performance. A primary optimization experiment using spray pressure, spray speed, spray distance, and the number of sprays as variables showed that spray pressure and number of sprays had the most significant influence on the kinetic energy and distribution of dust particles. Notably, the interaction between spray distance and number of sprays—related to curvature effects—was identified as a key variable increasing process sensitivity. A supplementary experiment, which added spray angle as a variable, indicated that while spray pressure remained the most significant factor, spray angle and number of sprays had an indirect influence through interaction terms. Both experiments used the same response variable (point cloud data) interactions to stepwise analyze particle transfer and spatial diffusion. The resulting optimal conditions offer a standard basis for evaluating LiDAR cleaning performance and may help improve cleaning efficiency and maintenance strategies. Full article

In the era of digital intelligence, digital media landscapes increasingly influence cultural tourism consumption. Consumerism capitalizes on tourists’ superficial aesthetic commonalities, constructing a homogenized media imagination that leads to collective convergence in travel decisions, which obscures aspects of local culture, poses safety risks, and results in fleeting local tourism booms. In this study, semistructured interviews were conducted with 36 tourists, and NVivo12.0 was used for three-level node coding in a qualitative analysis to explore the digital media attributions of conformist travel behavior. The findings indicate that digital media landscapes exert a “digital persuasion” effect by reconstructing self-experience models, directing the individual gaze, and projecting idealized self-images. These mechanisms drive tourists to follow digital traffic trends and engage in imitative behaviors, ultimately shaping the phenomenon of “group roaming”, grounded in the psychological effect of herd behavior. Full article

Emerging transportation challenges necessitate cutting-edge technologies for real-time infrastructure and traffic monitoring. To create a dynamic digital twin for intersection monitoring, data gathering, performance assessment, and calibration of microsimulation software, this study presents a state-of-the-art platform that combines high-resolution LiDAR sensor data with VISSIM simulation software. Intending to track traffic flow and evaluate important factors, including congestion, delays, and lane configurations, the platform gathers and analyzes real-time data. The technology allows proactive actions to improve safety and reduce interruptions by utilizing the comprehensive information that LiDAR provides, such as vehicle trajectories, speed profiles, and lane changes. The digital twin technique offers unparalleled precision in traffic and infrastructure state monitoring by fusing real data streams with simulation-based performance analysis. The results show how the platform can transform real-time monitoring and open the door to data-driven decision-making, safer intersections, and more intelligent traffic data collection methods. Using the proposed platform, this study calibrated a VISSIM simulation network to optimize the driving behavior parameters in the software. This study addresses current issues in urban traffic management with real-time solutions, demonstrating the revolutionary impact of emerging technology in intelligent infrastructure monitoring. Full article

Vehicle-to-everything (V2X) communication is a fundamental technology in the development of intelligent transportation systems, encompassing vehicle-to-vehicle (V2V), infrastructure (V2I), and pedestrian (V2P) communications. This technology enables connected and autonomous vehicles (CAVs) to interact with their surroundings, significantly enhancing road safety, traffic efficiency, and driving comfort. However, as V2X communication becomes more widespread, it becomes a prime target for adversarial and persistent cyberattacks, posing significant threats to the security and privacy of CAVs. These challenges are compounded by the dynamic nature of vehicular networks and the stringent requirements for real-time data processing and decision-making. Much research is on using novel technologies such as machine learning, blockchain, and cryptography to secure V2X communications. Our survey highlights the security challenges faced by V2X communications and assesses current ML and blockchain-based solutions, revealing significant gaps and opportunities for improvement. Specifically, our survey focuses on studies integrating ML, blockchain, and multi-access edge computing (MEC) for low latency, robust, and dynamic security in V2X networks. Based on our findings, we outline a conceptual framework that synergizes ML, blockchain, and MEC to address some of the identified security challenges. This integrated framework demonstrates the potential for real-time anomaly detection, decentralized data sharing, and enhanced system scalability. The survey concludes by identifying future research directions and outlining the remaining challenges for securing V2X communications in the face of evolving threats. Full article

The accurate perception of visual stimuli in human–machine systems is crucial for improving system safety, usability, and task performance. The widespread adoption of digital technology has significantly increased the importance of visual interfaces and information. Therefore, it is essential to design visual interfaces and information with user characteristics in mind to ensure accurate perception of visual information. This study employed the Cognitive Perceptual Assessment for Driving (CPAD) to evaluate and compare gender differences in the ability to perceive visual signals within complex visual stimuli. The experimental setup included a computer with CPAD installed, along with a touch monitor, mouse, joystick, and keyboard. The participants included 11 male and 20 female students, with an average age of 22 for males and 21 for females. Prior to the experiment, participants were instructed to determine whether a signal stimulus was present: if a square, presented as the signal, was included in the visual stimulus, they moved the joystick to the left; otherwise, they moved it to the right. Each participant performed a total of 40 trials. The entire experiment was recorded on video to measure overall response times. The experiment measured the number of correct detections of signal presence, response times, the number of misses (failing to detect the signal when present), and false alarms (detecting the signal when absent). The analysis of experimental data revealed no significant differences in perceptual ability or response times for visual stimuli between genders. However, males demonstrated slightly superior perceptual ability and marginally shorter response times compared to females. Analyses of sensitivity and response bias, based on signal detection theory, also indicated a slightly higher perceptual ability in males. In conclusion, although these differences were not statistically significant, males demonstrated a slightly better perception ability for visual stimuli. The findings of this study can inform the design of information, user interfaces, and visual displays in human–machine systems, particularly in light of the recent trend of increased female participation in the industrial sector. Future research will focus on diverse types of visual information to further validate these findings. Full article

This paper introduces a robust and efficient Localized Spline-based Path-Planning (LSPP) algorithm designed to enhance autonomous vehicle navigation on highways. The LSPP approach prioritizes smooth maneuvering, obstacle avoidance, passenger comfort, and adherence to road constraints, including lane boundaries, through optimized trajectory generation using quintic spline functions and a dynamic speed profile. Leveraging real-time data from the vehicle’s sensor fusion module, the LSPP algorithm accurately interprets the positions of surrounding vehicles and obstacles, creating a safe, dynamically feasible path that is relayed to the Model Predictive Control (MPC) track-following module for precise execution. The theoretical distinction of LSPP lies in its modular integration of: (1) a finite state machine (FSM)-based decision-making layer that selects maneuver-specific goal states (e.g., keep lane, change lane left/right); (2) quintic spline optimization to generate smooth, jerk-minimized, and kinematically consistent trajectories; (3) a multi-objective cost evaluation framework that ranks competing paths according to safety, comfort, and efficiency; and (4) a closed-loop MPC controller to ensure real-time trajectory execution with robustness. Extensive simulations conducted in diverse highway scenarios and traffic conditions demonstrate LSPP’s effectiveness in delivering smooth, safe, and computationally efficient trajectories. Results show consistent improvements in lane-keeping accuracy, collision avoidance, enhanced materials wear performance, and planning responsiveness compared to traditional path-planning methods. These findings confirm LSPP’s potential as a practical and high-performance solution for autonomous highway driving. Full article

In intelligent vehicular networks, the accuracy of semantic segmentation in road scenes is crucial for vehicle-mounted artificial intelligence to achieve environmental perception, decision support, and safety control. Although deep learning methods have made significant progress, two main challenges remain: first, the difficulty in balancing global and local features leads to blurred object boundaries and misclassification; second, conventional convolutions have limited ability to perceive irregular objects, causing information loss and affecting segmentation accuracy. To address these issues, this paper proposes a global–local collaborative attention module and a spider web convolution module. The former enhances feature representation through bidirectional feature interaction and dynamic weight allocation, reducing false positives and missed detections. The latter introduces an asymmetric sampling topology and six-directional receptive field paths to effectively improve the recognition of irregular objects. Experiments on the Cityscapes, CamVid, and BDD100K datasets, collected using vehicle-mounted cameras, demonstrate that the proposed method performs excellently across multiple evaluation metrics, including mIoU, mRecall, mPrecision, and mAccuracy. Comparative experiments with classical segmentation networks, attention mechanisms, and convolution modules validate the effectiveness of the proposed approach. The proposed method demonstrates outstanding performance in sensor-based semantic segmentation tasks and is well-suited for environmental perception systems in autonomous driving. Full article

Intelligent driving technology is expected to reshape urban transportation, but its promotion is hindered by user acceptance challenges and diverse technical routes. This study proposes a comprehensive user acceptance evaluation framework for intelligent driving from the perspective of value engineering (VE). The novelty of this framework lies in three aspects: (1) It unifies behavioral theory and utility theory under the value engineering framework, and it extracts key indicators such as safety, travel efficiency, trust, comfort, and cost, thus addressing the issue of the lack of integration between subjective and objective factors in previous studies. (2) It establishes a systematic mapping mechanism from technical solutions to evaluation indicators, filling the gap of insufficient targeting at different technical routes in the existing literature. (3) It quantifies acceptance differences via VE’s core formula of V = F/C, overcoming the ambiguity of non-technical evaluation in prior research. A case study comparing single-vehicle intelligence vs. collaborative intelligence and different sensor combinations (vision-only, map fusion, and lidar fusion) shows that collaborative intelligence and vision-based solutions offer higher comprehensive acceptance due to balanced functionality and cost. This framework guides enterprises in technical strategy planning and assists governments in formulating industrial policies by quantifying acceptance differences across technical routes. Full article

With the continuous increase in urban population, the subway is the main way to alleviate traffic congestion. However, the construction environment of subway stations is complex, and the safety risks are extremely high. Therefore, it is of great practical significance to scientifically and systematically evaluate the safety vulnerability of subway station construction. This paper takes the Chengdu subway project as an example, and establishes a metro station construction safety vulnerability evaluation index system based on the driving forces–pressures–state–impacts–responses (DPSIR) theory with 5 first-level indexes and 23 second-level indexes, and adopts the fuzzy hierarchical analysis method (FAHP) to calculate the subjective weights, and the improved Harris Hawks optimization–projection pursuit method (HHO-PPM) to determine the objective weights, combined with game theory to calculate the comprehensive weights of the indicators, and finally uses the improved cloud model of Bayesian feedback to determine the vulnerability level of subway station construction safety. The study found that the combined empowerment–improvement cloud model assessment method is reliable, and the case study verifies that the vulnerability level of the project is “very low risk”, and the investigations of safety hazards and the pressure of surrounding traffic are the key influencing factors, allowing for the proposal of more scientific and effective management strategies for the construction of subway stations. Full article

High-viscosity stratified strata, characterized by complex geotechnical properties such as strong cohesion, low permeability, and pronounced layered structures, exhibit significant lateral friction resistance and high-end resistance during steel sheet pile installation. These factors substantially increase construction difficulty and may even cause structural damage. This study addresses two critical mechanical challenges during vibratory pile driving in Fujian Province’s hydraulic engineering project: prolonged high-frequency driving durations, and severe U-shaped steel sheet pile head damage in high-viscosity stratified soils. Employing the Coupled Eulerian–Lagrangian (CEL) numerical method, a systematic investigation was conducted into the penetration resistance, stress distribution, and damage patterns during vibratory pile driving under varying conditions of cohesive soil layer thickness, predrilled hole spacing, and aperture dimensions. The correlation between pile stress and penetration depth was established, with the influence mechanisms of key factors on driving-induced damage in high-viscosity stratified strata under multi-factor coupling effects elucidated. Finally, the feasibility of predrilling techniques for resistance reduction was explored. This study applies the damage prediction model based on the CEL method to U-shaped sheet piles in high-viscosity stratified formations, solving the problem of mesh distortion in traditional finite element methods. The findings provide scientific guidance for steel sheet pile construction in high-viscosity stratified formations, offering significant implications for enhancing construction efficiency, ensuring operational safety, and reducing costs in such challenging geological conditions. Full article

Given the severe global traffic safety challenges—including threats to human lives and socioeconomic impacts—this study analyzes visual behavior to promote sustainable transportation, improve road safety, and reduce resource waste and pollution caused by accidents. Four typical road sections, namely, turning, straight ahead, uphill, and downhill, were selected, and the eye movement data of 23 drivers in different driving stages were collected by aSee Glasses eye-tracking device to analyze the visual gaze characteristics of the drivers and their transfer patterns in each road section. Using Markov chain theory, the probability of staying at each gaze point and the transfer probability distribution between gaze points were investigated. The results of the study showed that drivers’ visual behaviors in different road sections showed significant differences: drivers in the turning section had the largest percentage of fixation on the near front, with a fixation duration and frequency of 29.99% and 28.80%, respectively; the straight ahead section, on the other hand, mainly focused on the right side of the road, with 31.57% of fixation duration and 19.45% of frequency of fixation; on the uphill section, drivers’ fixation duration on the left and right roads was more balanced, with 24.36% of fixation duration on the left side of the road and 25.51% on the right side of the road; drivers on the downhill section looked more frequently at the distance ahead, with a total fixation frequency of 23.20%, while paying higher attention to the right side of the road environment, with a fixation duration of 27.09%. In terms of visual fixation, the fixation shift in the turning road section was mainly concentrated between the near and distant parts of the road ahead and frequently turned to the left and right sides; the straight road section mainly showed a shift between the distant parts of the road ahead and the dashboard; the uphill road section was concentrated on the shift between the near parts of the road ahead and the two sides of the road, while the downhill road section mainly occurred between the distant parts of the road ahead and the rearview mirror. Although drivers’ fixations on the front of the road were most concentrated under the four road sections, with an overall fixation stability probability exceeding 67%, there were significant differences in fixation smoothness between different road sections. Through this study, this paper not only reveals the laws of drivers’ visual behavior under different driving environments but also provides theoretical support for behavior-based traffic safety improvement strategies. Full article