Search Results (866)

With the rapid advancement of autonomous driving technology, Traffic Sign Detection and Recognition (TSDR) has become a critical component for ensuring vehicle safety. However, existing TSDR systems still face significant challenges in accurately detecting partially occluded traffic signs, which poses a substantial risk in real-world applications. To address this issue, this study proposes a comprehensive solution from three perspectives: data augmentation, model architecture enhancement, and dataset construction. We propose an innovative network framework tailored for occluded traffic sign detection. The framework enhances feature representation through a dual-path convolutional mechanism (DualConv) that preserves information flow even when parts of the sign are blocked, and employs a spatial attention module (SEAM) that helps the model focus on visible sign regions while ignoring occluded areas. Finally, we construct the Jinzhou Traffic Sign (JZTS) occlusion dataset to provide targeted training and evaluation samples. Extensive experiments on the public Tsinghua-Tencent 100K (TT-100K) dataset and our JZTS dataset demonstrate the superior performance and strong generalisation capability of our model under occlusion conditions. This work not only advances the robustness of TSDR systems for autonomous driving but also provides a valuable benchmark for future research. Full article

Recent advances in Large Vision–Language Models (LVLMs) have demonstrated strong cross-modal reasoning capabilities, offering new opportunities for decision-making in autonomous driving. However, existing end-to-end approaches still suffer from limited semantic consistency, weak task controllability, and insufficient interpretability. To address these challenges, we propose SemAlign-E2E (Semantic-Aligned End-to-End), a semantic-aligned multimodal LVLM framework that unifies visual, LiDAR, and task-oriented textual inputs through cross-modal attention. This design enables end-to-end reasoning from scene understanding to high-level driving command generation. Beyond producing structured control instructions, the framework also provides natural-language explanations to enhance interpretability. We conduct extensive evaluations on the nuScenes dataset and CARLA simulation platform. Experimental results show that SemAlign-E2E achieves substantial improvements in driving stability, safety, multi-task generalization, and semantic comprehension, consistently outperforming state-of-the-art baselines. Notably, the framework exhibits superior behavioral consistency and risk-aware decision-making in complex traffic scenarios. These findings highlight the potential of LVLM-driven semantic reasoning for autonomous driving and provide a scalable pathway toward future semantic-enhanced end-to-end driving systems. Full article

Sensor degradation poses a significant challenge in autonomous driving. During heavy rainfall, interference from raindrops can adversely affect the quality of LiDAR point clouds, resulting in, for instance, inaccurate point measurements. This, in turn, can potentially lead to safety concerns if autonomous driving systems are not weather-aware, i.e., if they are unable to discern such changes. In this study, we release a new, large-scale, multi-modal emulated rain dataset, REHEARSE-3D, to promote research advancements in 3D point cloud de-raining. Distinct from the most relevant competitors, our dataset is unique in several respects. First, it is the largest point-wise annotated dataset (9.2 billion annotated points), and second, it is the only one with high-resolution LiDAR data (LiDAR-256) enriched with 4D RADAR point clouds logged in both daytime and nighttime conditions in a controlled weather environment. Furthermore, REHEARSE-3D involves rain-characteristic information, which is of significant value not only for sensor noise modeling but also for analyzing the impact of weather at the point level. Leveraging REHEARSE-3D, we benchmark raindrop detection and removal in fused LiDAR and 4D RADAR point clouds. Our comprehensive study further evaluates the performance of various statistical and deep learning models, where SalsaNext and 3D-OutDet achieve above $94\%$ IoU for raindrop detection. Full article

The rapid advancement of autonomous vehicle systems (AVS) has introduced complex challenges to road safety. While some studies have investigated the contribution of factors influencing AV-involved crashes, few have focused on the impact of vehicle-specific factors within AVS on crash outcomes, a focus that gains importance due to the absence of a human driver. To address this gap, the advanced machine learning algorithm, LightGBM (v4.4.0), is employed to quantify the potential effects of vehicle factors on crash severity and collision types based on the Autonomous Vehicle Operation Incident Dataset (AVOID). The joint effects of different vehicle factors and the interactive effects of vehicle factors and environmental factors are studied. Compared with other frequently utilized machine learning techniques, LightGBM demonstrates superior performance. Furthermore, the SHapley Additive exPlanation (SHAP) approach is employed to interpret the results of LightGBM. The analysis of crash severity revealed the importance of investigating the vehicle characteristics of AVs. Operator type is the most predictive factor. For road types, highways and streets show a positive association with the model’s prediction of serious crashes. Crashes involving vulnerable road users can be attributed to different factors. The road type is the most significant factor, followed by precrash speed and mileage. This study identifies key predictive associations for the development of safer AV systems and provides data-driven insights to support regulatory strategies for autonomous driving technologies. Full article

To achieve the automation and intelligence of mining equipment, it is essential to address the challenge of autonomous driving, with the core task being how to navigate safely from the starting point to the mining area endpoint. This paper proposes a boundary-aware multi-point preview control algorithm to tackle the strong dependency on predefined paths and the lack of foresight in the autonomous driving of underground articulated mining vehicles in highly confined underground spaces. The algorithm determines the driving direction by calculating the vehicle’s real-time state and LiDAR data, previewing road conditions without relying on preset path planning. Experiments conducted in a ROS Noetic/GAZEBO 11 simulation environment compared the proposed method with single-point and two-point preview algorithms, validating the effectiveness of the boundary-aware multi-point preview control. The results show that the proposed control strategy yields the lowest lateral deviation and the highest steering smoothness compared to single-point and two-point preview algorithms; it also outperforms the standard multi-point preview algorithm. This demonstrates its superior performance. Specifically, the proposed boundary-aware multi-point preview algorithm outperformed other methods in terms of steering smoothness and stability, significantly enhancing the vehicle system’s adaptability, robustness, and safety. Full article

The advancement of automated driving technologies offers potential safety and efficiency gains, yet safety remains the primary barrier to higher-level deployment. Failures in automated driving systems rarely result from a single technical malfunction. Instead, they emerge from coupled organizational, technical, human, and environmental factors, particularly in partial and conditional automation where human supervision and intervention remain critical. This study systematically identifies safety failures in automated driving systems and analyzes how they propagate across system layers and human–machine interactions. A qualitative case-based analytical approach is adopted by integrating the Swiss Cheese model and the SHELL model. The Swiss Cheese model is used to represent multilayer defensive structures, including governance and policy, perception, planning and decision-making, control and actuation, and human–machine interfaces. The SHELL model structures interaction failures between liveware and software, hardware, environment, and other liveware. The results reveal recurrent cross-layer failure pathways in which interface-level mismatches, such as low-salience alerts, sensor miscalibration, adverse environmental conditions, and inadequate handover communication, align with latent system weaknesses to produce unsafe outcomes. These findings demonstrate that autonomous driving safety failures are predominantly socio-technical in nature rather than purely technological. The proposed hybrid framework provides actionable insights for system designers, operators, and regulators by identifying critical intervention points for improving interface design, operational procedures, and policy-level safeguards in autonomous driving systems. Full article

Intelligent vehicle networks powered by machine learning, AI and blockchain are transforming various sectors beyond transportation. In this context, being able to remote drive a vehicle is key for enhancing autonomous driving systems. After deploying end-to-end teleoperated driving systems under 5G networks, the need to address complex challenges in other critical areas arises. These challenges belong to different technologies that need to be integrated in this particular system: video transmission and visualization technologies, artificial intelligence techniques, and network optimization features, incorporating haptic devices and critical data security. This article explores how these technologies can enhance the teleoperated driving activity experiences already executed in real-life environments by analyzing the quality of the video which is transmitted over the network, exploring its correlation with the current state-of-the-art AI object detection algorithms, analyzing the extended reality and digital twin paradigms, obtaining the maximum possible performance of forthcoming 6G networks and proposing decentralized security schema for ensuring the privacy and safety of the end-users of teleoperated driving infrastructures. An integrated set of conclusions and recommendations will be given to outline the future teleoperated driving systems design in the forthcoming years. Full article

Autonomous vehicles (AVs) must dynamically maintain sufficient safe distances from surrounding vehicles to ensure safety. Many existing studies have focused on collisions avoidance, such as the safety ranges in a rectangular shape that consider only longitudinal safe distance. A safety envelope is proposed herein, which is geometrically constructed from four quarter ellipses that account for longitudinal and lateral safe distances. The origin of the safety envelope is placed at the AV’s center of gravity. Using the safety envelope, a potential collision is identified when any surrounding vehicle enters it. To sustain the safety envelope even under hazardous situations, a collision avoidance strategy is introduced. In this strategy, the AV dynamically adjusts its velocity or changes lanes with velocity adjusting by assessing the risk level, complexity level, and riding comfort. For the lane-changing maneuvers, a virtual vehicle is introduced to be placed in the target lane to guide the AV’s movement. The efficacy of this strategy is verified via a simulation under a hazardous situation involving an AV and six human-driven vehicles driving on a highway. Results show that the proposed collision avoidance strategy utilizing safety envelope effectively ensures the safety of AV and surrounding vehicles, even under hazardous situations. Full article

Human–machine shared control has been widely adopted to enhance driving performance and facilitate smooth transitions between manual and fully autonomous driving. However, existing authority allocation strategies often neglect real-time assessment of scenario complexity and driver workload. To address this gap, we leverage non-invasive eye-tracking devices and the 3D virtual driving simulator Car Learning to Act (CARLA) to collect multimodal data—including physiological measures and vehicle dynamics—for the real-time classification of scenario complexity and cognitive workload. Feature importance is quantified using the SHAP (SHapley Additive exPlanations) values derived from Random Forest classifiers, enabling robust feature selection. Building upon a Hidden Markov Model (HMM) for workload inference and a Model Predictive Control (MPC) framework, we propose a novel human–machine shared control architecture with adaptive authority allocation. Human-in-the-loop validation experiments under both high- and low-workload conditions demonstrate that the proposed strategy significantly improves driving safety, stability, and overall performance. Notably, under high-workload scenarios, it achieves substantially greater reductions in Time to Collision (TTC) and Time to Lane Crossing (TLC) compared to low-workload conditions. Moreover, the adaptive approach yields lower controller load than alternative authority allocation methods, thereby minimizing human–machine conflict. Full article

Ensuring the safety of autonomous driving systems (ADS) requires scenario-based testing that reflects the complexity and variability of real-world driving conditions. However, the nondeterministic nature of actual traffic environments makes physical testing costly and limited in scope, particularly for rare and safety-critical scenarios. To address this, simulation has become a core component in validation by providing scalable, controllable, and repeatable testing environments. This study proposes a trajectory-based scenario classification framework that emphasizes both generality and interpretability. Specifically, we define a set of rule-based maneuver classification criteria using lateral acceleration patterns and apply them to simulated urban driving scenarios modeled with OpenSCENARIO. To address overlapping maneuver characteristics, a priority ordering of classification rules is introduced to resolve ambiguities. The proposed method was evaluated on a dataset comprising 7 types of maneuvers, including straight driving, lane changes, turns, roundabouts, and U-turns. Experimental results demonstrate the effectiveness of rule-driven classification based on vehicle trajectory dynamics and highlight the potential of this approach for structured scenario definition and validation in ADS simulation environments. Full article

Automated driving technologies have the capability to significantly increase road safety by decreasing accidents and increasing travel efficiency. This research presents a decision-making strategy for automated vehicles that models both lane changing and double lane changing maneuvers and is supported by a Deep Reinforcement Learning (DRL) algorithm. To capture realistic driving challenges, a highway driving scenario was designed using the professional multi-body simulation tool IPG Carmaker software, version 11 with realistic weather simulations to include aspects of rainy weather by incorporating vehicles with explicitly reduced tire–road friction while the ego vehicle is attempting to safely and perform efficient maneuvers in highway and merged merges. The hierarchical control system both creates an operational structure for planning and decision-making processes in highway maneuvers and articulates between higher-level driving decisions and lower-level autonomous motion control processes. As a result, a Duel Deep Deterministic Policy Gradient (Duel-DDPG) agent was created as the DRL approach to achieving decision-making in adverse driving conditions, which was built in MATLAB version 2021, designed, and tested. The study thoroughly explains both the Duel-DDPG and standard Deep Deterministic Policy Gradient (DDPG) algorithms, and we provide a direct performance comparative analysis. The discussion continues with simulation experiments of traffic complexity with uncertainty relating to weather conditions, which demonstrate the effectiveness of the Duel-DDPG algorithm. Full article

The growing deployment of connected and autonomous vehicles (CAVs) requires coordinated control strategies that jointly address safety, mobility, and energy efficiency. This paper presents a novel two-stage cooperative control framework for autonomous hybrid electric vehicle (HEV) platoons based on machine learning. In the first stage, a metric learning-based distributed model predictive control (ML-DMPC) strategy is proposed to enable cooperative longitudinal control among heterogeneous vehicles, explicitly incorporating inter-vehicle interactions to improve speed tracking, ride comfort, and platoon-level energy efficiency. In the second stage, a multi-agent twin-delayed deep deterministic policy gradient (MATD3) algorithm is developed for real-time energy management, achieving an optimal power split between the engine and battery while reducing Q-value overestimation and accelerating learning convergence. Simulation results across multiple standard driving cycles demonstrate that the proposed framework outperforms conventional distributed model predictive control (DMPC) and multi-agent deep deterministic policy gradient (MADDPG)-based methods in fuel economy, stability, and convergence speed, while maintaining battery state of charge (SOC) within safe limits. To facilitate future experimental validation, a dSPACE-based hardware-in-the-loop (HIL) architecture is designed to enable real-time deployment and testing of the proposed control framework. Full article

The high computational demand of dense voxel representations severely limits current vision-centric 3D semantic occupancy prediction methods, despite their capacity for granular scene understanding. This challenge is particularly acute in safety-critical applications like autonomous driving, where accurately perceiving dynamic instances often takes precedence over capturing the static background. This paper challenges the paradigm of dense prediction for such instance-focused tasks. We introduce the LSON-IP, a framework that strategically avoids the computational expense of dense 3D grids. LSON-IP operates on a sparse set of 3D instance queries, which are initialized directly from multi-view 2D images. These queries are then refined by our novel Sparse Instance Aggregator (SIA), an attention-based module. The SIA incorporates rich multi-view features while simultaneously modeling inter-query relationships to construct coherent object representations. Furthermore, to obviate the need for costly 3D annotations, we pioneer a Differentiable Sparse Rendering (DSR) technique. DSR innovatively defines a continuous field from the sparse voxel output, establishing a differentiable bridge between our sparse 3D representation and 2D supervision signals through volume rendering. Extensive experiments on major autonomous driving benchmarks, including SemanticKITTI and nuScenes, validate our approach. LSON-IP achieves strong performance on key dynamic instance categories and competitive overall semantic completion, all while reducing computational overhead by over 60% compared to dense baselines. Our work thus paves the way for efficient, high-fidelity instance-aware 3D perception. Full article

Aiming at the problems of large search space, unstable computational efficiency, and lack of safety of generated paths in complex environments of traditional HybridA* algorithms, this paper proposes an improved HybridA* algorithm based on Voronoi diagrams and safe corridors (GCHybridA*) to overcome these challenges. The method first reduces ineffective node expansion by constructing a Voronoi path away from obstacles and smoothing it, followed by selecting key guidance points to provide stage-like goals for path search. Then, an innovative safe corridor is generated and the path search is restricted to the safe corridor area to guarantee the safety of the path, and an adaptive step-size mechanism is designed to balance the search efficiency and path quality. The experimental results show that the GCHybridA* algorithm significantly outperforms the conventional HybridA* algorithm, with an average reduction of 83.7% in node expansions while maintaining zero potential collision points across all four typical maps. This study provides an innovative and robust solution for efficient and safe path planning in autonomous driving systems. This study provides an innovative and robust solution for global path planning in autonomous driving systems, focusing on static environment navigation with safety guarantees. Full article

Artificial intelligence (AI) technology, serving as an indispensable component within intelligent construction systems, has become a cornerstone for driving the digital and intelligent transformation of the construction industry. Although AI demonstrates autonomous decision-making capabilities in specific operational contexts, because of the dynamic and often unforeseeable nature of construction workflows, human–AI collaboration (HAIC) still dominates the operational paradigm. This study undertakes a systematic review of the prior research on human–AI collaboration in intelligent construction. Through a bibliometric search, scientometric analysis, and in-depth literature classification, 191 highly cited articles in the past five years, which are in the top 10% by citation count within the dataset (as of May 2025, based on Scopus, Google Scholar, and WOS), were screened, and four research streams were formed based on a co-citation analysis and clustering, namely, construction robotics, productivity and safety, intelligent algorithms and modelling, and factors related to construction workers. Finally, a three-dimensional knowledge framework covering the technical layer, application layer, and management layer was constructed. Through this comprehensive synthesis, the study developed a human–AI collaboration knowledge framework in the field of construction science that integrates technology, scenarios, and management dimensions, revealing the co-evolutionary path of artificial intelligence technology and industry digital transformation. Full article