Sikai Chen

Yansong Qu, Zihao Sheng, Zilin Huang et al.

Reinforcement Learning (RL) has emerged as a dominant paradigm for end-to-end autonomous driving (AD). However, RL suffers from sample inefficiency and a lack of semantic interpretability in complex scenarios. Foundation Models, particularly Vision-Language Models (VLMs), can mitigate this by offering rich, context-aware knowledge, yet their high inference latency hinders deployment in high-frequency RL training loops. To bridge this gap, we present Found-RL, a platform tailored to efficiently enhance RL for AD using foundation models. A core innovation is the asynchronous batch inference framework, which decouples heavy VLM reasoning from the simulation loop, effectively resolving latency bottlenecks to support real-time learning. We introduce diverse supervision mechanisms: Value-Margin Regularization (VMR) and Advantage-Weighted Action Guidance (AWAG) to effectively distill expert-like VLM action suggestions into the RL policy. Additionally, we adopt high-throughput CLIP for dense reward shaping. We address CLIP's dynamic blindness via Conditional Contrastive Action Alignment, which conditions prompts on discretized speed/command and yields a normalized, margin-based bonus from context-specific action-anchor scoring. Found-RL provides an end-to-end pipeline for fine-tuned VLM integration and shows that a lightweight RL model can achieve near-VLM performance compared with billion-parameter VLMs while sustaining real-time inference (approx. 500 FPS). Code, data, and models will be publicly available at https://github.com/ys-qu/found-rl.

Junwei You, Pei Li, Zhuoyu Jiang et al.

Multimodal large language models (MLLMs) have shown strong potential for autonomous driving, yet existing benchmarks remain largely ego-centric and therefore cannot systematically assess model performance in infrastructure-centric and cooperative driving conditions. In this work, we introduce V2X-QA, a real-world dataset and benchmark for evaluating MLLMs across vehicle-side, infrastructure-side, and cooperative viewpoints. V2X-QA is built around a view-decoupled evaluation protocol that enables controlled comparison under vehicle-only, infrastructure-only, and cooperative driving conditions within a unified multiple-choice question answering (MCQA) framework. The benchmark is organized into a twelve-task taxonomy spanning perception, prediction, and reasoning and planning, and is constructed through expert-verified MCQA annotation to enable fine-grained diagnosis of viewpoint-dependent capabilities. Benchmark results across ten representative state-of-the-art proprietary and open-source models show that viewpoint accessibility substantially affects performance, and infrastructure-side reasoning supports meaningful macroscopic traffic understanding. Results also indicate that cooperative reasoning remains challenging since it requires cross-view alignment and evidence integration rather than simply additional visual input. To address these challenges, we introduce V2X-MoE, a benchmark-aligned baseline with explicit view routing and viewpoint-specific LoRA experts. The strong performance of V2X-MoE further suggests that explicit viewpoint specialization is a promising direction for multi-view reasoning in autonomous driving. Overall, V2X-QA provides a foundation for studying multi-perspective reasoning, reliability, and cooperative physical intelligence in connected autonomous driving. The dataset and V2X-MoE resources are publicly available at: https://github.com/junwei0001/V2X-QA.

Jiqian Dong, Sikai Chen, Samuel Labi

With ongoing development of autonomous driving systems and increasing desire for deployment, researchers continue to seek reliable approaches for ADS systems. The virtual simulation test (VST) has become a prominent approach for testing autonomous driving systems (ADS) and advanced driver assistance systems (ADAS) due to its advantages of fast execution, low cost, and high repeatability. However, the success of these simulation-based experiments heavily relies on the realism of the testing scenarios. It is needed to create more flexible and high-fidelity testing scenarios in VST in order to increase the safety and reliabilityof ADS and ADAS.To address this challenge, this paper introduces the "Transfusor" model, which leverages the transformer and diffusor models (two cutting-edge deep learning generative technologies). The primary objective of the Transfusor model is to generate highly realistic and controllable human-like lane-changing trajectories in highway scenarios. Extensive experiments were carried out, and the results demonstrate that the proposed model effectively learns the spatiotemporal characteristics of humans' lane-changing behaviors and successfully generates trajectories that closely mimic real-world human driving. As such, the proposed model can play a critical role of creating more flexible and high-fidelity testing scenarios in the VST, ultimately leading to safer and more reliable ADS and ADAS.

Tianyi Wang, Tianyi Zeng, Zimo Zeng et al.

Advanced driver assistance systems (ADAS) play an important role in modern automotive intelligence, significantly enhancing vehicle safety and stability. The performance of ADAS critically relies on accurate and reliable vehicle state estimation, particularly from vehicle dynamic sensors. Among these signals, wheel load is a key variable for chassis control and safety-critical functions, yet it remains difficult to estimate robustly due to complex suspension geometry, nonlinear dynamics, and measurement noise. To address this issue, we propose DBPnet, a Bayesian physics-informed neural network (PINN) with a physics-aware embedding module inspired by damper characteristics. First, this paper presents a suspension linkage-level modeling (SLLM) approach that constructs a nonlinear instantaneous dynamic model by explicitly considering the complex geometric structure of the suspension. Building upon SLLM, Bayesian inference is integrated into the PINN to effectively cope with noise and uncertainty in the vehicle chassis system, thereby improving the model's robustness. Then, a physics-informed loss function is employed to ensure consistency with fundamental physical principles, while the damper characteristics-inspired embedding module extracts temporal variation features of input signals and incorporates them into each layer of the PINN, ensuring that physical observations guide the neural network without being constrained by fixed physical models. Extensive evaluations on high-fidelity simulations and real-world experiments demonstrate that our DBPnet consistently achieves lower RMSE and MaxError than baseline methods. These results highlight the potential of our DBPnet to advance wheel load estimation and contribute to the development of more reliable ADAS actuator functions.

Zihao Sheng, Zilin Huang, Sikai Chen

To drive safely in complex traffic environments, autonomous vehicles need to make an accurate prediction of the future trajectories of nearby heterogeneous traffic agents (i.e., vehicles, pedestrians, bicyclists, etc). Due to the interactive nature, human drivers are accustomed to infer what the future situations will become if they are going to execute different maneuvers. To fully exploit the impacts of interactions, this paper proposes a ego-planning guided multi-graph convolutional network (EPG-MGCN) to predict the trajectories of heterogeneous agents using both historical trajectory information and ego vehicle's future planning information. The EPG-MGCN first models the social interactions by employing four graph topologies, i.e., distance graphs, visibility graphs, planning graphs and category graphs. Then, the planning information of the ego vehicle is encoded by both the planning graph and the subsequent planning-guided prediction module to reduce uncertainty in the trajectory prediction. Finally, a category-specific gated recurrent unit (CS-GRU) encoder-decoder is designed to generate future trajectories for each specific type of agents. Our network is evaluated on two real-world trajectory datasets: ApolloScape and NGSIM. The experimental results show that the proposed EPG-MGCN achieves state-of-the-art performance compared to existing methods.

Zihao Sheng, Zilin Huang, Sikai Chen

Model-based reinforcement learning (RL) is anticipated to exhibit higher sample efficiency compared to model-free RL by utilizing a virtual environment model. However, it is challenging to obtain sufficiently accurate representations of the environmental dynamics due to uncertainties in complex systems and environments. An inaccurate environment model may degrade the sample efficiency and performance of model-based RL. Furthermore, while model-based RL can improve sample efficiency, it often still requires substantial training time to learn from scratch, potentially limiting its advantages over model-free approaches. To address these challenges, this paper introduces a knowledge-informed model-based residual reinforcement learning framework aimed at enhancing learning efficiency by infusing established expert knowledge into the learning process and avoiding the issue of beginning from zero. Our approach integrates traffic expert knowledge into a virtual environment model, employing the Intelligent Driver Model (IDM) for basic dynamics and neural networks for residual dynamics, thus ensuring adaptability to complex scenarios. We propose a novel strategy that combines traditional control methods with residual RL, facilitating efficient learning and policy optimization without the need to learn from scratch. The proposed approach is applied to CAV trajectory control tasks for the dissipation of stop-and-go waves in mixed traffic flow. Experimental results demonstrate that our proposed approach enables the CAV agent to achieve superior performance in trajectory control compared to the baseline agents in terms of sample efficiency, traffic flow smoothness and traffic mobility. The source code and supplementary materials are available at: https://zihaosheng.github.io/traffic-expertise-RL/.

Jiaming Wang, Jiqian Dong, Sikai Chen et al.

The exponential growth of electric vehicles (EVs) presents novel challenges in preserving battery health and in addressing the persistent problem of vehicle range anxiety. To address these concerns, wireless charging, particularly, Mobile Energy Disseminators (MEDs) have emerged as a promising solution. The MED is mounted behind a large vehicle and charges all participating EVs within a radius upstream of it. Unfortuantely, during such V2V charging, the MED and EVs inadvertently form platoons, thereby occupying multiple lanes and impairing overall corridor travel efficiency. In addition, constrained budgets for MED deployment necessitate the development of an effective dispatching strategy to determine optimal timing and locations for introducing the MEDs into traffic. This paper proposes a deep reinforcement learning (DRL) based methodology to develop a vehicle dispatching framework. In the first component of the framework, we develop a realistic reinforcement learning environment termed "ChargingEnv" which incorporates a reliable charging simulation system that accounts for common practical issues in wireless charging deployment, specifically, the charging panel misalignment. The second component, the Proximal-Policy Optimization (PPO) agent, is trained to control MED dispatching through continuous interactions with ChargingEnv. Numerical experiments were carried out to demonstrate the demonstrate the efficacy of the proposed MED deployment decision processor. The experiment results suggest that the proposed model can significantly enhance EV travel range while efficiently deploying a optimal number of MEDs. The proposed model is found to be not only practical in its applicability but also has promises of real-world effectiveness. The proposed model can help travelers to maximize EV range and help road agencies or private-sector vendors to manage the deployment of MEDs efficiently.

Zilin Huang, Zihao Sheng, Sikai Chen

In the field of autonomous driving, developing safe and trustworthy autonomous driving policies remains a significant challenge. Recently, Reinforcement Learning with Human Feedback (RLHF) has attracted substantial attention due to its potential to enhance training safety and sampling efficiency. Nevertheless, existing RLHF-enabled methods often falter when faced with imperfect human demonstrations, potentially leading to training oscillations or even worse performance than rule-based approaches. Inspired by the human learning process, we propose Physics-enhanced Reinforcement Learning with Human Feedback (PE-RLHF). This novel framework synergistically integrates human feedback (e.g., human intervention and demonstration) and physics knowledge (e.g., traffic flow model) into the training loop of reinforcement learning. The key advantage of PE-RLHF is its guarantee that the learned policy will perform at least as well as the given physics-based policy, even when human feedback quality deteriorates, thus ensuring trustworthy safety improvements. PE-RLHF introduces a Physics-enhanced Human-AI (PE-HAI) collaborative paradigm for dynamic action selection between human and physics-based actions, employs a reward-free approach with a proxy value function to capture human preferences, and incorporates a minimal intervention mechanism to reduce the cognitive load on human mentors. Extensive experiments across diverse driving scenarios demonstrate that PE-RLHF significantly outperforms traditional methods, achieving state-of-the-art (SOTA) performance in safety, efficiency, and generalizability, even with varying quality of human feedback. The philosophy behind PE-RLHF not only advances autonomous driving technology but can also offer valuable insights for other safety-critical domains. Demo video and code are available at: \https://zilin-huang.github.io/PE-RLHF-website/

Keke Long, Zihao Sheng, Haotian Shi et al.

In vehicle trajectory prediction, physics models and data-driven models are two predominant methodologies. However, each approach presents its own set of challenges: physics models fall short in predictability, while data-driven models lack interpretability. Addressing these identified shortcomings, this paper proposes a novel framework, the Physics-Enhanced Residual Learning (PERL) model. PERL integrates the strengths of physics-based and data-driven methods for traffic state prediction. PERL contains a physics model and a residual learning model. Its prediction is the sum of the physics model result and a predicted residual as a correction to it. It preserves the interpretability inherent to physics-based models and has reduced data requirements compared to data-driven methods. Experiments were conducted using a real-world vehicle trajectory dataset. We proposed a PERL model, with the Intelligent Driver Model (IDM) as its physics car-following model and Long Short-Term Memory (LSTM) as its residual learning model. We compare this PERL model with the physics car-following model, data-driven model, and other physics-informed neural network (PINN) models. The result reveals that PERL achieves better prediction with a small dataset, compared to the physics model, data-driven model, and PINN model. Second, the PERL model showed faster convergence during training, offering comparable performance with fewer training samples than the data-driven model and PINN model. Sensitivity analysis also proves comparable performance of PERL using another residual learning model and a physics car-following model.

Zilin Huang, Zihao Sheng, Zhengyang Wan et al.

Ensuring safe decision-making in autonomous vehicles remains a fundamental challenge despite rapid advances in end-to-end learning approaches. Traditional reinforcement learning (RL) methods rely on manually engineered rewards or sparse collision signals, which fail to capture the rich contextual understanding required for safe driving and make unsafe exploration unavoidable in real-world settings. Recent vision-language models (VLMs) offer promising semantic understanding capabilities; however, their high inference latency and susceptibility to hallucination hinder direct application to real-time vehicle control. To address these limitations, this paper proposes DriveVLM-RL, a neuroscience-inspired framework that integrates VLMs into RL through a dual-pathway architecture for safe and deployable autonomous driving. The framework decomposes semantic reward learning into a Static Pathway for continuous spatial safety assessment using CLIP-based contrasting language goals, and a Dynamic Pathway for attention-gated multi-frame semantic risk reasoning using a lightweight detector and a large VLM. A hierarchical reward synthesis mechanism fuses semantic signals with vehicle states, while an asynchronous training pipeline decouples expensive VLM inference from environment interaction. All VLM components are used only during offline training and are removed at deployment, ensuring real-time feasibility. Experiments in the CARLA simulator show significant improvements in collision avoidance, task success, and generalization across diverse traffic scenarios, including strong robustness under settings without explicit collision penalties. These results demonstrate that DriveVLM-RL provides a practical paradigm for integrating foundation models into autonomous driving without compromising real-time feasibility. Demo video and code are available at: https://zilin-huang.github.io/DriveVLM-RL-website/

Zihao Sheng, Xin Ye, Jingru Luo et al.

End-to-end autonomous driving models based on Vision-Language-Action (VLA) architectures have shown promising results by learning driving policies through behavior cloning on expert demonstrations. However, imitation learning inherently limits the model to replicating observed behaviors without exploring diverse driving strategies, leaving it brittle in novel or out-of-distribution scenarios. Reinforcement learning (RL) offers a natural remedy by enabling policy exploration beyond the expert distribution. Yet VLA models, typically trained on offline datasets, lack directly observable state transitions, necessitating a learned world model to anticipate action consequences. In this work, we propose a unified understanding-and-generation framework that leverages world modeling to simultaneously enable meaningful exploration and provide dense supervision. Specifically, we augment trajectory prediction with future RGB and depth image generation as dense world modeling objectives, requiring the model to learn fine-grained visual and geometric representations that substantially enrich the planning backbone. Beyond serving as a supervisory signal, the world model further acts as a source of intrinsic reward for policy exploration: its image prediction uncertainty naturally measures a trajectory's novelty relative to the training distribution, where high uncertainty indicates out-of-distribution scenarios that, if safe, represent valuable learning opportunities. We incorporate this exploration signal into a safety-gated reward and optimize the policy via Group Relative Policy Optimization (GRPO). Experiments on the NAVSIM and nuScenes benchmarks demonstrate the effectiveness of our approach, achieving a state-of-the-art PDMS score of 93.7 and an EPDMS of 88.8 on NAVSIM. The code and demo will be publicly available at https://zihaosheng.github.io/ExploreVLA/.

Xiangyu Li, Tianyi Wang, Xi Cheng et al.

Accurate classification of autonomous vehicle (AV) driving behaviors is critical for safety validation, performance diagnosis, and traffic integration analysis. However, existing approaches primarily rely on numerical time-series modeling and often lack semantic abstraction, limiting interpretability and robustness in complex traffic environments. This paper presents LLM-MLFFN, a novel large language model (LLM)-enhanced multi-level feature fusion network designed to address the complexities of multi-dimensional driving data. The proposed LLM-MLFFN framework integrates priors from largescale pre-trained models and employs a multi-level approach to enhance classification accuracy. LLM-MLFFN comprises three core components: (1) a multi-level feature extraction module that extracts statistical, behavioral, and dynamic features to capture the quantitative aspects of driving behaviors; (2) a semantic description module that leverages LLMs to transform raw data into high-level semantic features; and (3) a dual-channel multi-level feature fusion network that combines numerical and semantic features using weighted attention mechanisms to improve robustness and prediction accuracy. Evaluation on the Waymo open trajectory dataset demonstrates the superior performance of the proposed LLM-MLFFN, achieving a classification accuracy of over 94%, surpassing existing machine learning models. Ablation studies further validate the critical contributions of multi-level fusion, feature extraction strategies, and LLM-derived semantic reasoning. These results suggest that integrating structured feature modeling with language-driven semantic abstraction provides a principled and interpretable pathway for robust autonomous driving behavior classification.

Sicong Jiang, Zilin Huang, Kangan Qian et al.

The rapid progress of multimodal large language models (MLLM) has paved the way for Vision-Language-Action (VLA) paradigms, which integrate visual perception, natural language understanding, and control within a single policy. Researchers in autonomous driving are actively adapting these methods to the vehicle domain. Such models promise autonomous vehicles that can interpret high-level instructions, reason about complex traffic scenes, and make their own decisions. However, the literature remains fragmented and is rapidly expanding. This survey offers the first comprehensive overview of VLA for Autonomous Driving (VLA4AD). We (i) formalize the architectural building blocks shared across recent work, (ii) trace the evolution from early explainer to reasoning-centric VLA models, and (iii) compare over 20 representative models according to VLA's progress in the autonomous driving domain. We also consolidate existing datasets and benchmarks, highlighting protocols that jointly measure driving safety, accuracy, and explanation quality. Finally, we detail open challenges - robustness, real-time efficiency, and formal verification - and outline future directions of VLA4AD. This survey provides a concise yet complete reference for advancing interpretable socially aligned autonomous vehicles. Github repo is available at \href{https://github.com/JohnsonJiang1996/Awesome-VLA4AD}{SicongJiang/Awesome-VLA4AD}.

Tianyi Zeng, Tianyi Wang, Jiaru Zhang et al.

Diffusion models have emerged as powerful generative tools for modeling complex data distributions, yet their purely data-driven nature limits applicability in practical engineering and scientific problems where physical laws need to be followed. This paper proposes Physics-Informed Learning via Diffusion (PILD), a framework that unifies diffusion modeling and first-principles physical constraints by introducing a virtual residual observation sampled from a Laplace distribution to supervise generation during training. To further integrate physical laws, a conditional embedding module is incorporated to inject physical information into the denoising network at multiple layers, ensuring consistent guidance throughout the diffusion process. The proposed PILD framework is concise, modular, and broadly applicable to problems governed by ordinary differential equations, partial differential equations, as well as algebraic equations or inequality constraints. Extensive experiments across engineering and scientific tasks including estimating vehicle trajectories, tire forces, Darcy flow and plasma dynamics, demonstrate that our PILD substantially improves accuracy, stability, and generalization over existing physics-informed and diffusion-based baselines.

Zilin Huang, Zhengyang Wan, Zihao Sheng et al.

Deploying reinforcement learning policies trained in simulation to real autonomous vehicles remains a fundamental challenge, particularly for VLM-guided RL frameworks whose policies are typically learned with simulator-native observations and simulator-coupled action semantics that are unavailable on physical platforms. This paper presents Sim2Real-AD, a modular framework for zero-shot sim-to-real transfer of CARLA-trained VLM-guided RL policies to full-scale vehicles without any real-world RL training data. The framework decomposes the transfer problem into four components: a Geometric Observation Bridge (GOB) that converts monocular front-view images into simulator-compatible bird's-eye-view (BEV) observations, a Physics-Aware Action Mapping (PAM) that translates policy outputs into platform-agnostic physical commands, a Two-Phase Progressive Training (TPT) strategy that stabilizes adaptation by separating action-space and observation-space transfer, and a Real-time Deployment Pipeline (RDP) that integrates perception, policy inference, control conversion, and safety monitoring for closed-loop execution. Simulation experiments show that the framework preserves the relative performance ordering of representative RL algorithms across different reward paradigms and validate the contribution of each module. Zero-shot deployment on a full-scale Ford E-Transit achieves success rates of 90%, 80%, and 75% in car-following, obstacle avoidance, and stop-sign interaction scenarios, respectively. To the best of our knowledge, this study is among the first to demonstrate zero-shot closed-loop deployment of a CARLA-trained VLM-guided RL policy on a full-scale real vehicle without any real-world RL training data. The demo video and code are available at: https://zilin-huang.github.io/Sim2Real-AD-website/.

Zilin Huang, Zihao Sheng, Chengyuan Ma et al.

Despite significant progress in autonomous vehicles (AVs), the development of driving policies that ensure both the safety of AVs and traffic flow efficiency has not yet been fully explored. In this paper, we propose an enhanced human-in-the-loop reinforcement learning method, termed the Human as AI mentor-based deep reinforcement learning (HAIM-DRL) framework, which facilitates safe and efficient autonomous driving in mixed traffic platoon. Drawing inspiration from the human learning process, we first introduce an innovative learning paradigm that effectively injects human intelligence into AI, termed Human as AI mentor (HAIM). In this paradigm, the human expert serves as a mentor to the AI agent. While allowing the agent to sufficiently explore uncertain environments, the human expert can take control in dangerous situations and demonstrate correct actions to avoid potential accidents. On the other hand, the agent could be guided to minimize traffic flow disturbance, thereby optimizing traffic flow efficiency. In detail, HAIM-DRL leverages data collected from free exploration and partial human demonstrations as its two training sources. Remarkably, we circumvent the intricate process of manually designing reward functions; instead, we directly derive proxy state-action values from partial human demonstrations to guide the agents' policy learning. Additionally, we employ a minimal intervention technique to reduce the human mentor's cognitive load. Comparative results show that HAIM-DRL outperforms traditional methods in driving safety, sampling efficiency, mitigation of traffic flow disturbance, and generalizability to unseen traffic scenarios. The code and demo videos for this paper can be accessed at: https://zilin-huang.github.io/HAIM-DRL-website/

Zilin Huang, Zihao Sheng, Yansong Qu et al.

In recent years, reinforcement learning (RL)-based methods for learning driving policies have gained increasing attention in the autonomous driving community and have achieved remarkable progress in various driving scenarios. However, traditional RL approaches rely on manually engineered rewards, which require extensive human effort and often lack generalizability. To address these limitations, we propose \textbf{VLM-RL}, a unified framework that integrates pre-trained Vision-Language Models (VLMs) with RL to generate reward signals using image observation and natural language goals. The core of VLM-RL is the contrasting language goal (CLG)-as-reward paradigm, which uses positive and negative language goals to generate semantic rewards. We further introduce a hierarchical reward synthesis approach that combines CLG-based semantic rewards with vehicle state information, improving reward stability and offering a more comprehensive reward signal. Additionally, a batch-processing technique is employed to optimize computational efficiency during training. Extensive experiments in the CARLA simulator demonstrate that VLM-RL outperforms state-of-the-art baselines, achieving a 10.5\% reduction in collision rate, a 104.6\% increase in route completion rate, and robust generalization to unseen driving scenarios. Furthermore, VLM-RL can seamlessly integrate almost any standard RL algorithms, potentially revolutionizing the existing RL paradigm that relies on manual reward engineering and enabling continuous performance improvements. The demo video and code can be accessed at: https://zilin-huang.github.io/VLM-RL-website.

Zihao Sheng, Zilin Huang, Yansong Qu et al.

Ensuring safety in autonomous driving systems remains a critical challenge, particularly in handling rare but potentially catastrophic safety-critical scenarios. While existing research has explored generating safety-critical scenarios for autonomous vehicle (AV) testing, there is limited work on effectively incorporating these scenarios into policy learning to enhance safety. Furthermore, developing training curricula that adapt to an AV's evolving behavioral patterns and performance bottlenecks remains largely unexplored. To address these challenges, we propose CurricuVLM, a novel framework that leverages Vision-Language Models (VLMs) to enable personalized curriculum learning for autonomous driving agents. Our approach uniquely exploits VLMs' multimodal understanding capabilities to analyze agent behavior, identify performance weaknesses, and dynamically generate tailored training scenarios for curriculum adaptation. Through comprehensive analysis of unsafe driving situations with narrative descriptions, CurricuVLM performs in-depth reasoning to evaluate the AV's capabilities and identify critical behavioral patterns. The framework then synthesizes customized training scenarios targeting these identified limitations, enabling effective and personalized curriculum learning. Extensive experiments on the Waymo Open Motion Dataset show that CurricuVLM outperforms state-of-the-art baselines across both regular and safety-critical scenarios, achieving superior performance in terms of navigation success, driving efficiency, and safety metrics. Further analysis reveals that CurricuVLM serves as a general approach that can be integrated with various RL algorithms to enhance autonomous driving systems. The code and demo video are available at: https://zihaosheng.github.io/CurricuVLM/.

Rui Gan, Junyi Ma, Pei Li et al.

Cooperative autonomous driving requires traffic scene understanding from both vehicle and infrastructure perspectives. While vision-language models (VLMs) show strong general reasoning capabilities, their performance in safety-critical traffic scenarios remains insufficiently evaluated due to the ego-vehicle focus of existing benchmarks. To bridge this gap, we present \textbf{CrashSight}, a large-scale vision-language benchmark for roadway crash understanding using real-world roadside camera data. The dataset comprises 250 crash videos, annotated with 13K multiple-choice question-answer pairs organized under a two-tier taxonomy. Tier 1 evaluates the visual grounding of scene context and involved parties, while Tier 2 probes higher-level reasoning, including crash mechanics, causal attribution, temporal progression, and post-crash outcomes. We benchmark 8 state-of-the-art VLMs and show that, despite strong scene description capabilities, current models struggle with temporal and causal reasoning in safety-critical scenarios. We provide a detailed analysis of failure scenarios and discuss directions for improving VLM crash understanding. The benchmark provides a standardized evaluation framework for infrastructure-assisted perception in cooperative autonomous driving. The CrashSight benchmark, including the full dataset and code, is accessible at https://mcgrche.github.io/crashsight.

Yansong Qu, Zilin Huang, Zihao Sheng et al.

Reinforcement learning (RL)-based autonomous driving policy learning faces critical limitations such as low sample efficiency and poor generalization; its reliance on online interactions and trial-and-error learning is especially unacceptable in safety-critical scenarios. Existing methods including safe RL often fail to capture the true semantic meaning of "safety" in complex driving contexts, leading to either overly conservative driving behavior or constraint violations. To address these challenges, we propose VL-SAFE, a world model-based safe RL framework with Vision-Language model (VLM)-as-safety-guidance paradigm, designed for offline safe policy learning. Specifically, we construct offline datasets containing data collected by expert agents and labeled with safety scores derived from VLMs. A world model is trained to generate imagined rollouts together with safety estimations, allowing the agent to perform safe planning without interacting with the real environment. Based on these imagined trajectories and safety evaluations, actor-critic learning is conducted under VLM-based safety guidance to optimize the driving policy more safely and efficiently. Extensive evaluations demonstrate that VL-SAFE achieves superior sample efficiency, generalization, safety, and overall performance compared to existing baselines. To the best of our knowledge, this is the first work that introduces a VLM-guided world model-based approach for safe autonomous driving. The demo video and code can be accessed at: https://ys-qu.github.io/vlsafe-website/

Yansong Qu, Zixuan Xu, Zilin Huang et al.

Semantic scene completion (SSC) is essential for achieving comprehensive perception in autonomous driving systems. However, existing SSC methods often overlook the high deployment costs in real-world applications. Traditional architectures, such as 3D Convolutional Neural Networks (3D CNNs) and self-attention mechanisms, face challenges in efficiently capturing long-range dependencies within 3D voxel grids, limiting their effectiveness. To address these issues, we introduce MetaSSC, a novel meta-learning-based framework for SSC that leverages deformable convolution, large-kernel attention, and the Mamba (D-LKA-M) model. Our approach begins with a voxel-based semantic segmentation (SS) pretraining task, aimed at exploring the semantics and geometry of incomplete regions while acquiring transferable meta-knowledge. Using simulated cooperative perception datasets, we supervise the perception training of a single vehicle using aggregated sensor data from multiple nearby connected autonomous vehicles (CAVs), generating richer and more comprehensive labels. This meta-knowledge is then adapted to the target domain through a dual-phase training strategy that does not add extra model parameters, enabling efficient deployment. To further enhance the model's capability in capturing long-sequence relationships within 3D voxel grids, we integrate Mamba blocks with deformable convolution and large-kernel attention into the backbone network. Extensive experiments demonstrate that MetaSSC achieves state-of-the-art performance, significantly outperforming competing models while also reducing deployment costs.

Zihao Sheng, Zilin Huang, Yansong Qu et al.

Multimodal large language models (MLLMs) have achieved remarkable progress across a range of vision-language tasks and demonstrate strong potential for traffic accident understanding. However, existing MLLMs in this domain primarily focus on coarse-grained image-level or video-level comprehension and often struggle to handle fine-grained visual details or localized scene components, limiting their applicability in complex accident scenarios. To address these limitations, we propose SafePLUG, a novel framework that empowers MLLMs with both Pixel-Level Understanding and temporal Grounding for comprehensive traffic accident analysis. SafePLUG supports both arbitrary-shaped visual prompts for region-aware question answering and pixel-level segmentation based on language instructions, while also enabling the recognition of temporally anchored events in traffic accident scenarios. To advance the development of MLLMs for traffic accident understanding, we curate a new dataset containing multimodal question-answer pairs centered on diverse accident scenarios, with detailed pixel-level annotations and temporal event boundaries. Experimental results show that SafePLUG achieves strong performance on multiple tasks, including region-based question answering, pixel-level segmentation, temporal event localization, and accident event understanding. These capabilities lay a foundation for fine-grained understanding of complex traffic scenes, with the potential to improve driving safety and enhance situational awareness in smart transportation systems. The code, dataset, and model checkpoints will be made publicly available at: https://zihaosheng.github.io/SafePLUG

Zilin Huang, Zihao Sheng, Zhengyang Wan et al.

Recent advances in autonomous system simulation platforms have significantly enhanced the safe and scalable testing of driving policies. However, existing simulators do not yet fully meet the needs of future transportation research-particularly in enabling effective human-AI collaboration and modeling socially-aware driving agents. This paper introduces Sky-Drive, a novel distributed multi-agent simulation platform that addresses these limitations through four key innovations: (a) a distributed architecture for synchronized simulation across multiple terminals; (b) a multi-modal human-in-the-loop framework integrating diverse sensors to collect rich behavioral data; (c) a human-AI collaboration mechanism supporting continuous and adaptive knowledge exchange; and (d) a digital twin framework for constructing high-fidelity virtual replicas of real-world transportation environments. Sky-Drive supports diverse applications such as autonomous vehicle-human road users interaction modeling, human-in-the-loop training, socially-aware reinforcement learning, personalized driving development, and customized scenario generation. Future extensions will incorporate foundation models for context-aware decision support and hardware-in-the-loop testing for real-world validation. By bridging scenario generation, data collection, algorithm training, and hardware integration, Sky-Drive has the potential to become a foundational platform for the next generation of human-centered and socially-aware autonomous transportation systems research. The demo video and code are available at:https://sky-lab-uw.github.io/Sky-Drive-website/

Junwei You, Rui Gan, Weizhe Tang et al.

Vehicle trajectory prediction is crucial for advancing autonomous driving and advanced driver assistance systems (ADAS). Although deep learning-based approaches - especially those utilizing transformer-based and generative models - have markedly improved prediction accuracy by capturing complex, non-linear patterns in vehicle dynamics and traffic interactions, they frequently overlook detailed car-following behaviors and the inter-vehicle interactions critical for real-world driving applications, particularly in fully autonomous or mixed traffic scenarios. To address the issue, this study introduces a scaled noise conditional diffusion model for car-following trajectory prediction, which integrates detailed inter-vehicular interactions and car-following dynamics into a generative framework, improving both the accuracy and plausibility of predicted trajectories. The model utilizes a novel pipeline to capture historical vehicle dynamics by scaling noise with encoded historical features within the diffusion process. Particularly, it employs a cross-attention-based transformer architecture to model intricate inter-vehicle dependencies, effectively guiding the denoising process and enhancing prediction accuracy. Experimental results on diverse real-world driving scenarios demonstrate the state-of-the-art performance and robustness of the proposed method.

Yuhao Luo, Yuang Zhang, Kehua Chen et al.

Accurate pedestrian trajectory prediction is crucial for ensuring safety and efficiency in autonomous driving and human-robot interaction scenarios. Earlier studies primarily utilized sufficient observational data to predict future trajectories. However, in real-world scenarios, such as pedestrians suddenly emerging from blind spots, sufficient observational data is often unavailable (i.e. momentary trajectory), making accurate prediction challenging and increasing the risk of traffic accidents. Therefore, advancing research on pedestrian trajectory prediction under extreme scenarios is critical for enhancing traffic safety. In this work, we propose a novel framework termed Diffusion^2, tailored for momentary trajectory prediction. Diffusion^2 consists of two sequentially connected diffusion models: one for backward prediction, which generates unobserved historical trajectories, and the other for forward prediction, which forecasts future trajectories. Given that the generated unobserved historical trajectories may introduce additional noise, we propose a dual-head parameterization mechanism to estimate their aleatoric uncertainty and design a temporally adaptive noise module that dynamically modulates the noise scale in the forward diffusion process. Empirically, Diffusion^2 sets a new state-of-the-art in momentary trajectory prediction on ETH/UCY and Stanford Drone datasets.

Junwei You, Haotian Shi, Keshu Wu et al.

Vehicle trajectory prediction is crucial for advancing autonomous driving and advanced driver assistance systems (ADAS), enhancing road safety and traffic efficiency. While traditional methods have laid foundational work, modern deep learning techniques, particularly transformer-based models and generative approaches, have significantly improved prediction accuracy by capturing complex and non-linear patterns in vehicle motion and traffic interactions. However, these models often overlook the detailed car-following behaviors and inter-vehicle interactions essential for real-world driving scenarios. This study introduces a Cross-Attention Transformer Enhanced Conditional Diffusion Model (Crossfusor) specifically designed for car-following trajectory prediction. Crossfusor integrates detailed inter-vehicular interactions and car-following dynamics into a robust diffusion framework, improving both the accuracy and realism of predicted trajectories. The model leverages a novel temporal feature encoding framework combining GRU, location-based attention mechanisms, and Fourier embedding to capture historical vehicle dynamics. It employs noise scaled by these encoded historical features in the forward diffusion process, and uses a cross-attention transformer to model intricate inter-vehicle dependencies in the reverse denoising process. Experimental results on the NGSIM dataset demonstrate that Crossfusor outperforms state-of-the-art models, particularly in long-term predictions, showcasing its potential for enhancing the predictive capabilities of autonomous driving systems.

Shuya Zong, Sikai Chen, Majed Alinizzi et al.

Assessing collision risk is a critical challenge to effective traffic safety management. The deployment of unmanned aerial vehicles (UAVs) to address this issue has shown much promise, given their wide visual field and movement flexibility. This research demonstrates the application of UAVs and V2X connectivity to track the movement of road users and assess potential collisions at intersections. The study uses videos captured by UAVs. The proposed method combines deep-learning based tracking algorithms and time-to-collision tasks. The results not only provide beneficial information for vehicle's recognition of potential crashes and motion planning but also provided a valuable tool for urban road agencies and safety management engineers.

Shuya Zong, Sikai Chen, Samuel Labi

Video monitoring of traffic is useful for traffic management and control, traffic counting, and traffic law enforcement. However, traffic monitoring during inclement weather such as rain is a challenging task because video quality is corrupted by streaks of falling rain on the video image, and this hinders reliable characterization not only of the road environment but also of road-user behavior during such adverse weather events. This study proposes a two-stage self-supervised learning method to remove rain streaks in traffic videos. The first and second stages address intra- and inter-frame noise, respectively. The results indicated that the model exhibits satisfactory performance in terms of the image visual quality and the Peak Signal-Noise Ratio value.

Paul, Ha, Sikai Chen et al.

Optimizing traffic signal control (TSC) at intersections continues to pose a challenging problem, particularly for large-scale traffic networks. It has been shown in past research that it is feasible to optimize the operations of individual TSC systems or a small number of such systems. However, it has been computationally difficult to scale these solution approaches to large networks partly due to the curse of dimensionality that is encountered as the number of intersections increases. Fortunately, recent studies have recognized the potential of exploiting advancements in deep and reinforcement learning to address this problem, and some preliminary successes have been achieved in this regard. However, facilitating such intelligent solution approaches may require large amounts of infrastructural investments such as roadside units (RSUs) and drones in order to ensure thorough connectivity across all intersections in large networks, an investment that may be burdensome for agencies to undertake. As such, this study builds on recent work to present a scalable TSC model that may reduce the number of required enabling infrastructure. This is achieved using graph attention networks (GATs) to serve as the neural network for deep reinforcement learning, which aids in maintaining the graph topology of the traffic network while disregarding any irrelevant or unnecessary information. A case study is carried out to demonstrate the effectiveness of the proposed model, and the results show much promise. The overall research outcome suggests that by decomposing large networks using fog-nodes, the proposed fog-based graphic RL (FG-RL) model can be easily applied to scale into larger traffic networks.

Jiqian Dong, Sikai Chen, Shuya Zong et al.

In the last decade, deep learning (DL) approaches have been used successfully in computer vision (CV) applications. However, DL-based CV models are generally considered to be black boxes due to their lack of interpretability. This black box behavior has exacerbated user distrust and therefore has prevented widespread deployment DLCV models in autonomous driving tasks even though some of these models exhibit superiority over human performance. For this reason, it is essential to develop explainable DL models for autonomous driving task. Explainable DL models can not only boost user trust in autonomy but also serve as a diagnostic approach to identify anydefects and weaknesses of the model during the system development phase. In this paper, we propose an explainable end-to-end autonomous driving system based on "Transformer", a state-of-the-art (SOTA) self-attention based model, to map visual features from images collected by onboard cameras to guide potential driving actions with corresponding explanations. The model achieves a soft attention over the global features of the image. The results demonstrate the efficacy of our proposed model as it exhibits superior performance (in terms of correct prediction of actions and explanations) compared to the benchmark model by a significant margin with lower computational cost.

Jiqian Dong, Sikai Chen, Samuel Labi

It is anticipated that the era of fully autonomous vehicle operations will be preceded by a lengthy "Transition Period" where the traffic stream will be mixed, that is, consisting of connected autonomous vehicles (CAVs), human-driven vehicles (HDVs) and connected human-driven vehicles (CHDVs). In recognition of the fact that public acceptance of CAVs will hinge on safety performance of automated driving systems, and that there will likely be safety challenges in the early part of the transition period, significant research efforts have been expended in the development of safety-conscious automated driving systems. Yet still, there appears to be a lacuna in the literature regarding the handling of the crash-imminent situations that are caused by errant human driven vehicles (HDVs) in the vicinity of the CAV during operations on the roadway. In this paper, we develop a simple model-based Reinforcement Learning (RL) based system that can be deployed in the CAV to generate trajectories that anticipate and avoid potential collisions caused by drivers of the HDVs. The model involves an end-to-end data-driven approach that contains a motion prediction model based on deep learning, and a fast trajectory planning algorithm based on model predictive control (MPC). The proposed system requires no prior knowledge or assumption about the physical environment including the vehicle dynamics, and therefore represents a general approach that can be deployed on any type of vehicle (e.g., truck, buse, motorcycle, etc.). The framework is trained and tested in the CARLA simulator with multiple collision imminent scenarios, and the results indicate the proposed model can avoid the collision at high successful rate (>85%) even in highly compact and dangerous situations.

Sikai Chen, Jiqian Dong, Runjia Du et al.

Deep learning (DL) based computer vision (CV) models are generally considered as black boxes due to poor interpretability. This limitation impedes efficient diagnoses or predictions of system failure, thereby precluding the widespread deployment of DLCV models in safety-critical tasks such as autonomous driving. This study is motivated by the need to enhance the interpretability of DL model in autonomous driving and therefore proposes an explainable DL-based framework that generates textual descriptions of the driving environment and makes appropriate decisions based on the generated descriptions. The proposed framework imitates the learning process of human drivers by jointly modeling the visual input (images) and natural language, while using the language to induce the visual attention in the image. The results indicate strong explainability of autonomous driving decisions obtained by focusing on relevant features from visual inputs. Furthermore, the output attention maps enhance the interpretability of the model not only by providing meaningful explanation to the model behavior but also by identifying the weakness of and potential improvement directions for the model.

Runjia Du, Sikai Chen, Jiqian Dong et al.

Past research and practice have demonstrated that dynamic rerouting framework is effective in mitigating urban traffic congestion and thereby improve urban travel efficiency. It has been suggested that dynamic rerouting could be facilitated using emerging technologies such as fog-computing which offer advantages of low-latency capabilities and information exchange between vehicles and roadway infrastructure. To address this question, this study proposes a two-stage model that combines GAQ (Graph Attention Network - Deep Q Learning) and EBkSP (Entropy Based k Shortest Path) using a fog-cloud architecture, to reroute vehicles in a dynamic urban environment and therefore to improve travel efficiency in terms of travel speed. First, GAQ analyzes the traffic conditions on each road and for each fog area, and then assigns a road index based on the information attention from both local and neighboring areas. Second, EBkSP assigns the route for each vehicle based on the vehicle priority and route popularity. A case study experiment is carried out to investigate the efficacy of the proposed model. At the model training stage, different methods are used to establish the vehicle priorities, and their impact on the results is assessed. Also, the proposed model is tested under various scenarios with different ratios of rerouting and background (non-rerouting) vehicles. The results demonstrate that vehicle rerouting using the proposed model can help attain higher speed and reduces possibility of severe congestion. This result suggests that the proposed model can be deployed by urban transportation agencies for dynamic rerouting and ultimately, to reduce urban traffic congestion.

Yu Qiao, Sikai Chen, Majed Alinizzi et al.

Surface roughness is primary measure of pavement performance that has been associated with ride quality and vehicle operating costs. Of all the surface roughness indicators, the International Roughness Index (IRI) is the most widely used. However, it is costly to measure IRI, and for this reason, certain road classes are excluded from IRI measurements at a network level. Higher levels of distresses are generally associated with higher roughness. However, for a given roughness level, pavement data typically exhibits a great deal of variability in the distress types, density, and severity. It is hypothesized that it is feasible to estimate the IRI of a pavement section given its distress types and their respective densities and severities. To investigate this hypothesis, this paper uses data from in-service pavements and machine learning methods to ascertain the extent to which IRI can be predicted given a set of pavement attributes. The results suggest that machine learning can be used reliably to estimate IRI based on the measured distress types and their respective densities and severities. The analysis also showed that IRI estimated this way depends on the pavement type and functional class. The paper also includes an exploratory section that addresses the reverse situation, that is, estimating the probability of pavement distress type distribution and occurrence severity/extent based on a given roughness level.

Jiqian Dong, Sikai Chen, Paul Young Joun Ha et al.

Connected Autonomous Vehicle (CAV) Network can be defined as a collection of CAVs operating at different locations on a multilane corridor, which provides a platform to facilitate the dissemination of operational information as well as control instructions. Cooperation is crucial in CAV operating systems since it can greatly enhance operation in terms of safety and mobility, and high-level cooperation between CAVs can be expected by jointly plan and control within CAV network. However, due to the highly dynamic and combinatory nature such as dynamic number of agents (CAVs) and exponentially growing joint action space in a multiagent driving task, achieving cooperative control is NP hard and cannot be governed by any simple rule-based methods. In addition, existing literature contains abundant information on autonomous driving's sensing technology and control logic but relatively little guidance on how to fuse the information acquired from collaborative sensing and build decision processor on top of fused information. In this paper, a novel Deep Reinforcement Learning (DRL) based approach combining Graphic Convolution Neural Network (GCN) and Deep Q Network (DQN), namely Graphic Convolution Q network (GCQ) is proposed as the information fusion module and decision processor. The proposed model can aggregate the information acquired from collaborative sensing and output safe and cooperative lane changing decisions for multiple CAVs so that individual intention can be satisfied even under a highly dynamic and partially observed mixed traffic. The proposed algorithm can be deployed on centralized control infrastructures such as road-side units (RSU) or cloud platforms to improve the CAV operation.

Paul Young Joun Ha, Sikai Chen, Jiqian Dong et al.

Active Traffic Management strategies are often adopted in real-time to address such sudden flow breakdowns. When queuing is imminent, Speed Harmonization (SH), which adjusts speeds in upstream traffic to mitigate traffic showckwaves downstream, can be applied. However, because SH depends on driver awareness and compliance, it may not always be effective in mitigating congestion. The use of multiagent reinforcement learning for collaborative learning, is a promising solution to this challenge. By incorporating this technique in the control algorithms of connected and autonomous vehicle (CAV), it may be possible to train the CAVs to make joint decisions that can mitigate highway bottleneck congestion without human driver compliance to altered speed limits. In this regard, we present an RL-based multi-agent CAV control model to operate in mixed traffic (both CAVs and human-driven vehicles (HDVs)). The results suggest that even at CAV percent share of corridor traffic as low as 10%, CAVs can significantly mitigate bottlenecks in highway traffic. Another objective was to assess the efficacy of the RL-based controller vis-à-vis that of the rule-based controller. In addressing this objective, we duly recognize that one of the main challenges of RL-based CAV controllers is the variety and complexity of inputs that exist in the real world, such as the information provided to the CAV by other connected entities and sensed information. These translate as dynamic length inputs which are difficult to process and learn from. For this reason, we propose the use of Graphical Convolution Networks (GCN), a specific RL technique, to preserve information network topology and corresponding dynamic length inputs. We then use this, combined with Deep Deterministic Policy Gradient (DDPG), to carry out multi-agent training for congestion mitigation using the CAV controllers.

Jiqian Dong, Sikai Chen, Yujie Li et al.

The connectivity aspect of connected autonomous vehicles (CAV) is beneficial because it facilitates dissemination of traffic-related information to vehicles through Vehicle-to-External (V2X) communication. Onboard sensing equipment including LiDAR and camera can reasonably characterize the traffic environment in the immediate locality of the CAV. However, their performance is limited by their sensor range (SR). On the other hand, longer-range information is helpful for characterizing imminent conditions downstream. By contemporaneously coalescing the short- and long-range information, the CAV can construct comprehensively its surrounding environment and thereby facilitate informed, safe, and effective movement planning in the short-term (local decisions including lane change) and long-term (route choice). In this paper, we describe a Deep Reinforcement Learning based approach that integrates the data collected through sensing and connectivity capabilities from other vehicles located in the proximity of the CAV and from those located further downstream, and we use the fused data to guide lane changing, a specific context of CAV operations. In addition, recognizing the importance of the connectivity range (CR) to the performance of not only the algorithm but also of the vehicle in the actual driving environment, the paper carried out a case study. The case study demonstrates the application of the proposed algorithm and duly identifies the appropriate CR for each level of prevailing traffic density. It is expected that implementation of the algorithm in CAVs can enhance the safety and mobility associated with CAV driving operations. From a general perspective, its implementation can provide guidance to connectivity equipment manufacturers and CAV operators, regarding the default CR settings for CAVs or the recommended CR setting in a given traffic environment.