arXiv subjects

Zheng Fu

Publications and source records attributed to Zheng Fu.

16 recordsLinked to original sources

arXiv2026

Dual-Flow Reinforcement Learning with State-Aware Exploration

In complex continuous-control reinforcement learning tasks, multimodal optimal actions often coincide with uncertain, multimodal return distributions, making reliable value estimation and multimodal exploration challenging. Existing value estimation methods using unimodal Gaussians restrict expressiveness and yield biased estimates. Recent generative policies can represent multimodal actions but often collapse to a few modes and under-explore high-value areas of the action space. Motivated by these challenges, we propose Dual-Flow RL, a unified actor-critic framework that jointly models a continuous return distribution and a multimodal policy distribution using conditional flow matching (CFM). This design supports reliable value estimation and sustained multimodal exploration. To further enhance exploration, we introduce an Entropy-Covariance Exploration Regulator (ECER) that enables state-aware exploration regulation leveraging policy entropy and action-uncertainty covariance. Experiments on DeepMind Control Suite and Humanoid-Bench show that Dual-Flow RL achieves state-of-the-art performance on most tasks, significantly outperforming prior diffusion-based and flow-based methods.

cs.LG↗

arXiv2026

Envision4D: Envisioning Visual Futures via Feed-forward 4D Gaussian Splatting for Autonomous Driving

Forecasting the future evolution of dynamic scenes is crucial in autonomous driving. However, existing feed-forward paradigms are primarily designed for interpolation. When extended to future extrapolation, they suffer from ghosting artifacts under large displacements and are constrained by simplified motion assumptions or strict future priors. To overcome these challenges, we propose Envision4D, a fully self-supervised feed-forward framework for pose-free future extrapolation. Specifically, we introduce a Future Pose Prediction module that infers future camera parameters via an iterative denoising process. Furthermore, to capture non-linear dynamics, we propose In-layer Temporal Attention and employ Conditioned Motion Lifting, which transforms the highly uncertain extrapolation process into robust relational mappings. Finally, a Progressive Training Strategy is utilized to stabilize unsupervised motion learning against error accumulation. Extensive experiments demonstrate that Envision4D achieves state-of-the-art performance, significantly outperforming existing methods in future view synthesis.

cs.CV↗

arXiv2026

Physics-informed Deep Mixture-of-Koopmans Vehicle Dynamics Model with Dual-branch Encoder for Distributed Electric-drive Trucks

Advanced autonomous driving systems require accurate vehicle dynamics modeling. However, identifying a precise dynamics model remains challenging due to strong nonlinearities and the coupled longitudinal and lateral dynamic characteristics. Previous research has employed physics-based analytical models or neural networks to construct vehicle dynamics representations. Nevertheless, these approaches often struggle to simultaneously achieve satisfactory performance in terms of system identification efficiency, modeling accuracy, and compatibility with linear control strategies. In this paper, we propose a fully data-driven dynamics modeling method tailored for complex distributed electric-drive trucks (DETs), leveraging Koopman operator theory to represent highly nonlinear dynamics in a lifted linear embedding space. To achieve high-precision modeling, we first propose a novel dual-branch encoder which encodes dynamic states and provides a powerful basis for the proposed Koopman-based methods entitled KODE. A physics-informed supervision mechanism, grounded in the geometric consistency of temporal vehicle motion, is incorporated into the training process to facilitate effective learning of both the encoder and the Koopman operator. Furthermore, to accommodate the diverse driving patterns of DETs, we extend the vanilla Koopman operator to a mixture-of-Koopman operator framework, enhancing modeling capability. Simulations conducted in a high-fidelity TruckSim environment and real-world experiments demonstrate that the proposed approach achieves state-of-the-art performance in long-term dynamics state estimation.

cs.RO↗

arXiv2026

ForSim: Stepwise Forward Simulation for Traffic Policy Fine-Tuning

As the foundation of closed-loop training and evaluation in autonomous driving, traffic simulation still faces two fundamental challenges: covariate shift introduced by open-loop imitation learning and limited capacity to reflect the multimodal behaviors observed in real-world traffic. Although recent frameworks such as RIFT have partially addressed these issues through group-relative optimization, their forward simulation procedures remain largely non-reactive, leading to unrealistic agent interactions within the virtual domain and ultimately limiting simulation fidelity. To address these issues, we propose ForSim, a stepwise closed-loop forward simulation paradigm. At each virtual timestep, the traffic agent propagates the virtual candidate trajectory that best spatiotemporally matches the reference trajectory through physically grounded motion dynamics, thereby preserving multimodal behavioral diversity while ensuring intra-modality consistency. Other agents are updated with stepwise predictions, yielding coherent and interaction-aware evolution. When incorporated into the RIFT traffic simulation framework, ForSim operates in conjunction with group-relative optimization to fine-tune traffic policy. Extensive experiments confirm that this integration consistently improves safety while maintaining efficiency, realism, and comfort. These results underscore the importance of modeling closed-loop multimodal interactions within forward simulation and enhance the fidelity and reliability of traffic simulation for autonomous driving. Project Page: https://currychen77.github.io/ForSim/

cs.RO↗

arXiv2025

DTCCL: Disengagement-Triggered Contrastive Continual Learning for Autonomous Bus Planners

Autonomous buses run on fixed routes but must operate in open, dynamic urban environments. Disengagement events on these routes are often geographically concentrated and typically arise from planner failures in highly interactive regions. Such policy-level failures are difficult to correct using conventional imitation learning, which easily overfits to sparse disengagement data. To address this issue, this paper presents a Disengagement-Triggered Contrastive Continual Learning (DTCCL) framework that enables autonomous buses to improve planning policies through real-world operation. Each disengagement triggers cloud-based data augmentation that generates positive and negative samples by perturbing surrounding agents while preserving route context. Contrastive learning refines policy representations to better distinguish safe and unsafe behaviors, and continual updates are applied in a cloud-edge loop without human supervision. Experiments on urban bus routes demonstrate that DTCCL improves overall planning performance by 48.6 percent compared with direct retraining, validating its effectiveness for scalable, closed-loop policy improvement in autonomous public transport.

cs.RO↗

arXiv2025

MTRDrive: Memory-Tool Synergistic Reasoning for Robust Autonomous Driving in Corner Cases

Vision-Language Models(VLMs) have demonstrated significant potential for end-to-end autonomous driving, yet a substantial gap remains between their current capabilities and the reliability necessary for real-world deployment. A critical challenge is their fragility, characterized by hallucinations and poor generalization in out-of-distribution (OOD) scenarios. To bridge this gap, we introduce MTRDrive, a novel framework that integrates procedural driving experiences with a dynamic toolkit to enhance generalization and proactive decision-making. MTRDrive addresses these limitations through a closed-loop system that combines a memory-based experience retrieval mechanism with dynamic toolkits. This synergy enables the model to interact more effectively with its environment, improving both reasoning and decision-making capabilities with the help of our memory-tool synergistic reasoning. Additionally, we introduce a new benchmark based on complex Roadwork construction scenarios to rigorously evaluate zero-shot generalization. Extensive experiments demonstrate the superior effectiveness of our approach. On the public NAVSIM benchmark, our 3B-parameter MTRDrive model achieves an exceptional PDMS of 88.3 without chain-of-thought and sets a state-of-the-art performance bar on high-level planning, with a driving metric score of 79.8\% and a planning accuracy of 82.6\%. Rigorous zero-shot evaluation on the new Roadwork-VLM benchmark shows a strong ability to reason robustly in unseen scenarios, achieving a driving metric score of 80.2\%. These results highlight MTRDrive's potential to advance autonomous driving toward safer and more reliable systems.

cs.RO↗

arXiv2025

EvaDrive: Evolutionary Adversarial Policy Optimization for End-to-End Autonomous Driving

Autonomous driving faces significant challenges in achieving human-like iterative decision-making, which continuously generates, evaluates, and refines trajectory proposals. Current generation-evaluation frameworks isolate trajectory generation from quality assessment, preventing iterative refinement essential for planning, while reinforcement learning methods collapse multi-dimensional preferences into scalar rewards, obscuring critical trade-offs and yielding scalarization bias.To overcome these issues, we present EvaDrive, a novel multi-objective reinforcement learning framework that establishes genuine closed-loop co-evolution between trajectory generation and evaluation via adversarial optimization. EvaDrive frames trajectory planning as a multi-round adversarial game. In this game, a hierarchical generator continuously proposes candidate paths by combining autoregressive intent modeling for temporal causality with diffusion-based refinement for spatial flexibility. These proposals are then rigorously assessed by a trainable multi-objective critic that explicitly preserves diverse preference structures without collapsing them into a single scalarization bias.This adversarial interplay, guided by a Pareto frontier selection mechanism, enables iterative multi-round refinement, effectively escaping local optima while preserving trajectory diversity.Extensive experiments on NAVSIM and Bench2Drive benchmarks demonstrate SOTA performance, achieving 94.9 PDMS on NAVSIM v1 (surpassing DiffusionDrive by 6.8, DriveSuprim by 5.0, and TrajHF by 0.9) and 64.96 Driving Score on Bench2Drive. EvaDrive generates diverse driving styles via dynamic weighting without external preference data, introducing a closed-loop adversarial framework for human-like iterative decision-making, offering a novel scalarization-free trajectory optimization approach.

cs.LG↗

arXiv2025

A Survey on Vision-Language-Action Models for Autonomous Driving

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}.

cs.CV↗

arXiv2025

AgentThink: A Unified Framework for Tool-Augmented Chain-of-Thought Reasoning in Vision-Language Models for Autonomous Driving

Vision-Language Models (VLMs) show promise for autonomous driving, yet their struggle with hallucinations, inefficient reasoning, and limited real-world validation hinders accurate perception and robust step-by-step reasoning. To overcome this, we introduce \textbf{AgentThink}, a pioneering unified framework that integrates Chain-of-Thought (CoT) reasoning with dynamic, agent-style tool invocation for autonomous driving tasks. AgentThink's core innovations include: \textbf{(i) Structured Data Generation}, which establishes an autonomous driving tool library to automatically construct structured, self-verified reasoning data explicitly incorporating tool usage for diverse driving scenarios; \textbf{(ii) A Two-stage Training Pipeline}, employing Supervised Fine-Tuning (SFT) with Group Relative Policy Optimization (GRPO) to equip VLMs with the capability for autonomous tool invocation; and \textbf{(iii) Agent-style Tool-Usage Evaluation}, introducing a novel multi-tool assessment protocol to rigorously evaluate the model's tool invocation and utilization. Experiments on the DriveLMM-o1 benchmark demonstrate that AgentThink significantly boosts overall reasoning scores by \textbf{53.91%} and enhances answer accuracy by \textbf{33.54%}, while markedly improving reasoning quality and consistency. Furthermore, ablation studies and robust zero-shot/few-shot generalization experiments across various benchmarks underscore its powerful capabilities. These findings highlight a promising trajectory for developing trustworthy and tool-aware autonomous driving models. Code is available at https://github.com/curryqka/AgentThink.

cs.RO↗

arXiv2025

FASIONAD++ : Integrating High-Level Instruction and Information Bottleneck in FAt-Slow fusION Systems for Enhanced Safety in Autonomous Driving with Adaptive Feedback

Ensuring safe, comfortable, and efficient planning is crucial for autonomous driving systems. While end-to-end models trained on large datasets perform well in standard driving scenarios, they struggle with complex low-frequency events. Recent Large Language Models (LLMs) and Vision Language Models (VLMs) advancements offer enhanced reasoning but suffer from computational inefficiency. Inspired by the dual-process cognitive model "Thinking, Fast and Slow", we propose $\textbf{FASIONAD}$ -- a novel dual-system framework that synergizes a fast end-to-end planner with a VLM-based reasoning module. The fast system leverages end-to-end learning to achieve real-time trajectory generation in common scenarios, while the slow system activates through uncertainty estimation to perform contextual analysis and complex scenario resolution. Our architecture introduces three key innovations: (1) A dynamic switching mechanism enabling slow system intervention based on real-time uncertainty assessment; (2) An information bottleneck with high-level plan feedback that optimizes the slow system's guidance capability; (3) A bidirectional knowledge exchange where visual prompts enhance the slow system's reasoning while its feedback refines the fast planner's decision-making. To strengthen VLM reasoning, we develop a question-answering mechanism coupled with reward-instruct training strategy. In open-loop experiments, FASIONAD achieves a $6.7\%$ reduction in average $L2$ trajectory error and $28.1\%$ lower collision rate.

cs.RO↗

arXiv2025

LEGO-Motion: Learning-Enhanced Grids with Occupancy Instance Modeling for Class-Agnostic Motion Prediction

Accurate and reliable spatial and motion information plays a pivotal role in autonomous driving systems. However, object-level perception models struggle with handling open scenario categories and lack precise intrinsic geometry. On the other hand, occupancy-based class-agnostic methods excel in representing scenes but fail to ensure physics consistency and ignore the importance of interactions between traffic participants, hindering the model's ability to learn accurate and reliable motion. In this paper, we introduce a novel occupancy-instance modeling framework for class-agnostic motion prediction tasks, named LEGO-Motion, which incorporates instance features into Bird's Eye View (BEV) space. Our model comprises (1) a BEV encoder, (2) an Interaction-Augmented Instance Encoder, and (3) an Instance-Enhanced BEV Encoder, improving both interaction relationships and physics consistency within the model, thereby ensuring a more accurate and robust understanding of the environment. Extensive experiments on the nuScenes dataset demonstrate that our method achieves state-of-the-art performance, outperforming existing approaches. Furthermore, the effectiveness of our framework is validated on the advanced FMCW LiDAR benchmark, showcasing its practical applicability and generalization capabilities. The code will be made publicly available to facilitate further research.

cs.CV↗

arXiv2025

RAVE: End-to-end Hierarchical Visual Localization with Rasterized and Vectorized HD map

Accurate localization serves as an important component in autonomous driving systems. Traditional rule-based localization involves many standalone modules, which is theoretically fragile and requires costly hyperparameter tuning, therefore sacrificing the accuracy and generalization. In this paper, we propose an end-to-end visual localization system, RAVE, in which the surrounding images are associated with the HD map data to estimate pose. To ensure high-quality observations for localization, a low-rank flow-based prior fusion module (FLORA) is developed to incorporate misaligned map prior into the perceived BEV features. Pursuing a balance among efficiency, interpretability, and accuracy, a hierarchical localization module is proposed, which efficiently estimates poses through a decoupled BEV neural matching-based pose solver (DEMA) using rasterized HD map, and then refines the estimation through a Transformer-based pose regressor (POET) using vectorized HD map. The experimental results demonstrate that our method can perform robust and accurate localization under varying environmental conditions while running efficiently.

cs.RO↗

arXiv2024

How Cars Move: Analyzing Driving Dynamics for Safer Urban Traffic

Understanding the spatial dynamics of cars within urban systems is essential for optimizing infrastructure management and resource allocation. Recent empirical approaches for analyzing traffic patterns have gained traction due to their applicability to city-scale policy development. However, conventional methodologies often rely on fragmented grid-based techniques, which may overlook critical interdependencies among spatial elements and temporal continuity. These limitations can compromise analytical effectiveness in complex urban environments. To address these challenges, we propose PriorMotion, a data integration framework designed to systematically uncover movement patterns through driving dynamics analysis. Our approach combines multi-scale empirical observations with customized analytical tools to capture evolving spatial-temporal trends in urban traffic. Comprehensive evaluations demonstrate that PriorMotion significantly enhances analytical outcomes, including increased accuracy in traffic pattern analysis, improved adaptability to heterogeneous data environments, and reduced long-term projection errors. Validation confirms its effectiveness for urban infrastructure management applications requiring precise characterization of complex spatial-temporal interactions.

cs.CV↗

arXiv2024

FASIONAD : FAst and Slow FusION Thinking Systems for Human-Like Autonomous Driving with Adaptive Feedback

Ensuring safe, comfortable, and efficient navigation is a critical goal for autonomous driving systems. While end-to-end models trained on large-scale datasets excel in common driving scenarios, they often struggle with rare, long-tail events. Recent progress in large language models (LLMs) has introduced enhanced reasoning capabilities, but their computational demands pose challenges for real-time decision-making and precise planning. This paper presents FASIONAD, a novel dual-system framework inspired by the cognitive model "Thinking, Fast and Slow." The fast system handles routine navigation tasks using rapid, data-driven path planning, while the slow system focuses on complex reasoning and decision-making in challenging or unfamiliar situations. A dynamic switching mechanism based on score distribution and feedback allows seamless transitions between the two systems. Visual prompts generated by the fast system enable human-like reasoning in the slow system, which provides high-quality feedback to enhance the fast system's decision-making. To evaluate FASIONAD, we introduce a new benchmark derived from the nuScenes dataset, specifically designed to differentiate fast and slow scenarios. FASIONAD achieves state-of-the-art performance on this benchmark, establishing a new standard for frameworks integrating fast and slow cognitive processes in autonomous driving. This approach paves the way for more adaptive, human-like autonomous driving systems.

cs.RO↗

arXiv2024

High Definition Map Mapping and Update: A General Overview and Future Directions

Along with the rapid growth of autonomous vehicles (AVs), more and more demands are required for environment perception technology. Among others, HD mapping has become one of the more prominent roles in helping the vehicle realize essential tasks such as localization and path planning. While increasing research efforts have been directed toward HD Map development. However, a comprehensive overview of the overall HD map mapping and update framework is still lacking. This article introduces the development and current state of the algorithm involved in creating HD map mapping and its maintenance. As part of this study, the primary data preprocessing approach of processing raw data to information ready to feed for mapping and update purposes, semantic segmentation, and localization are also briefly reviewed. Moreover, the map taxonomy, ontology, and quality assessment are extensively discussed, the map data's general representation method is presented, and the mapping algorithm ranging from SLAM to transformers learning-based approaches are also discussed. The development of the HD map update algorithm, from change detection to the update methods, is also presented. Finally, the authors discuss possible future developments and the remaining challenges in HD map mapping and update technology. This paper simultaneously serves as a position paper and tutorial to those new to HD map mapping and update domains.

cs.RO↗

arXiv2023

Poses as Queries: Image-to-LiDAR Map Localization with Transformers

High-precision vehicle localization with commercial setups is a crucial technique for high-level autonomous driving tasks. Localization with a monocular camera in LiDAR map is a newly emerged approach that achieves promising balance between cost and accuracy, but estimating pose by finding correspondences between such cross-modal sensor data is challenging, thereby damaging the localization accuracy. In this paper, we address the problem by proposing a novel Transformer-based neural network to register 2D images into 3D LiDAR map in an end-to-end manner. Poses are implicitly represented as high-dimensional feature vectors called pose queries and can be iteratively updated by interacting with the retrieved relevant information from cross-model features using attention mechanism in a proposed POse Estimator Transformer (POET) module. Moreover, we apply a multiple hypotheses aggregation method that estimates the final poses by performing parallel optimization on multiple randomly initialized pose queries to reduce the network uncertainty. Comprehensive analysis and experimental results on public benchmark conclude that the proposed image-to-LiDAR map localization network could achieve state-of-the-art performances in challenging cross-modal localization tasks.

cs.RO↗