🚗 自动驾驶技术进展

Leaders from Waabi, Shield AI, and General Motors join the Real World AI Stage at TechCrunch Disrupt 2026 to talk building AI. Save up to $200 by Sept. 25, 11:59 p.m. PT. Get a second pass at 50% off.

Waabi、Shield AI 和通用汽车的负责人将登上 TechCrunch Disrupt 2026 的 Real World AI 舞台，探讨 AI 构建。9 月 25 日前购票最高省 200 美元，第二张票半价

We present S2Planner, a trajectory planner that combines three front-facing cameras with ego-motion history and the current driving command. A fine-tuned DINOv3 backbone and a Spatial Tuning Adapter p...

S2Planner 是一种轨迹规划器，结合三个前向摄像头、自车运动历史和当前驾驶指令进行规划。模型采用微调后的 DINOv3 backbone（自监督视觉基础模型）并引入空间调节适配器（Spatial Tuning Adapter）模块

Camera-LiDAR fusion has become a prevailing paradigm for 3D object detection in autonomous driving. However, existing fusion detectors often establish strong inter-modality dependencies by decoding ob...

相机-激光雷达融合已成为自动驾驶 3D 目标检测的主流范式。但现有融合检测器往往通过解码目标特征建立强模态间依赖，导致某一传感器退化时性能大幅下降。SARFusion 提出场景感知路由机制，根据各模态可靠性动态调整融合权重

Accurate digital maps are essential for Advanced Driver Assistance Systems (ADAS) or Autonomous Driving (AD), providing critical information such as road geometry, traffic signs and speed limits requi...

高精数字地图对 ADAS（高级驾驶辅助系统）和自动驾驶至关重要，提供道路几何、交通标志、限速等关键信息。本研究提出一种相异性度量方法，评估众包建图中车队规模对地图质量与更新覆盖率的影响

Federated learning (FL) can improve world model (WM)-based autonomous driving (AD) without centralizing raw private vehicle data, but it also turns model aggregation into a safety-critical integrity b...

联邦学习（FL，即各车辆本地训练、仅上传模型参数的分布式训练方式）可以在不集中原始车辆隐私数据的情况下改进基于世界模型（WM）的自动驾驶，但也使模型聚合成为安全关键环节。FedWM-Guard 旨在检测并阻止恶意车辆上传被投毒的世界模型参数

Automated vehicles must be able to interact with pedestrians safely and efficiently across diverse traffic situations. Although driving simulators offer a scalable testbed for learning such capabiliti...

自动驾驶车辆必须能在多样交通情境中与行人安全高效交互。虽然驾驶仿真器为此类能力提供了可扩展的测试环境……

While end-to-end autonomous driving systems show promise, their application to micromobility vehicles is hindered by simulators failing to capture specific kinematics, such as differential drives and ...

尽管端到端自动驾驶系统前景可观，但将其应用于微出行车辆（如电动滑板车、平衡车等）受限于现有仿真器无法准确模拟其特殊运动学特性（如差速驱动、欠驱动动力学等）。本工作构建了支持这些运动学建模的仿真学习环境

The Amazon-owned company launched an investigation after a worker submitted an OSHA complaint.

这家亚马逊旗下公司在一名工人向 OSHA（职业安全与健康管理局）投诉后启动了调查，并暂停了亚特兰大的测试车队

Vision-language models (VLMs) offer a promising approach to long-tail autonomous driving, but existing driving datasets provide limited supervision for connecting decision-critical visual evidence wit...

视觉语言模型为长尾自动驾驶提供了有前景的方案，但现有驾驶数据集对连接决策关键视觉证据与推理的监督十分有限

With the development of applications like autonomous driving, object detection has gained significant attention, while also highlighting critical vulnerabilities like backdoor attacks that severely co...

随着自动驾驶等应用的发展，目标检测备受关注，但也暴露出后门攻击等严重威胁安全的关键漏洞

Feds know of 5 crashes where cars using comma.ai devices hit slow or stopped vehicles.

联邦监管机构已知 5 起使用 comma.ai 设备的车辆撞击缓慢或静止车辆的事故

Vision-language-action models (VLA) promise human-steerable autonomous driving, but their training is bottlenecked by the scarcity of frames paired with natural-language instructions: while camera str...

视觉-语言-动作模型（VLA）承诺人类可干预的自动驾驶，但其训练受限于配对自然语言指令帧的稀缺：虽然相机数据流庞大……

Continuous Vision (CV) systems underpin real-time applications such as autonomous driving and augmented reality, where latency, throughput, and energy are tightly constrained on mobile platforms. Mode...

持续视觉系统支撑着自动驾驶和增强现实等实时应用，这些应用在移动平台上对延迟、吞吐和能耗有严格约束。而传统模型……

Automotive spinning FMCW radar provides dense, $360^\circ$ sensing and remains reliable under poor illumination and adverse weather, making it well-suited to autonomous navigation. Place recognition u...

车载旋转 FMCW 雷达提供密集的 360 度感知，在弱光和恶劣天气下依然可靠，适合自主导航。基于此的位置识别研究……

Reliable localization of non-line-of-sight (NLOS) pedestrians is critical for safe urban autonomous driving, yet it remains highly challenging in ego-dynamic outdoor environments, where ego-vehicle mo...

非视距（NLOS）行人的可靠定位对城市自动驾驶安全至关重要，但在自车运动的动态户外环境中仍极具挑战

Validating Autonomous Driving Systems (ADS) in simulation requires testing architectures that can discover rare, safety-critical failures while generating scenarios that are executable, diverse, and u...

在仿真中验证自动驾驶系统需要能发现罕见、安全攸关失败的测试架构，同时生成的场景需可执行、多样且有用

Most autonomous-driving datasets record only the action executed by a behavior policy and the single future that followed, providing limited supervision for comparing alternative ego decisions. We int...

多数自动驾驶数据集只记录行为策略实际执行的动作及其后发生的单一未来，为比较不同自车决策提供的监督有限。本文引入分支结构数据集

Faithfully evaluating end-to-end driving policies in simulation requires observations that are not merely photo-realistic, but preserve the scene features a policy relies on to make decisions. Existin...

在仿真中忠实评估端到端驾驶策略，需要观测不仅逼真，还要保留策略做决策所依赖的场景特征。现有方法难以兼顾

End-to-end (E2E) driving policies have progressed rapidly on open-loop (OL) benchmarks, yet OL evaluation cannot reveal whether a policy withstands compounding errors, recovers from failures, or inter...

端到端驾驶策略在开环基准上进展迅速，但开环评估无法揭示策略是否能承受误差累积、从失败中恢复

Mercedes-Benz has signed a definitive production agreement with British self-driving startup Wayve to build its “AI Driver” into future Mercedes vehicles, with integration starting within the next two...

梅赛德斯-奔驰与英国自动驾驶初创公司 Wayve 签署正式量产协议，将在未来梅赛德斯车型中搭载其“AI Driver”端到端自动驾驶系统，两年内开始集成

This paper introduces Dr-LiSA, a first-of-its-kind direct method for localizing 2D spinning radar intensity measurements in $SE(3)$ against 3D lidar maps. Radar-lidar localization combines the complem...

本文提出 Dr-LiSA，首创将 2D 旋转雷达强度测量直接在 SE(3) 上与 3D 激光雷达地图配准的定位方法。雷达-激光雷达定位结合两者互补优势

Camera-based 3D object detection and online vectorized HD mapping provide compact scene representations for autonomous driving, but both depend on accurate metric geometry and remain limited by depth ...

基于相机的 3D 检测和在线矢量化高精地图为自动驾驶提供紧凑场景表示，但都依赖精确度量几何并受深度估计误差限制，本文引入视觉点云地图先验来弥补

Existing latent world models are typically optimized for future predictability, yet the resulting representations are not necessarily useful for planning in autonomous driving. Predictions are commonl...

现有潜在世界模型通常针对未来可预测性优化，但得到的表示未必对自动驾驶规划有用。预测常见地准确拟合了与决策无关的动态

Existing latent world models are typically optimized for future predictability, yet the resulting representations are not necessarily useful for planning in autonomous driving. Predictions are commonl...

现有潜在世界模型通常针对未来可预测性优化，但得到的表示未必对自动驾驶规划有用。预测常见地准确拟合了与决策无关的动态

The field of Artificial Intelligence has been adopted for many application domains. Vision Language Models are one of the recently advanced AI techniques that have been explored to support automotive ...

人工智能已被应用于众多领域。视觉语言模型（VLM）是近期被探索用于支持汽车感知的先进 AI 技术之一

When we evaluate the performance of our odometry, it is common practice to score the estimated track against a ground truth. Unfortunately, scoring uses point metrics, such as the root mean square err...

评估里程计性能时，常见做法是将估计轨迹与真值对比打分。但遗憾的是，这种打分通常使用点级指标，如均方根误差（RMSE）等，可能无法全面反映轨迹质量

Sparse instance representations provide a compact interface for spatial LiDAR-camera and temporal past-current interaction in multimodal perception and E2EAD. Effective interaction requires reliable i...

稀疏实例表示为多模态感知和端到端自动驾驶（E2EAD）中的 LiDAR-相机空间交互与时序交互提供了紧凑接口。有效交互需要可靠的实例匹配

Recent VLM-based autonomous driving planners adopt GRPO-style reinforcement learning to optimize driving performance. However, existing GRPO recipes either optimize driving efficiency, risking progres...

近期基于视觉语言模型（VLM）的自动驾驶规划器采用 GRPO 风格的强化学习来优化驾驶性能。但现有 GRPO 方案要么优化行驶效率、 risking 进展受限……

This short report documents our entry to the COMFORT Localization Benchmark (IROS 2026), evaluated on the GrandTour dataset recorded with the Boxi payload. It extends MOLA-LO into a LiDAR-inertial sys...

这份短报告记录了其参加 IROS 2026 COMFORT 定位基准的方案，在 Boxi 载荷采集的 GrandTour 数据集上评估，将 MOLA-LO 扩展为激光雷达-惯性系统

Standard deployment-ready object detectors for autonomous vehicles degrade in adverse weather and lighting conditions without being trained on extensive domain-specific data. While large-scale vision ...

面向量产的标准目标检测器在恶劣天气和光照条件下性能退化，除非用大量领域特定数据训练。而大规模视觉基础模型可以用于自动标注来弥补

LiDAR point clouds acquired in underground environments exhibit severe geometric incompleteness due to occlusions and limited sensor viewpoints, making reliable point cloud completion challenging with...

地下环境中采集的 LiDAR 点云因遮挡和传感器视角受限存在严重几何不完整，ZVeC 提出零样本框架做实例级车辆提取和点云补全

Autonomous parking in nonconvex and narrow environments remains challenging. Although optimal-control methods can explicitly enforce vehicle dynamics and collision constraints, nonconvexity compromise...

非凸狭窄环境下的自动泊车仍有挑战。最优控制方法能显式满足车辆动力学和碰撞约束，但非凸性损害求解质量，本文用大语言模型引导语义决策辅助最优控制求解

Latent world models learn predictive representations for autonomous driving, but the relational semantics these states preserve often remain implicit. We investigate whether traffic scene graphs can s...

潜在世界模型为自动驾驶学习预测性表示，但其保留的关系语义往往隐式存在。本文研究交通场景图能否作为显式的结构化先验来锚定潜在状态的关系语义

Intelligent transportation systems require Incremental Learning (IL) to continually improve their overall performance in dynamic environments. However, most edge devices lack the computational resourc...

智能交通系统需要增量学习（IL）在动态环境中持续提升性能，但多数边缘设备算力有限。MECAIL 用仅 14.6 KB 的时空专家模块实现通信感知的增量目标检测

Recent 4D LiDAR language models aim to reason about objects and their evolving spatial relationships. Yet, in our evaluation, always selecting the same option nearly matches the multiple-choice accura...

近期的 4D LiDAR 语言模型旨在推理目标及其演变的空间关系，但评测发现总是选同一个选项几乎能达到多项选择题的准确率，说明这些模型可能并未真正理解时空关系

Pedestrian head orientation recognition plays an important role in autonomous driving by providing valuable cues for understanding pedestrian attention and anticipating potential crossing behavior. Ho...

行人头部朝向识别能提供行人注意力和过街意图的重要线索，对自动驾驶交互安全很重要，本文提出轻量级网络方案

This paper evaluates Nash- and Stackelberg-based decision-making controllers for autonomous intersection crossing using a three-stage evaluation pipeline culminating in physical Quanser QCar 2 experim...

本文评测基于纳什均衡和斯塔克尔伯格（Stackelberg）博弈的自动驾驶无保护路口通行决策控制器，采用三阶段评测管线，最终在物理 Quanser QCar 2 实验车上验证

Multimodal 3D object detection is fundamental to robust perception in autonomous driving because it integrates complementary information from LiDAR and camera sensors. However, existing methods often ...

多模态 3D 检测融合 LiDAR 与相机互补信息，是自动驾驶鲁棒感知的基础，但现有方法常在单传感器退化时性能大幅下降，本文利用视觉基础模型提升鲁棒性

End-to-end autonomous driving maps current observations directly to future trajectories, yet those trajectories must remain valid as the scene evolves. Future state modeling aims to address this tempo...

端到端驾驶将观测直接映射为未来轨迹，但轨迹需随场景演化保持有效。FeasibleFlow 用一步流（Flow）模型联合传输配置可行性与轨迹，解决时序一致性问题

End-to-end driving requires planning-relevant bird's-eye-view (BEV) representations, but existing pretraining approaches often rely on task annotations or dense scene reconstruction. We introduce PAVE...

端到端驾驶需要与规划相关的鸟瞰图（BEV）表示，但现有预训练方法依赖任务标注或稠密场景重建。PAVE 引入稀疏动作条件目标，让 BEV 预训练直接对齐下游规划需求

Dynamic object segmentation and ego-motion estimation are closely coupled problems in autonomous driving, as accurate ego-motion estimation typically requires static scene observations, while identify...

动态目标分割与自车运动估计是强耦合问题：准确的自车运动估计需要静态场景观测，而识别动态目标又依赖运动估计。本文用雷达-图像融合联合求解

Image-to-LiDAR registration estimates the camera pose of an image with respect to a LiDAR point cloud. It has diverse applications in autonomous driving, robot navigation etc. However, state-of-the-ar...

图像到 LiDAR 配准估计图像相对于点云的相机位姿，在自动驾驶、机器人导航等场景有广泛应用。但现有最优方法依赖特定场景训练，ZIL 实现零样本泛化

Traffic sign recognition (TSR) is an important perception task for autonomous driving and advanced driver-assistance systems, where a system must both localize traffic signs and determine their semant...

交通标志识别（TSR）是自动驾驶和高级驾驶辅助系统（ADAS）的重要感知任务，需要同时定位交通标志并确定其语义类别，本文用分支 YOLOv2 结构结合几何特征提升识别

Current Vision-Language-Action (VLA) models for autonomous driving operate primarily through feedforward inference across the perception--reasoning--planning hierarchy. While modern architectures main...

当前自动驾驶的视觉-语言-动作（VLA）模型主要采用感知-推理-规划的前馈推理，本文引入情境记忆嵌入提供感知反馈，形成更接近闭环的推理结构

3D object detection from LiDAR point clouds faces a fundamental dilemma: voxel-based methods achieve efficiency at the cost of geometric quantization, while point-based methods preserve fidelity but s...

LiDAR 点云 3D 检测存在两难：体素化方法高效但损失几何精度，点级方法保真但计算量大。PointLAM 用 Mamba（一种线性复杂度的状态空间模型）做局部注意力实现两者兼顾

Fully autonomous tractor--trailer systems are increasingly deployed in logistics, agriculture, and industrial environments, where precise and robust path-tracking capabilities are essential. However, ...

全自动驾驶卡车-挂车系统在物流、农业和工业环境中部署日益增多，需要精确鲁棒的路径跟踪能力，本文提出针对铰接车辆的路径跟踪算法

Generative world models offer controllable and repeatable closed-loop simulation for end-to-end and vision-language-action driving policies, but production deployment exposes three unresolved requirem...

生成式世界模型为端到端和视觉-语言-动作（VLA）驾驶策略提供可控可复现的闭环仿真，但量产部署暴露出三大未解决需求，ZYT-World 针对实时性和可控性给出方案

Generative world models offer controllable and repeatable closed-loop simulation for end-to-end and vision-language-action driving policies, but production deployment exposes three unresolved requirem...

生成式世界模型为端到端和视觉-语言-动作驾驶策略提供了可控可复现的闭环仿真，但量产部署暴露出三个未解决的需求

Self-play in high-throughput simulators yields driving policies with robust closed-loop performance, but improvement per unit of simulation diminishes as training scales. Policies learn to handle comm...

在高吞吐量模拟器中的自博弈（self-play，即策略与自身历史版本对弈训练）能产生鲁棒的闭环驾驶策略，但随着训练规模扩大单位模拟量的收益递减，本文通过反事实后训练提升样本效率

Multimodal trajectory prediction improves behavioral coverage in end-to-end autonomous driving, but existing methods remain limited by sparse scene representations. Incomplete evidence leads to low-qu...

多模态轨迹预测提升了端到端驾驶的行为覆盖度，但现有方法受限于稀疏场景表示，证据不完整导致低质量轨迹。本文提出基于寄存器（register）的稠密表示并结合风险推理

Sparse representation formulates the environment perception for the end-to-end driving system as a set of discrete elements like objects and lane lines. This formulation meets safety risks in crowded,...

稀疏表示（以目标物、车道线等离散元素描述环境）在拥挤场景下面临安全风险，本文提出风险感知的占用栅格（Occupancy）表示方法来提升端到端驾驶系统的安全性

Conventional end-to-end driving systems model the environment with sparse objects and lane elements. While efficient, this paradigm discards planning-critical information in crowded and occluded scena...

传统端到端驾驶系统用稀疏目标物和车道元素建模环境。虽然高效，但这种范式在拥挤和遮挡场景中丢弃了规划关键信息

End-to-End autonomous driving models commonly predict future waypoints from sensor inputs and convert them into vehicle control commands through a downstream controller. However, conventional waypoint...

端到端驾驶模型通常从传感器输入预测未来航点再转为控制指令，但常规航点预测的输出轨迹不稳定。本文用交互多模型（IMM）滤波构造平滑教师信号来稳定训练

Cooperative and connected automated vehicles (CAVs) rely on Signal Phase and Timing (SPaT) messages to cross signalized intersections; a denial-of-service (DoS) flood that blocks SPaT forces CAVs into...

协作式互联自动驾驶车辆（CAV）依赖信号相位与配时（SPaT）消息通过信号灯路口；阻断 SPaT 的拒绝服务（DoS）洪泛攻击会迫使 CAV 降级运行，本文研究不同气质的人类驾驶员如何影响攻击后果

Recent advances in large language models have improved their effectiveness as back-end components for voice assistants, particularly in intent understanding and context-aware input classification. How...

大语言模型提升了语音助手在意图理解和上下文分类上的后端能力，本文为自动驾驶赛车构建可在边缘设备部署的开源语音助手框架

The company suspended service in the city in April after one of its robotaxis got swept away.

今年 4 月该公司一辆 Robotaxi 被洪水冲走后，Waymo 暂停了在该市的服务，现已重启。

As scaling pre-training data alone yields diminishing returns, post-training is becoming increasingly important across physical AI domains such as autonomous driving. End-to-end driving policies are p...

随着仅扩大预训练数据收益递减，后训练在自动驾驶等物理 AI 领域愈发重要。OPTED 用免渲染的教师模型对端到端驾驶策略做在线策略（on-policy）微调。

Reinforcement learning constitutes a promising approach owing to its potential for superhuman performance and self-learned policies. However, its application to real-world autonomous driving remains s...

强化学习（RL）因具备超人性能和自学习策略的潜力而受关注，但用于真实世界自动驾驶仍面临困难。MILER 通过语义中间层表征改善非结构化场景下 RL 的仿真到现实迁移。

Connected vehicles use millimeter-wave (mmWave) sidelinks for the data rates cooperative driving demands, and emerging designs select the carrier band from sensed rainfall. We show that this weather a...

网联车辆使用毫米波（mmWave）直连链路满足协同驾驶所需的数据速率，新兴设计会根据感知到的降雨选择载波频段。本文证明这种天气自适应机制可被虚假天气数据欺骗。

Connected-vehicle safety evaluations rely on coupled traffic and network simulations, but standard channel models ignore radio resource competition in 5G NR sidelink Mode-2, reporting unrealistically ...

网联车辆安全评估依赖交通与网络联合仿真，但标准信道模型忽略了 5G NR 直连链路 Mode-2 下的无线资源竞争，报告出不切实际的性能。NS3Learn 将 ns-3 的真实接收模型迁移到 Veins/SUMO 联合仿真栈中。

Occlusion creates fundamental uncertainty in autonomous driving. Existing methods often propagate frame-wise hypotheses or optimize ego behavior against prescribed hidden-agent predictions, leaving th...

遮挡给自动驾驶带来根本性的不确定性。现有方法多逐帧传播假设或按预设的隐车预测优化自车行为，本文直接在时空遮挡区域内搜索最坏情况的隐藏车辆轨迹。

Autonomous driving involves coupled decision-making and scene evolution under multi-mode uncertainty. To capture this coupling and uncertainty, we introduce MM-Future, a world-action model that genera...

自动驾驶涉及多模式不确定性下的决策与场景演化耦合。MM-Future 是一个世界-动作联合模型，同时生成多模式的未来场景和自车动作，以捕捉这种耦合与不确定性。

Lidl has started using a cab-less autonomous electric truck for regular store deliveries in Germany, assigning it a daily replenishment route between its Edermünde distribution centre and a nearby sto...

Lidl 开始使用一辆无驾驶室的自动驾驶电动卡车为德国门店做常规配送，每天在 Edermünde 配送中心与附近门店之间跑补货路线。

Accurate simulation is crucial for autonomous driving development, yet capturing real-world traffic complexity remains challenging. Existing simulators that rely on predefined rules or static data pla...

准确的仿真对自动驾驶研发至关重要，但捕捉真实交通复杂性仍然困难。依赖预定义规则或静态数据的现有仿真器表现不佳，REARL 利用真实交通数据和 LLM 增强闭环仿真的真实度。

Predicting the future trajectories of surrounding vehicles in autonomous driving is important for collision risk assessment and safe ego-vehicle path planning. Conventional neural network-based trajec...

预测周围车辆的未来轨迹对碰撞风险评估和自车路径规划很重要。传统神经网络预测缺乏物理一致性，本文将多个扩展卡尔曼滤波器（EKF）产生的轨迹候选用神经网络融合。

Modern vehicular embedded systems increasingly combine ROS 2-based autonomous-driving stacks with AUTOSAR Adaptive Platform (AUTOSAR AP). Such mixed stacks make scenario-based evaluation hard to inter...

现代车载嵌入式系统越来越多地将基于 ROS 2 的自动驾驶栈与 AUTOSAR 自适应平台（AUTOSAR AP）混合使用，这种混合架构让场景化评估的追溯与解释变得困难，BA-TRACE 通过边界感知的轨迹重构来解决。

Cooperative autonomous driving in the IoT-to-Edge-to-Cloud continuum requires system-level validation across vehicles, infrastructure sensors, edge-side Dynamic Map services, and in-vehicle autonomous...

物联网-边缘-云协同的协作式自动驾驶需要在车辆、路侧传感器、边缘侧动态地图服务和车载自动驾驶系统之间进行系统级验证，VAST 提供了覆盖全链路的验证工具链。

Work zones alter lane geometry through temporary traffic controls and closures that may be absent from on-board maps, challenging autonomous vehicle (AV) perception and planning. Generalization is als...

施工区域通过临时交通管制和封路改变车道几何，这些变化往往不在车载地图中，给自动驾驶感知和规划带来挑战。WZPlanner 专注施工区域的端到端安全路径规划。

Three-dimensional object detection for autonomous driving is dominated by detectors trained on large corpora of human-annotated 3D boxes. Such a detector learns a fixed category list, and everything o...

自动驾驶 3D 目标检测目前由基于大规模人工标注 3D 框训练的检测器主导，这类检测器只能识别固定类别列表，本文结合可提示分割实现开放词汇检测，突破固定类别的限制。

The deal could net the asset-light robotaxi company more than $300 million in funding.

这笔交易可能为这家轻资产运营的 Robotaxi 公司带来超过 3 亿美元融资。SPAC 即特殊目的收购公司，是绕过传统 IPO 的快捷上市方式。

We propose SEAM, a LiDAR-based lifelong mapping framework. Instead of relying on a single anchor spanning the entire session, SEAM generates evidence based on a trajectory optimized with submap-level ...

SEAM 是一个基于 LiDAR 的终身建图框架。它不依赖贯穿整个会话的单一锚点，而是基于子图级优化后的轨迹生成证据，解决长期建图中的轨迹形变问题。

In interactions with uncertain opponents, dual model predictive control (MPC) can improve performance through information-seeking actions that reduce uncertainty about opponents' behavior. Its recent ...

在与行为不确定的对手交互时，对偶模型预测控制（MPC）可通过主动探索信息来降低对对手行为的不确定度。本文用图神经网络（GNN）加速混合整数对偶 MPC 的求解。

State-of-the-art vision-language-action models (VLA) for autonomous driving face critical limitations: excessive parameter counts, inefficient high-resolution image processing, and lack of temporal me...

当前最先进的视觉-语言-动作模型（VLA）在自动驾驶中存在参数量过大、高分辨率图像处理低效、缺乏时序记忆等关键局限。FIVE-VLA 引入循环动作记忆来提升速度和效率。

Research on 4D millimeter-wave radar perception algorithms has flourished in recent years, extending from signal processing and object detection to semantic segmentation, motion estimation, occupancy ...

近年来 4D 毫米波雷达感知算法研究蓬勃发展，从信号处理和目标检测扩展到语义分割、运动估计、占据栅格预测等方向，本文对相关算法进行了系统综述。

Track 2 of the AI City Challenge 2026 requires both visual question answering (VQA) and traffic event description generation under a challenging synthetic-to real domain shift. Existing vision-languag...

AI City Challenge 2026 第二赛道要求在合成到真实（sim-to-real）的域偏移下完成视觉问答和交通事件描述生成。本文用 V-JEPA 视频表征将语义理解与文本生成分离以提升泛化能力。

4D radar has emerged as a promising next-generation sensor for improving the robustness of autonomous driving perception systems because of its stable sensing capability under adverse weather conditio...

4D 毫米波雷达因在恶劣天气下感知稳定，成为提升自动驾驶感知鲁棒性的下一代传感器。本文研究兼顾检测精度和实时性的雷达点云预处理方法。

In autonomous driving, perception models often struggle to generalize to new environments due to domain shifts. While unsupervised model adaptation offers a feasible solution without labor-intensive m...

在自动驾驶中，感知模型常因域偏移（domain shift，即训练环境与部署环境差异）难以泛化到新环境。无监督模型自适应提供了一种无需大量人工标注的可行方案，本文利用多车协同感知来辅助模型适应。

Safe motion planning in automated driving requires anticipating evolving traffic risks and deciding when to revise the current planned trajectory. We introduce RiskWorld, a risk-aware world modeling f...

自动驾驶的安全运动规划需要预判不断演变的交通风险并决定何时修改当前轨迹。RiskWorld 是一个风险感知世界建模框架，基于流引导的占据栅格演化来估计风险。

Lie down in a car, and you’re more likely to get seriously hurt in a crash. But a driverless future is forcing governments to revisit long-standing safety issues.

如果你躺卧在车内，碰撞中受重伤的概率更高。而无人驾驶的未来正在迫使各国政府重新审视长期沿用的汽车安全标准。