Tuopu Wen

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}.

Yining Shi, Kun Jiang, Qiang Meng et al. · tsinghua

World models are critical for autonomous driving to simulate environmental dynamics and generate synthetic data. Existing methods struggle to disentangle ego-vehicle motion (perspective shifts) from scene evolvement (agent interactions), leading to suboptimal predictions. Instead, we propose to separate environmental changes from ego-motion by leveraging the scene-centric coordinate systems. In this paper, we introduce COME: a framework that integrates scene-centric forecasting Control into the Occupancy world ModEl. Specifically, COME first generates ego-irrelevant, spatially consistent future features through a scene-centric prediction branch, which are then converted into scene condition using a tailored ControlNet. These condition features are subsequently injected into the occupancy world model, enabling more accurate and controllable future occupancy predictions. Experimental results on the nuScenes-Occ3D dataset show that COME achieves consistent and significant improvements over state-of-the-art (SOTA) methods across diverse configurations, including different input sources (ground-truth, camera-based, fusion-based occupancy) and prediction horizons (3s and 8s). For example, under the same settings, COME achieves 26.3% better mIoU metric than DOME and 23.7% better mIoU metric than UniScene. These results highlight the efficacy of disentangled representation learning in enhancing spatio-temporal prediction fidelity for world models. Code and videos will be available at https://github.com/synsin0/COME.

Peijin Jia, Ziang Luo, Tuopu Wen et al.

In autonomous driving, recent advances in lane segment perception provide autonomous vehicles with a comprehensive understanding of driving scenarios. Moreover, incorporating prior information input into such perception model represents an effective approach to ensure the robustness and accuracy. However, utilizing diverse sources of prior information still faces three key challenges: the acquisition of high-quality prior information, alignment between prior and online perception, efficient integration. To address these issues, we investigate prior augmentation from a novel perspective of trajectory priors. In this paper, we initially extract crowdsourcing trajectory data from Argoverse2 motion forecasting dataset and encode trajectory data into rasterized heatmap and vectorized instance tokens, then we incorporate such prior information into the online mapping model through different ways. Besides, with the purpose of mitigating the misalignment between prior and online perception, we design a confidence-based fusion module that takes alignment into account during the fusion process. We conduct extensive experiments on OpenLane-V2 dataset. The results indicate that our method's performance significantly outperforms the current state-of-the-art methods. Code is released is at https://github.com/wowlza/TrajTopo

Peijin Jia, Tuopu Wen, Ziang Luo et al.

Constructing high-definition (HD) maps is a crucial requirement for enabling autonomous driving. In recent years, several map segmentation algorithms have been developed to address this need, leveraging advancements in Bird's-Eye View (BEV) perception. However, existing models still encounter challenges in producing realistic and consistent semantic map layouts. One prominent issue is the limited utilization of structured priors inherent in map segmentation masks. In light of this, we propose DiffMap, a novel approach specifically designed to model the structured priors of map segmentation masks using latent diffusion model. By incorporating this technique, the performance of existing semantic segmentation methods can be significantly enhanced and certain structural errors present in the segmentation outputs can be effectively rectified. Notably, the proposed module can be seamlessly integrated into any map segmentation model, thereby augmenting its capability to accurately delineate semantic information. Furthermore, through extensive visualization analysis, our model demonstrates superior proficiency in generating results that more accurately reflect real-world map layouts, further validating its efficacy in improving the quality of the generated maps.

Chandan Kumar Sah, Ankit Kumar Shaw, Xiaoli Lian et al.

Autonomous vehicles (AVs) require reliable traffic sign recognition and robust lane detection capabilities to ensure safe navigation in complex and dynamic environments. This paper introduces an integrated approach combining advanced deep learning techniques and Multimodal Large Language Models (MLLMs) for comprehensive road perception. For traffic sign recognition, we systematically evaluate ResNet-50, YOLOv8, and RT-DETR, achieving state-of-the-art performance of 99.8% with ResNet-50, 98.0% accuracy with YOLOv8, and achieved 96.6% accuracy in RT-DETR despite its higher computational complexity. For lane detection, we propose a CNN-based segmentation method enhanced by polynomial curve fitting, which delivers high accuracy under favorable conditions. Furthermore, we introduce a lightweight, Multimodal, LLM-based framework that directly undergoes instruction tuning using small yet diverse datasets, eliminating the need for initial pretraining. This framework effectively handles various lane types, complex intersections, and merging zones, significantly enhancing lane detection reliability by reasoning under adverse conditions. Despite constraints in available training resources, our multimodal approach demonstrates advanced reasoning capabilities, achieving a Frame Overall Accuracy (FRM) of 53.87%, a Question Overall Accuracy (QNS) of 82.83%, lane detection accuracies of 99.6% in clear conditions and 93.0% at night, and robust performance in reasoning about lane invisibility due to rain (88.4%) or road degradation (95.6%). The proposed comprehensive framework markedly enhances AV perception reliability, thus contributing significantly to safer autonomous driving across diverse and challenging road scenarios.

Jiusi Li, Jackson Jiang, Jinyu Miao et al.

Corner cases are crucial for training and validating autonomous driving systems, yet collecting them from the real world is often costly and hazardous. Editing objects within captured sensor data offers an effective alternative for generating diverse scenarios, commonly achieved through 3D Gaussian Splatting or image generative models. However, these approaches often suffer from limited visual fidelity or imprecise pose control. To address these issues, we propose G^2Editor, a framework designed for photorealistic and precise object editing in driving videos. Our method leverages a 3D Gaussian representation of the edited object as a dense prior, injected into the denoising process to ensure accurate pose control and spatial consistency. A scene-level 3D bounding box layout is employed to reconstruct occluded areas of non-target objects. Furthermore, to guide the appearance details of the edited object, we incorporate hierarchical fine-grained features as additional conditions during generation. Experiments on the Waymo Open Dataset demonstrate that G^2Editor effectively supports object repositioning, insertion, and deletion within a unified framework, outperforming existing methods in both pose controllability and visual quality, while also benefiting downstream data-driven tasks.

Ankit Kumar Shaw, Kun Jiang, Tuopu Wen et al. · tsinghua

The rapid growth of intelligent connected vehicles (ICVs) and integrated vehicle-road-cloud systems has increased the demand for accurate, real-time HD map updates. However, ensuring map reliability remains challenging due to inconsistencies in crowdsourced data, which suffer from motion blur, lighting variations, adverse weather, and lane marking degradation. This paper introduces CleanMAP, a Multimodal Large Language Model (MLLM)-based distillation framework designed to filter and refine crowdsourced data for high-confidence HD map updates. CleanMAP leverages an MLLM-driven lane visibility scoring model that systematically quantifies key visual parameters, assigning confidence scores (0-10) based on their impact on lane detection. A novel dynamic piecewise confidence-scoring function adapts scores based on lane visibility, ensuring strong alignment with human evaluations while effectively filtering unreliable data. To further optimize map accuracy, a confidence-driven local map fusion strategy ranks and selects the top-k highest-scoring local maps within an optimal confidence range (best score minus 10%), striking a balance between data quality and quantity. Experimental evaluations on a real-world autonomous vehicle dataset validate CleanMAP's effectiveness, demonstrating that fusing the top three local maps achieves the lowest mean map update error of 0.28m, outperforming the baseline (0.37m) and meeting stringent accuracy thresholds (<= 0.32m). Further validation with real-vehicle data confirms 84.88% alignment with human evaluators, reinforcing the model's robustness and reliability. This work establishes CleanMAP as a scalable and deployable solution for crowdsourced HD map updates, ensuring more precise and reliable autonomous navigation. The code will be available at https://Ankit-Zefan.github.io/CleanMap/

Xuewei Tang, Mengmeng Yang, Tuopu Wen et al.

With the growing interest in autonomous driving, there is an increasing demand for accurate and reliable road perception technologies. In complex environments without high-definition map support, autonomous vehicles must independently interpret their surroundings to ensure safe and robust decision-making. However, these scenarios pose significant challenges due to the large number, complex geometries, and frequent occlusions of road elements. A key limitation of existing approaches lies in their insufficient exploitation of the structured priors inherently present in road elements, resulting in irregular, inaccurate predictions. To address this, we propose PriorFusion, a unified framework that effectively integrates semantic, geometric, and generative priors to enhance road element perception. We introduce an instance-aware attention mechanism guided by shape-prior features, then construct a data-driven shape template space that encodes low-dimensional representations of road elements, enabling clustering to generate anchor points as reference priors. We design a diffusion-based framework that leverages these prior anchors to generate accurate and complete predictions. Experiments on large-scale autonomous driving datasets demonstrate that our method significantly improves perception accuracy, particularly under challenging conditions. Visualization results further confirm that our approach produces more accurate, regular, and coherent predictions of road elements.

Yining Shi, Kun Jiang, Xin Zhao et al. · tsinghua

LiDAR-based 3D object detection is a fundamental task in the field of autonomous driving. This paper explores the unique advantage of Frequency Modulated Continuous Wave (FMCW) LiDAR in autonomous perception. Given a single frame FMCW point cloud with radial velocity measurements, we expect that our object detector can detect the short-term future locations of objects using only the current frame sensor data and demonstrate a fast ability to respond to intermediate danger. To achieve this, we extend the standard object detection task to a novel task named predictive object detection (POD), which aims to predict the short-term future location and dimensions of objects based solely on current observations. Typically, a motion prediction task requires historical sensor information to process the temporal contexts of each object, while our detector's avoidance of multi-frame historical information enables a much faster response time to potential dangers. The core advantage of FMCW LiDAR lies in the radial velocity associated with every reflected point. We propose a novel POD framework, the core idea of which is to generate a virtual future point using a ray casting mechanism, create virtual two-frame point clouds with the current and virtual future frames, and encode these two-frame voxel features with a sparse 4D encoder. Subsequently, the 4D voxel features are separated by temporal indices and remapped into two Bird's Eye View (BEV) features: one decoded for standard current frame object detection and the other for future predictive object detection. Extensive experiments on our in-house dataset demonstrate the state-of-the-art standard and predictive detection performance of the proposed POD framework.

Yining Shi, Kun Jiang, Jinyu Miao et al.

3D occupancy prediction (3DOcc) is a rapidly rising and challenging perception task in the field of autonomous driving. Existing 3D occupancy networks (OccNets) are both computationally heavy and label-hungry. In terms of model complexity, OccNets are commonly composed of heavy Conv3D modules or transformers at the voxel level. Moreover, OccNets are supervised with expensive large-scale dense voxel labels. Model and data inefficiencies, caused by excessive network parameters and label annotation requirements, severely hinder the onboard deployment of OccNets. This paper proposes an EFFicient Occupancy learning framework, EFFOcc, that targets minimal network complexity and label requirements while achieving state-of-the-art accuracy. We first propose an efficient fusion-based OccNet that only uses simple 2D operators and improves accuracy to the state-of-the-art on three large-scale benchmarks: Occ3D-nuScenes, Occ3D-Waymo, and OpenOccupancy-nuScenes. On the Occ3D-nuScenes benchmark, the fusion-based model with ResNet-18 as the image backbone has 21.35M parameters and achieves 51.49 in terms of mean Intersection over Union (mIoU). Furthermore, we propose a multi-stage occupancy-oriented distillation to efficiently transfer knowledge to vision-only OccNet. Extensive experiments on occupancy benchmarks show state-of-the-art precision for both fusion-based and vision-based OccNets. For the demonstration of learning with limited labels, we achieve 94.38\% of the performance (mIoU = 28.38) of a 100\% labeled vision OccNet (mIoU = 30.07) using the same OccNet trained with only 40\% labeled sequences and distillation from the fusion-based OccNet.

Jinyu Miao, Kun Jiang, Yunlong Wang et al.

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.

Haibao Yu, Tuopu Wen, Guangliang Cheng et al.

Low-bit quantization is challenging to maintain high performance with limited model capacity (e.g., 4-bit for both weights and activations). Naturally, the distribution of both weights and activations in deep neural network are Gaussian-like. Nevertheless, due to the limited bitwidth of low-bit model, uniform-like distributed weights and activations have been proved to be more friendly to quantization while preserving accuracy~\cite{Han2015Learning}. Motivated by this, we propose Scale-Clip, a Distribution Reshaping technique that can reshape weights or activations into a uniform-like distribution in a dynamic manner. Furthermore, to increase the model capability for a low-bit model, a novel Group-based Quantization algorithm is proposed to split the filters into several groups. Different groups can learn different quantization parameters, which can be elegantly merged in to batch normalization layer without extra computational cost in the inference stage. Finally, we integrate Scale-Clip technique with Group-based Quantization algorithm and propose the Group-based Distribution Reshaping Quantization (GDQR) framework to further improve the quantization performance. Experiments on various networks (e.g. VGGNet and ResNet) and vision tasks (e.g. classification, detection and segmentation) demonstrate that our framework achieves good performance.

Zhongyang Xiao, Kun Jiang, Shichao Xie et al.

High precision localization is a crucial requirement for the autonomous driving system. Traditional positioning methods have some limitations in providing stable and accurate vehicle poses, especially in an urban environment. Herein, we propose a novel self-localizing method using a monocular camera and a 3D compact semantic map. Pre-collected information of the road landmarks is stored in a self-defined map with a minimal amount of data. We recognize landmarks using a deep neural network, followed with a geometric feature extraction process which promotes the measurement accuracy. The vehicle location and posture are estimated by minimizing a self-defined re-projection residual error to evaluate the map-to-image registration, together with a robust association method. We validate the effectiveness of our approach by applying this method to localize a vehicle in an open dataset, achieving the RMS accuracy of 0.345 meter with reduced sensor setup and map storage compared to the state of art approaches. We also evaluate some key steps and discuss the contribution of the subsystems.