XianPeng Lang

Jingyu Li, Junjie Wu, Dongnan Hu et al.

Recent end-to-end autonomous driving approaches have leveraged Vision-Language Models (VLMs) to enhance planning capabilities in complex driving scenarios. However, VLMs are inherently trained as generalist models, lacking specialized understanding of driving-specific reasoning in 3D space and time. When applied to autonomous driving, these models struggle to establish structured spatial-temporal representations that capture geometric relationships, scene context, and motion patterns critical for safe trajectory planning. To address these limitations, we propose SGDrive, a novel framework that explicitly structures the VLM's representation learning around driving-specific knowledge hierarchies. Built upon a pre-trained VLM backbone, SGDrive decomposes driving understanding into a scene-agent-goal hierarchy that mirrors human driving cognition: drivers first perceive the overall environment (scene context), then attend to safety-critical agents and their behaviors, and finally formulate short-term goals before executing actions. This hierarchical decomposition provides the structured spatial-temporal representation that generalist VLMs lack, integrating multi-level information into a compact yet comprehensive format for trajectory planning. Extensive experiments on the NAVSIM benchmark demonstrate that SGDrive achieves state-of-the-art performance among camera-only methods on both PDMS and EPDMS, validating the effectiveness of hierarchical knowledge structuring for adapting generalist VLMs to autonomous driving.

Xinyang Wang, Qian Liu, Wenjie Ding et al.

Vision-Language-Action (VLA) models are emerging as a promising paradigm for end-to-end autonomous driving, valued for their potential to leverage world knowledge and reason about complex driving scenes. However, existing methods suffer from two critical limitations: a persistent misalignment between language instructions and action outputs, and the inherent inefficiency of typical auto-regressive action generation. In this paper, we introduce LinkVLA, a novel architecture that directly addresses these challenges to enhance both alignment and efficiency. First, we establish a structural link by unifying language and action tokens into a shared discrete codebook, processed within a single multi-modal model. This structurally enforces cross-modal consistency from the ground up. Second, to create a deep semantic link, we introduce an auxiliary action understanding objective that trains the model to generate descriptive captions from trajectories, fostering a bidirectional language-action mapping. Finally, we replace the slow, step-by-step generation with a two-step coarse-to-fine generation method C2F that efficiently decodes the action sequence, saving 86% inference time. Experiments on closed-loop driving benchmarks show consistent gains in instruction following accuracy and driving performance, alongside reduced inference latency.

Wenzhao Zheng, Junjie Wu, Yao Zheng et al.

Vision-based autonomous driving shows great potential due to its satisfactory performance and low costs. Most existing methods adopt dense representations (e.g., bird's eye view) or sparse representations (e.g., instance boxes) for decision-making, which suffer from the trade-off between comprehensiveness and efficiency. This paper explores a Gaussian-centric end-to-end autonomous driving (GaussianAD) framework and exploits 3D semantic Gaussians to extensively yet sparsely describe the scene. We initialize the scene with uniform 3D Gaussians and use surrounding-view images to progressively refine them to obtain the 3D Gaussian scene representation. We then use sparse convolutions to efficiently perform 3D perception (e.g., 3D detection, semantic map construction). We predict 3D flows for the Gaussians with dynamic semantics and plan the ego trajectory accordingly with an objective of future scene forecasting. Our GaussianAD can be trained in an end-to-end manner with optional perception labels when available. Extensive experiments on the widely used nuScenes dataset verify the effectiveness of our end-to-end GaussianAD on various tasks including motion planning, 3D occupancy prediction, and 4D occupancy forecasting. Code: https://github.com/wzzheng/GaussianAD.

Tao Tang, Enhui Ma, xia zhou et al.

Autonomous driving has seen remarkable advancements, largely driven by extensive real-world data collection. However, acquiring diverse and corner-case data remains costly and inefficient. Generative models have emerged as a promising solution by synthesizing realistic sensor data. However, existing approaches primarily focus on single-modality generation, leading to inefficiencies and misalignment in multimodal sensor data. To address these challenges, we propose OminiGen, which generates aligned multimodal sensor data in a unified framework. Our approach leverages a shared Bird\u2019s Eye View (BEV) space to unify multimodal features and designs a novel generalizable multimodal reconstruction method, UAE, to jointly decode LiDAR and multi-view camera data. UAE achieves multimodal sensor decoding through volume rendering, enabling accurate and flexible reconstruction. Furthermore, we incorporate a Diffusion Transformer (DiT) with a ControlNet branch to enable controllable multimodal sensor generation. Our comprehensive experiments demonstrate that OminiGen achieves desired performances in unified multimodal sensor data generation with multimodal consistency and flexible sensor adjustments.

Chengen Xie, Bin Sun, Tianyu Li et al.

End-to-end autonomous driving models trained on largescale datasets perform well in common scenarios but struggle with rare, long-tail situations due to limited scenario diversity. Recent Vision-Language-Action (VLA) models leverage broad knowledge from pre-trained visionlanguage models to address this limitation, yet face critical challenges: (1) numerical imprecision in trajectory prediction due to discrete tokenization, (2) heavy reliance on language annotations that introduce linguistic bias and annotation burden, and (3) computational inefficiency from multi-step chain-of-thought reasoning hinders real-time deployment. We propose LatentVLA, a novel framework that employs self-supervised latent action prediction to train VLA models without language annotations, eliminating linguistic bias while learning rich driving representations from unlabeled trajectory data. Through knowledge distillation, LatentVLA transfers the generalization capabilities of VLA models to efficient vision-based networks, achieving both robust performance and real-time efficiency. LatentVLA establishes a new state-of-the-art on the NAVSIM benchmark with a PDMS score of 92.4 and demonstrates strong zeroshot generalization on the nuScenes benchmark.

Pengxuan Yang, Ben Lu, Zhongpu Xia et al.

Latent World Models enhance scene representation through temporal self-supervised learning, presenting a perception annotation-free paradigm for end-to-end autonomous driving. However, the reconstruction-oriented representation learning tangles perception with planning tasks, leading to suboptimal optimization for planning. To address this challenge, we propose WorldRFT, a planning-oriented latent world model framework that aligns scene representation learning with planning via a hierarchical planning decomposition and local-aware interactive refinement mechanism, augmented by reinforcement learning fine-tuning (RFT) to enhance safety-critical policy performance. Specifically, WorldRFT integrates a vision-geometry foundation model to improve 3D spatial awareness, employs hierarchical planning task decomposition to guide representation optimization, and utilizes local-aware iterative refinement to derive a planning-oriented driving policy. Furthermore, we introduce Group Relative Policy Optimization (GRPO), which applies trajectory Gaussianization and collision-aware rewards to fine-tune the driving policy, yielding systematic improvements in safety. WorldRFT achieves state-of-the-art (SOTA) performance on both open-loop nuScenes and closed-loop NavSim benchmarks. On nuScenes, it reduces collision rates by 83% (0.30% -> 0.05%). On NavSim, using camera-only sensors input, it attains competitive performance with the LiDAR-based SOTA method DiffusionDrive (87.8 vs. 88.1 PDMS).

Tianyi Yan, Junbo Yin, Xianpeng Lang et al.

To enhance autonomous driving safety in complex scenarios, various methods have been proposed to simulate LiDAR point cloud data. Nevertheless, these methods often face challenges in producing high-quality, diverse, and controllable foreground objects. To address the needs of object-aware tasks in 3D perception, we introduce OLiDM, a novel framework capable of generating high-fidelity LiDAR data at both the object and the scene levels. OLiDM consists of two pivotal components: the Object-Scene Progressive Generation (OPG) module and the Object Semantic Alignment (OSA) module. OPG adapts to user-specific prompts to generate desired foreground objects, which are subsequently employed as conditions in scene generation, ensuring controllable outputs at both the object and scene levels. This also facilitates the association of user-defined object-level annotations with the generated LiDAR scenes. Moreover, OSA aims to rectify the misalignment between foreground objects and background scenes, enhancing the overall quality of the generated objects. The broad effectiveness of OLiDM is demonstrated across various LiDAR generation tasks, as well as in 3D perception tasks. Specifically, on the KITTI-360 dataset, OLiDM surpasses prior state-of-the-art methods such as UltraLiDAR by 17.5 in FPD. Additionally, in sparse-to-dense LiDAR completion, OLiDM achieves a significant improvement over LiDARGen, with a 57.47\% increase in semantic IoU. Moreover, OLiDM enhances the performance of mainstream 3D detectors by 2.4\% in mAP and 1.9\% in NDS, underscoring its potential in advancing object-aware 3D tasks. Code is available at: https://yanty123.github.io/OLiDM.

Enhui Ma, Jiahuan Zhang, Guantian Zheng et al.

Multimodal Large Language Models (MLLMs) are rapidly becoming the intelligence brain of end-to-end autonomous driving systems. A key challenge is to assess whether MLLMs can truly understand and follow complex real-world traffic rules. However, existing benchmarks mainly focus on single-rule scenarios like traffic sign recognition, neglecting the complexity of multi-rule concurrency and conflicts in real driving. Consequently, models perform well on simple tasks but often fail or violate rules in real world complex situations. To bridge this gap, we propose DriveCombo, a text and vision-based benchmark for compositional traffic rule reasoning. Inspired by human drivers' cognitive development, we propose a systematic Five-Level Cognitive Ladder that evaluates reasoning from single-rule understanding to multi-rule integration and conflict resolution, enabling quantitative assessment across cognitive stages. We further propose a Rule2Scene Agent that maps language-based traffic rules to dynamic driving scenes through rule crafting and scene generation, enabling scene-level traffic rule visual reasoning. Evaluations of 14 mainstream MLLMs reveal performance drops as task complexity grows, particularly during rule conflicts. After splitting the dataset and fine-tuning on the training set, we further observe substantial improvements in both traffic rule reasoning and downstream planning capabilities. These results highlight the effectiveness of DriveCombo in advancing compliant and intelligent autonomous driving systems.

Xiaoyu Tian, Junru Gu, Bailin Li et al.

A primary hurdle of autonomous driving in urban environments is understanding complex and long-tail scenarios, such as challenging road conditions and delicate human behaviors. We introduce DriveVLM, an autonomous driving system leveraging Vision-Language Models (VLMs) for enhanced scene understanding and planning capabilities. DriveVLM integrates a unique combination of reasoning modules for scene description, scene analysis, and hierarchical planning. Furthermore, recognizing the limitations of VLMs in spatial reasoning and heavy computational requirements, we propose DriveVLM-Dual, a hybrid system that synergizes the strengths of DriveVLM with the traditional autonomous driving pipeline. Experiments on both the nuScenes dataset and our SUP-AD dataset demonstrate the efficacy of DriveVLM and DriveVLM-Dual in handling complex and unpredictable driving conditions. Finally, we deploy the DriveVLM-Dual on a production vehicle, verifying it is effective in real-world autonomous driving environments.

Yunzhi Yan, Haotong Lin, Chenxu Zhou et al.

This paper aims to tackle the problem of modeling dynamic urban streets for autonomous driving scenes. Recent methods extend NeRF by incorporating tracked vehicle poses to animate vehicles, enabling photo-realistic view synthesis of dynamic urban street scenes. However, significant limitations are their slow training and rendering speed. We introduce Street Gaussians, a new explicit scene representation that tackles these limitations. Specifically, the dynamic urban scene is represented as a set of point clouds equipped with semantic logits and 3D Gaussians, each associated with either a foreground vehicle or the background. To model the dynamics of foreground object vehicles, each object point cloud is optimized with optimizable tracked poses, along with a 4D spherical harmonics model for the dynamic appearance. The explicit representation allows easy composition of object vehicles and background, which in turn allows for scene editing operations and rendering at 135 FPS (1066 $\times$ 1600 resolution) within half an hour of training. The proposed method is evaluated on multiple challenging benchmarks, including KITTI and Waymo Open datasets. Experiments show that the proposed method consistently outperforms state-of-the-art methods across all datasets. The code will be released to ensure reproducibility.

Chaojun Ni, Guosheng Zhao, Xiaofeng Wang et al.

Closed-loop simulation is crucial for end-to-end autonomous driving. Existing sensor simulation methods (e.g., NeRF and 3DGS) reconstruct driving scenes based on conditions that closely mirror training data distributions. However, these methods struggle with rendering novel trajectories, such as lane changes. Recent works have demonstrated that integrating world model knowledge alleviates these issues. Despite their efficiency, these approaches still encounter difficulties in the accurate representation of more complex maneuvers, with multi-lane shifts being a notable example. Therefore, we introduce ReconDreamer, which enhances driving scene reconstruction through incremental integration of world model knowledge. Specifically, DriveRestorer is proposed to mitigate artifacts via online restoration. This is complemented by a progressive data update strategy designed to ensure high-quality rendering for more complex maneuvers. To the best of our knowledge, ReconDreamer is the first method to effectively render in large maneuvers. Experimental results demonstrate that ReconDreamer outperforms Street Gaussians in the NTA-IoU, NTL-IoU, and FID, with relative improvements by 24.87%, 6.72%, and 29.97%. Furthermore, ReconDreamer surpasses DriveDreamer4D with PVG during large maneuver rendering, as verified by a relative improvement of 195.87% in the NTA-IoU metric and a comprehensive user study.

Yunzhi Yan, Zhen Xu, Haotong Lin et al.

This paper aims to tackle the problem of photorealistic view synthesis from vehicle sensor data. Recent advancements in neural scene representation have achieved notable success in rendering high-quality autonomous driving scenes, but the performance significantly degrades as the viewpoint deviates from the training trajectory. To mitigate this problem, we introduce StreetCrafter, a novel controllable video diffusion model that utilizes LiDAR point cloud renderings as pixel-level conditions, which fully exploits the generative prior for novel view synthesis, while preserving precise camera control. Moreover, the utilization of pixel-level LiDAR conditions allows us to make accurate pixel-level edits to target scenes. In addition, the generative prior of StreetCrafter can be effectively incorporated into dynamic scene representations to achieve real-time rendering. Experiments on Waymo Open Dataset and PandaSet demonstrate that our model enables flexible control over viewpoint changes, enlarging the view synthesis regions for satisfying rendering, which outperforms existing methods.

Qiao Sun, Huimin Wang, Jiahao Zhan et al.

Large real-world driving datasets have sparked significant research into various aspects of data-driven motion planners for autonomous driving. These include data augmentation, model architecture, reward design, training strategies, and planner pipelines. These planners promise better generalizations on complicated and few-shot cases than previous methods. However, experiment results show that many of these approaches produce limited generalization abilities in planning performance due to overly complex designs or training paradigms. In this paper, we review and benchmark previous methods focusing on generalizations. The experimental results indicate that as models are appropriately scaled, many design elements become redundant. We introduce StateTransformer-2 (STR2), a scalable, decoder-only motion planner that uses a Vision Transformer (ViT) encoder and a mixture-of-experts (MoE) causal Transformer architecture. The MoE backbone addresses modality collapse and reward balancing by expert routing during training. Extensive experiments on the NuPlan dataset show that our method generalizes better than previous approaches across different test sets and closed-loop simulations. Furthermore, we assess its scalability on billions of real-world urban driving scenarios, demonstrating consistent accuracy improvements as both data and model size grow.

Tao Tang, Dafeng Wei, Zhengyu Jia et al.

The rapid development of the autonomous driving industry has led to a significant accumulation of autonomous driving data. Consequently, there comes a growing demand for retrieving data to provide specialized optimization. However, directly applying previous image retrieval methods faces several challenges, such as the lack of global feature representation and inadequate text retrieval ability for complex driving scenes. To address these issues, firstly, we propose the BEV-TSR framework which leverages descriptive text as an input to retrieve corresponding scenes in the Bird's Eye View (BEV) space. Then to facilitate complex scene retrieval with extensive text descriptions, we employ a large language model (LLM) to extract the semantic features of the text inputs and incorporate knowledge graph embeddings to enhance the semantic richness of the language embedding. To achieve feature alignment between the BEV feature and language embedding, we propose Shared Cross-modal Embedding with a set of shared learnable embeddings to bridge the gap between these two modalities, and employ a caption generation task to further enhance the alignment. Furthermore, there lack of well-formed retrieval datasets for effective evaluation. To this end, we establish a multi-level retrieval dataset, nuScenes-Retrieval, based on the widely adopted nuScenes dataset. Experimental results on the multi-level nuScenes-Retrieval show that BEV-TSR achieves state-of-the-art performance, e.g., 85.78% and 87.66% top-1 accuracy on scene-to-text and text-to-scene retrieval respectively. Codes and datasets will be available.

Derun Li, Changye Li, Yue Wang et al.

Generating human-like and adaptive trajectories is essential for autonomous driving in dynamic environments. While generative models have shown promise in synthesizing feasible trajectories, they often fail to capture the nuanced variability of personalized driving styles due to dataset biases and distributional shifts. To address this, we introduce TrajHF, a human feedback-driven finetuning framework for generative trajectory models, designed to align motion planning with diverse driving styles. TrajHF incorporates multi-conditional denoiser and reinforcement learning with human feedback to refine multi-modal trajectory generation beyond conventional imitation learning. This enables better alignment with human driving preferences while maintaining safety and feasibility constraints. TrajHF achieves performance comparable to the state-of-the-art on NavSim benchmark. TrajHF sets a new paradigm for personalized and adaptable trajectory generation in autonomous driving.

Jian Zhu, Zhengyu Jia, Tian Gao et al.

Advanced end-to-end autonomous driving systems predict other vehicles' motions and plan ego vehicle's trajectory. The world model that can foresee the outcome of the trajectory has been used to evaluate the autonomous driving system. However, existing world models predominantly emphasize the trajectory of the ego vehicle and leave other vehicles uncontrollable. This limitation hinders their ability to realistically simulate the interaction between the ego vehicle and the driving scenario. In this paper, we propose a driving World Model named EOT-WM, unifying Ego-Other vehicle Trajectories in videos for driving simulation. Specifically, it remains a challenge to match multiple trajectories in the BEV space with each vehicle in the video to control the video generation. We first project ego-other vehicle trajectories in the BEV space into the image coordinate for vehicle-trajectory match via pixel positions. Then, trajectory videos are encoded by the Spatial-Temporal Variational Auto Encoder to align with driving video latents spatially and temporally in the unified visual space. A trajectory-injected diffusion Transformer is further designed to denoise the noisy video latents for video generation with the guidance of ego-other vehicle trajectories. In addition, we propose a metric based on control latent similarity to evaluate the controllability of trajectories. Extensive experiments are conducted on the nuScenes dataset, and the proposed model outperforms the state-of-the-art method by 30% in FID and 55% in FVD. The model can also predict unseen driving scenes with self-produced trajectories.

Titong Jiang, Xuefeng Jiang, Yuan Ma et al.

We present LightVLA, a simple yet effective differentiable token pruning framework for vision-language-action (VLA) models. While VLA models have shown impressive capability in executing real-world robotic tasks, their deployment on resource-constrained platforms is often bottlenecked by the heavy attention-based computation over large sets of visual tokens. LightVLA addresses this challenge through adaptive, performance-driven pruning of visual tokens: It generates dynamic queries to evaluate visual token importance, and adopts Gumbel softmax to enable differentiable token selection. Through fine-tuning, LightVLA learns to preserve the most informative visual tokens while pruning tokens which do not contribute to task execution, thereby improving efficiency and performance simultaneously. Notably, LightVLA requires no heuristic magic numbers and introduces no additional trainable parameters, making it compatible with modern inference frameworks. Experimental results demonstrate that LightVLA outperforms different VLA models and existing token pruning methods across diverse tasks on the LIBERO benchmark, achieving higher success rates with substantially reduced computational overhead. Specifically, LightVLA reduces FLOPs and latency by 59.1% and 38.2% respectively, with a 2.6% improvement in task success rate. Meanwhile, we also investigate the learnable query-based token pruning method LightVLA* with additional trainable parameters, which also achieves satisfactory performance. Our work reveals that as VLA pursues optimal performance, LightVLA spontaneously learns to prune tokens from a performance-driven perspective. To the best of our knowledge, LightVLA is the first work to apply adaptive visual token pruning to VLA tasks with the collateral goals of efficiency and performance, marking a significant step toward more efficient, powerful and practical real-time robotic systems.

Weicheng Zheng, Xiaofei Mao, Nanfei Ye et al.

The advent of Vision-Language Models (VLMs) has significantly advanced end-to-end autonomous driving, demonstrating powerful reasoning abilities for high-level behavior planning tasks. However, existing methods are often constrained by a passive perception paradigm, relying solely on text-based reasoning. This passivity restricts the model's capacity to actively seek crucial visual evidence when faced with uncertainty. To address this, we introduce DriveAgent-R1, the first autonomous driving agent capable of active perception for planning. In complex scenarios, DriveAgent-R1 proactively invokes tools to perform visual reasoning, firmly grounding its decisions in visual evidence, thereby enhancing both interpretability and reliability. Furthermore, we propose a hybrid thinking framework, inspired by human driver cognitive patterns, allowing the agent to adaptively switch between efficient text-only reasoning and robust tool-augmented visual reasoning based on scene complexity. This capability is cultivated through a three-stage progressive training strategy, featuring a core Cascaded Reinforcement Learning (Cascaded RL) phase. Extensive experiments on the Drive-Internal dataset, which is rich in long-tail scenarios, and the public nuScenes dataset show that, with only 3B parameters, DriveAgent-R1 achieves competitive performance comparable to top closed model systems such as GPT-5 and to human driving proficiency while remaining deployment-friendly, offering a proven path toward building more intelligent autonomous driving systems.

Junshan Hu, Jialiang Mao, Zhikang Liu et al.

Conventional Vision-Language Models(VLMs) typically utilize a fixed number of vision tokens, regardless of task complexity. This one-size-fits-all strategy introduces notable inefficiencies: using excessive tokens leads to unnecessary computational overhead in simpler tasks, whereas insufficient tokens compromise fine-grained visual comprehension in more complex contexts. To overcome these limitations, we present TokenFLEX, an innovative and adaptable vision-language framework that encodes images into a variable number of tokens for efficient integration with a Large Language Model (LLM). Our approach is underpinned by two pivotal innovations. Firstly, we present a novel training paradigm that enhances performance across varying numbers of vision tokens by stochastically modulating token counts during training. Secondly, we design a lightweight vision token projector incorporating an adaptive pooling layer and SwiGLU, allowing for flexible downsampling of vision tokens and adaptive selection of features tailored to specific token counts. Comprehensive experiments reveal that TokenFLEX consistently outperforms its fixed-token counterparts, achieving notable performance gains across various token counts enhancements of 1.6%, 1.0%, and 0.4% with 64, 144, and 256 tokens, respectively averaged over eight vision-language benchmarks. These results underscore TokenFLEX's remarkable flexibility while maintaining high-performance vision-language understanding.

Pengxiang Li, Yinan Zheng, Yue Wang et al.

End-to-End (E2E) solutions have emerged as a mainstream approach for autonomous driving systems, with Vision-Language-Action (VLA) models representing a new paradigm that leverages pre-trained multimodal knowledge from Vision-Language Models (VLMs) to interpret and interact with complex real-world environments. However, these methods remain constrained by the limitations of imitation learning, which struggles to inherently encode physical rules during training. Existing approaches often rely on complex rule-based post-refinement, employ reinforcement learning that remains largely limited to simulation, or utilize diffusion guidance that requires computationally expensive gradient calculations. To address these challenges, we introduce ReflectDrive, a novel learning-based framework that integrates a reflection mechanism for safe trajectory generation via discrete diffusion. We first discretize the two-dimensional driving space to construct an action codebook, enabling the use of pre-trained Diffusion Language Models for planning tasks through fine-tuning. Central to our approach is a safety-aware reflection mechanism that performs iterative self-correction without gradient computation. Our method begins with goal-conditioned trajectory generation to model multi-modal driving behaviors. Based on this, we apply local search methods to identify unsafe tokens and determine feasible solutions, which then serve as safe anchors for inpainting-based regeneration. Evaluated on the NAVSIM benchmark, ReflectDrive demonstrates significant advantages in safety-critical trajectory generation, offering a scalable and reliable solution for autonomous driving systems.

Xuefeng Jiang, Yuan Ma, Pengxiang Li et al.

In recent years, diffusion models have demonstrated remarkable potential across diverse domains, from vision generation to language modeling. Transferring its generative capabilities to modern end-to-end autonomous driving systems has also emerged as a promising direction. However, existing diffusion-based trajectory generative models often exhibit mode collapse where different random noises converge to similar trajectories after the denoising process.Therefore, state-of-the-art models often rely on anchored trajectories from pre-defined trajectory vocabulary or scene priors in the training set to mitigate collapse and enrich the diversity of generated trajectories, but such inductive bias are not available in real-world deployment, which can be challenged when generalizing to unseen scenarios. In this work, we investigate the possibility of effectively tackling the mode collapse challenge without the assumption of pre-defined trajectory vocabulary or pre-computed scene priors. Specifically, we propose TransDiffuser, an encoder-decoder based generative trajectory planning model, where the encoded scene information and motion states serve as the multi-modal conditional input of the denoising decoder. Different from existing approaches, we exploit a simple yet effective multi-modal representation decorrelation optimization mechanism during the denoising process to enrich the latent representation space which better guides the downstream generation. Without any predefined trajectory anchors or pre-computed scene priors, TransDiffuser achieves the PDMS of 94.85 on the closed-loop planning-oriented benchmark NAVSIM, surpassing previous state-of-the-art methods. Qualitative evaluation further showcases TransDiffuser generates more diverse and plausible trajectories which explore more drivable area.

Yuyin Chen, Yida Wang, Xueyang Zhang et al.

Urban scene reconstruction requires modeling both static infrastructure and dynamic elements while supporting diverse environmental conditions. We present \textbf{StyledStreets}, a multi-style street simulator that achieves instruction-driven scene editing with guaranteed spatial and temporal consistency. Building on a state-of-the-art Gaussian Splatting framework for street scenarios enhanced by our proposed pose optimization and multi-view training, our method enables photorealistic style transfers across seasons, weather conditions, and camera setups through three key innovations: First, a hybrid embedding scheme disentangles persistent scene geometry from transient style attributes, allowing realistic environmental edits while preserving structural integrity. Second, uncertainty-aware rendering mitigates supervision noise from diffusion priors, enabling robust training across extreme style variations. Third, a unified parametric model prevents geometric drift through regularized updates, maintaining multi-view consistency across seven vehicle-mounted cameras. Our framework preserves the original scene's motion patterns and geometric relationships. Qualitative results demonstrate plausible transitions between diverse conditions (snow, sandstorm, night), while quantitative evaluations show state-of-the-art geometric accuracy under style transfers. The approach establishes new capabilities for urban simulation, with applications in autonomous vehicle testing and augmented reality systems requiring reliable environmental consistency. Codes will be publicly available upon publication.

Yuru Wang, Pei Liu, Songtao Wang et al.

Open-world 3D scene understanding is a critical challenge that involves recognizing and distinguishing diverse objects and categories from 3D data, such as point clouds, without relying on manual annotations. Traditional methods struggle with this open-world task, especially due to the limitations of constructing extensive point cloud-text pairs and handling multimodal data effectively. In response to these challenges, we present UniPLV, a robust framework that unifies point clouds, images, and text within a single learning paradigm for comprehensive 3D scene understanding. UniPLV leverages images as a bridge to co-embed 3D points with pre-aligned images and text in a shared feature space, eliminating the need for labor-intensive point cloud-text pair crafting. Our framework achieves precise multimodal alignment through two innovative strategies: (i) Logit and feature distillation modules between images and point clouds to enhance feature coherence; (ii) A vision-point matching module that implicitly corrects 3D semantic predictions affected by projection inaccuracies from points to pixels. To further boost performance, we implement four task-specific losses alongside a two-stage training strategy. Extensive experiments demonstrate that UniPLV significantly surpasses state-of-the-art methods, with average improvements of 15.6% and 14.8% in semantic segmentation for Base-Annotated and Annotation-Free tasks, respectively. These results underscore UniPLV's efficacy in pushing the boundaries of open-world 3D scene understanding. We will release the code to support future research and development.

Zhewen Wan, Tianchen Song, Chen Lin et al.

Diffusion-based large multimodal models, such as LLaDA-V, have demonstrated impressive capabilities in vision-language understanding and generation. However, their bidirectional attention mechanism and diffusion-style iterative denoising paradigm introduce significant computational overhead, as visual tokens are repeatedly processed across all layers and denoising steps. In this work, we conduct an in-depth attention analysis and reveal that, unlike autoregressive decoders, LLaDA-V aggregates cross-modal information predominantly in middle-to-late layers, leading to delayed semantic alignment. Motivated by this observation, we propose a structured token pruning strategy inspired by FastV, selectively removing a proportion of visual tokens at designated layers to reduce FLOPs while preserving critical semantic information. To the best of our knowledge, this is the first work to investigate structured token pruning in diffusion-based large multimodal models. Unlike FastV, which focuses on shallow-layer pruning, our method targets the middle-to-late layers of the first denoising step to align with LLaDA-V's delayed attention aggregation to maintain output quality, and the first-step pruning strategy reduces the computation across all subsequent steps. Our framework provides an empirical basis for efficient LLaDA-V inference and highlights the potential of vision-aware pruning in diffusion-based multimodal models. Across multiple benchmarks, our best configuration reduces computational cost by up to 65% while preserving an average of 95% task performance.

Pei Liu, Songtao Wang, Lang Zhang et al.

Synthesizing high-fidelity and controllable 4D LiDAR data is crucial for creating scalable simulation environments for autonomous driving. This task is inherently challenging due to the sensor's unique spherical geometry, the temporal sparsity of point clouds, and the complexity of dynamic scenes. To address these challenges, we present LiSTAR, a novel generative world model that operates directly on the sensor's native geometry. LiSTAR introduces a Hybrid-Cylindrical-Spherical (HCS) representation to preserve data fidelity by mitigating quantization artifacts common in Cartesian grids. To capture complex dynamics from sparse temporal data, it utilizes a Spatio-Temporal Attention with Ray-Centric Transformer (START) that explicitly models feature evolution along individual sensor rays for robust temporal coherence. Furthermore, for controllable synthesis, we propose a novel 4D point cloud-aligned voxel layout for conditioning and a corresponding discrete Masked Generative START (MaskSTART) framework, which learns a compact, tokenized representation of the scene, enabling efficient, high-resolution, and layout-guided compositional generation. Comprehensive experiments validate LiSTAR's state-of-the-art performance across 4D LiDAR reconstruction, prediction, and conditional generation, with substantial quantitative gains: reducing generation MMD by a massive 76%, improving reconstruction IoU by 32%, and lowering prediction L1 Med by 50%. This level of performance provides a powerful new foundation for creating realistic and controllable autonomous systems simulations. Project link: https://ocean-luna.github.io/LiSTAR.gitub.io.

Kaiwen Cai, Xinze Liu, Xia Zhou et al.

The generation of realistic LiDAR point clouds plays a crucial role in the development and evaluation of autonomous driving systems. Although recent methods for 3D LiDAR point cloud generation have shown significant improvements, they still face notable limitations, including the lack of sequential generation capabilities and the inability to produce accurately positioned foreground objects and realistic backgrounds. These shortcomings hinder their practical applicability. In this paper, we introduce DriveLiDAR4D, a novel LiDAR generation pipeline consisting of multimodal conditions and a novel sequential noise prediction model LiDAR4DNet, capable of producing temporally consistent LiDAR scenes with highly controllable foreground objects and realistic backgrounds. To the best of our knowledge, this is the first work to address the sequential generation of LiDAR scenes with full scene manipulation capability in an end-to-end manner. We evaluated DriveLiDAR4D on the nuScenes and KITTI datasets, where we achieved an FRD score of 743.13 and an FVD score of 16.96 on the nuScenes dataset, surpassing the current state-of-the-art (SOTA) method, UniScene, with an performance boost of 37.2% in FRD and 24.1% in FVD, respectively.

Enhui Ma, Lijun Zhou, Tao Tang et al.

End-to-end planning methods are the de facto standard of the current autonomous driving system, while the robustness of the data-driven approaches suffers due to the notorious long-tail problem (i.e., rare but safety-critical failure cases). In this work, we explore whether recent diffusion-based video generation methods (a.k.a. world models), paired with structured 3D layouts, can enable a fully automated pipeline to self-correct such failure cases. We first introduce an agent to simulate the role of product manager, dubbed PM-Agent, which formulates data requirements to collect data similar to the failure cases. Then, we use a generative model that can simulate both data collection and annotation. However, existing generative models struggle to generate high-fidelity data conditioned on 3D layouts. To address this, we propose DriveSora, which can generate spatiotemporally consistent videos aligned with the 3D annotations requested by PM-Agent. We integrate these components into our self-correcting agentic system, CorrectAD. Importantly, our pipeline is an end-to-end model-agnostic and can be applied to improve any end-to-end planner. Evaluated on both nuScenes and a more challenging in-house dataset across multiple end-to-end planners, CorrectAD corrects 62.5% and 49.8% of failure cases, reducing collision rates by 39% and 27%, respectively.

Pei Liu, Qingtian Ning, Xinyan Lu et al.

Recent advances in vision-language models (VLMs) have demonstrated impressive spatial reasoning capabilities for autonomous driving, yet existing methods predominantly focus on static scene understanding while neglecting the essential temporal dimension of real-world driving scenarios. To address this critical limitation, we propose the OmniReason framework, which establishes robust spatiotemporal reasoning by jointly modeling dynamic 3D environments and their underlying decision-making processes. Our work makes two fundamental advances: (1) We introduce OmniReason-Data, two large-scale vision-language-action (VLA) datasets with dense spatiotemporal annotations and natural language explanations, generated through a novel hallucination-mitigated auto-labeling pipeline that ensures both physical plausibility and temporal coherence; (2) We develop the OmniReason-Agent architecture, which integrates a sparse temporal memory module for persistent scene context modeling and an explanation generator that produces human-interpretable decision rationales, facilitated by our spatiotemporal knowledge distillation approach that effectively captures spatiotemporal causal reasoning patterns. Comprehensive experiments demonstrate state-of-the-art performance, where OmniReason-Agent achieves significant improvements in both open-loop planning tasks and visual question answering (VQA) benchmarks, while establishing new capabilities for interpretable, temporally-aware autonomous vehicles operating in complex, dynamic environments.

Xingyue Peng, Yuandong Lyu, Lang Zhang et al.

Road surface reconstruction is essential for autonomous driving, supporting centimeter-accurate lane perception and high-definition mapping in complex urban environments.While recent methods based on mesh rendering or 3D Gaussian splatting (3DGS) achieve promising results under clean and static conditions, they remain vulnerable to occlusions from dynamic agents, visual clutter from static obstacles, and appearance degradation caused by lighting and weather changes. We present a robust reconstruction framework that integrates occlusion-aware 2D Gaussian surfels with semantic-guided color enhancement to recover clean, consistent road surfaces. Our method leverages a planar-adapted Gaussian representation for efficient large-scale modeling, employs segmentation-guided video inpainting to remove both dynamic and static foreground objects, and enhances color coherence via semantic-aware correction in HSV space. Extensive experiments on urban-scale datasets demonstrate that our framework produces visually coherent and geometrically faithful reconstructions, significantly outperforming prior methods under real-world conditions.

Yida Wang, Xueyang Zhang, Kun Zhan et al.

Neural surface reconstruction faces persistent challenges in reconciling geometric fidelity with photometric consistency under complex scene conditions. We present HiNeuS, a unified framework that holistically addresses three core limitations in existing approaches: multi-view radiance inconsistency, missing keypoints in textureless regions, and structural degradation from over-enforced Eikonal constraints during joint optimization. To resolve these issues through a unified pipeline, we introduce: 1) Differential visibility verification through SDF-guided ray tracing, resolving reflection ambiguities via continuous occlusion modeling; 2) Planar-conformal regularization via ray-aligned geometry patches that enforce local surface coherence while preserving sharp edges through adaptive appearance weighting; and 3) Physically-grounded Eikonal relaxation that dynamically modulates geometric constraints based on local radiance gradients, enabling detail preservation without sacrificing global regularity. Unlike prior methods that handle these aspects through sequential optimizations or isolated modules, our approach achieves cohesive integration where appearance-geometry constraints evolve synergistically throughout training. Comprehensive evaluations across synthetic and real-world datasets demonstrate state-of-the-art performance, including a 21.4% reduction in Chamfer distance over reflection-aware baselines and 2.32 dB PSNR improvement against neural rendering counterparts. Qualitative analyses reveal superior capability in recovering specular instruments, urban layouts with centimeter-scale infrastructure, and low-textured surfaces without local patch collapse. The method's generalizability is further validated through successful application to inverse rendering tasks, including material decomposition and view-consistent relighting.

Tao Tang, Lijun Zhou, Pengkun Hao et al.

3D multiple object tracking (MOT) plays a crucial role in autonomous driving perception. Recent end-to-end query-based trackers simultaneously detect and track objects, which have shown promising potential for the 3D MOT task. However, existing methods are still in the early stages of development and lack systematic improvements, failing to track objects in certain complex scenarios, like occlusions and the small size of target object's situations. In this paper, we first summarize the current end-to-end 3D MOT framework by decomposing it into three constituent parts: query initialization, query propagation, and query matching. Then we propose corresponding improvements, which lead to a strong yet simple tracker: S2-Track. Specifically, for query initialization, we present 2D-Prompted Query Initialization, which leverages predicted 2D object and depth information to prompt an initial estimate of the object's 3D location. For query propagation, we introduce an Uncertainty-aware Probabilistic Decoder to capture the uncertainty of complex environment in object prediction with probabilistic attention. For query matching, we propose a Hierarchical Query Denoising strategy to enhance training robustness and convergence. As a result, our S2-Track achieves state-of-the-art performance on nuScenes benchmark, i.e., 66.3% AMOTA on test split, surpassing the previous best end-to-end solution by a significant margin of 8.9% AMOTA. We achieve 1st place on the nuScenes tracking task leaderboard.

Enhui Ma, Lijun Zhou, Tao Tang et al.

Using generative models to synthesize new data has become a de-facto standard in autonomous driving to address the data scarcity issue. Though existing approaches are able to boost perception models, we discover that these approaches fail to improve the performance of planning of end-to-end autonomous driving models as the generated videos are usually less than 8 frames and the spatial and temporal inconsistencies are not negligible. To this end, we propose Delphi, a novel diffusion-based long video generation method with a shared noise modeling mechanism across the multi-views to increase spatial consistency, and a feature-aligned module to achieves both precise controllability and temporal consistency. Our method can generate up to 40 frames of video without loss of consistency which is about 5 times longer compared with state-of-the-art methods. Instead of randomly generating new data, we further design a sampling policy to let Delphi generate new data that are similar to those failure cases to improve the sample efficiency. This is achieved by building a failure-case driven framework with the help of pre-trained visual language models. Our extensive experiment demonstrates that our Delphi generates a higher quality of long videos surpassing previous state-of-the-art methods. Consequentially, with only generating 4% of the training dataset size, our framework is able to go beyond perception and prediction tasks, for the first time to the best of our knowledge, boost the planning performance of the end-to-end autonomous driving model by a margin of 25%.