自动驾驶之心行业交流群来了，关注L4赛道融资、技术进展、智驾落地、行业动态等方向~ 添加小助理微信AIDriver005，备注：昵称+机构/学校+进群。 ...

点击下方 卡片 ，关注" 自动驾驶之心 "公众号 戳我-> 领取 自动驾驶近30个 方向 学习 路线 最近一个月，星球内又更新了很多最新的技术&行业动态： 同时，还有很多的答疑解惑： 下周四柱哥也会邀请嘉宾在星球内部和大家聊一聊最近的一些技术进展，欢迎大家加入自动驾驶之心知识星 球。我们准备了大额新人优惠...... 扛内卷，一个足够有料的社区 对于很多想入门的同学来说，试错成本有点高。没时间和缺乏完整的体系是最大问题，这也容易导致行业壁 垒越来越高，如果想要卷赢那就更加困难了。 Waymo最新的基座模型分享，快慢双系统+数据飞轮； WorldSplat二十问，前馈GS趋势已成； 自动驾驶L4圆桌访谈，港大团队讲解特斯拉端到端三大挑战； 2025地平线技术生态大会上，苏箐关于自驾的一些insights； 自动驾驶世界模型论文与代码汇总； 英伟达2025年技术图鉴，自驾、具身、大模型全面开花； 传统规控实现给端到端大模型兜底的思路； 40个问题深度解析自动驾驶领域vla+wm的重磅工作：DriveVLA-W0； BEV融合如何能够提升盲区（很近范围）内3D Box的边界准确程度； 小鹏第二代VLA的延展讨论； ...

Core Insights - Mobileye (MBLY) is positioned as a significant player in the autonomous driving sector, with strong growth potential due to its technology being utilized by leading OEMs for Advanced Driver Assistance Systems (ADAS) [1] Company Overview - Mobileye's technology is integral to the development of ADAS, indicating its critical role in the future of autonomous driving [1] Market Position - The firm is recognized for its impressive business model and solid potential in the rapidly evolving autonomous driving market [1]

Core Insights - WeRide has launched its Robotaxi service Mini Program "WeRide Go" on WeChat, enhancing accessibility for users in China [1][3] - The integration with WeChat allows users to book Robotaxi rides without needing a separate app, streamlining the user experience [2][3] - WeRide aims to expand its Robotaxi fleet to tens of thousands by 2030, leveraging WeChat's extensive user base to boost ride volume and user retention [4] Company Overview - WeRide is a leader in the autonomous driving sector, operating over 1,000 Robotaxis globally, with fully driverless operations in major cities like Guangzhou and Beijing [4][5] - The company has received autonomous driving permits in eight markets, including China, the UAE, and the US, showcasing its regulatory compliance and market reach [5] - WeRide's technology platform, WeRide One, supports a range of autonomous driving products and services, addressing various transportation needs [5]

Core Insights - WeRide has launched its Robotaxi service Mini Program "WeRide Go" on WeChat, enhancing accessibility for users in China [1][3] - The integration with WeChat allows users to book Robotaxi rides without needing a separate app, streamlining the user experience [2][3] - WeRide aims to expand its Robotaxi fleet to tens of thousands by 2030, leveraging WeChat's extensive user base to boost ride volume and user retention [4] Company Overview - WeRide is a leader in the autonomous driving sector, operating over 1,000 Robotaxis globally, with fully driverless operations in major cities like Guangzhou and Beijing [4][5] - The company has received autonomous driving permits in eight markets, including China, the UAE, and the US, showcasing its regulatory compliance and market reach [5] - WeRide's technology platform, WeRide One, supports a range of autonomous driving products and services, addressing various transportation needs [5]

本工作提出了一种新型自动驾驶系统 KnowVal，该系统通过感知模块与知识检索模块的协同作用，实现视觉 - 语言推理能力。 团队构建了涵盖交通法规、防御性驾驶原则与道德考量的综合驾驶知识图谱，并为其开发了高效的基于大型语言模型的检索机制。通过设计集成世界模型与 价值模型的规划器，从而实现价值对齐决策。同时构建了人类偏好数据集用于训练价值模型。 一个智能驾驶系统，在迈向高阶自动驾驶的过程中，应当具备何种能力？除了基础的感知、预测、规划、决策能力，如何对三维空间进行更深入的理解？如 何具备包含法律法规、道德原则、防御性驾驶原则等知识？如何进行基本的视觉 - 语言推理？如何让智能系统具备世界观和价值观？ 来自北京大学王选计算机研究所王勇涛团队的最新工作 KnowVal 给出了一种有效可行的方案。 实验表明，KnowVal 兼容现有的端到端和 VLA 方法，在 nuScenes 数据集上实现了最低碰撞率，并在 Bench2Drive 基准测试中取得了最先进的性能表 现。 论文标题： KnowVal: A Knowledge-Augmented and Value-Guided Autonomous Driving S ...

2026年伊始，百度就给股东们递上了一份迟来的新年礼物。 据百度公告，旗下的AI芯片独角兽昆仑芯，已于元旦期间秘密向香港联交所递交了A1上市申请表。虽 然官方依然保持着"静默期"的克制，但资本市场的嗅觉从来不会出错。 这不仅仅是一个子公司上市的消息，更像是一个信号。 回望过去几年，中国互联网巨头在"分拆"这件事上走得并不顺遂。阿里著名的"1+6+N"变革，最终以阿 里云放弃分拆、菜鸟撤回IPO而由于各种复杂的内外因暂告段落。 当所有人都以为巨头们将回归"大一统"的保守策略时，李彦宏却选择了在这个节点，把百度体系内最硬 核的资产之一推向台前。 在这次分拆上市中，相信大家最关注的，一定是昆仑芯在百度体内和分拆上市后的估值差距。而无论怎 么估算，技术能力不如昆仑芯但市值一度超过百度整体的摩尔线程，给所有的股东耳边都敲响了一声闷 钟。 于是，这不禁让人产生一个大胆的疑问：昆仑芯的独立，究竟是百度的一次孤立行动，还是推倒多米诺 骨牌的第一指？ 如果百度真的决心通过分拆来重估价值，那么在昆仑芯之后，那个烧钱最多、技术积淀最深、商业化故 事最性感的"自动驾驶"业务，会不会是下一个？ 01 被"折叠"的价值，与不得不做的减法 ...

Core Viewpoint - The article discusses the challenges and advancements in reinforcement learning (RL) for autonomous driving, emphasizing the need for a balanced reward system to enhance both safety and efficiency in driving models [2][5]. Group 1: Challenges in Reinforcement Learning - Reinforcement learning faces significant issues such as reward hacking, where increased safety requirements can lead to decreased efficiency, and vice versa [2]. - Achieving a comprehensive performance improvement in RL models is challenging, with many companies not performing adequately [2]. - The complexity of autonomous driving requires adherence to various driving rules, making it essential to optimize through RL, especially in uncertain decision-making scenarios [2][5]. Group 2: Model Development and Talent Landscape - The current industry leaders have developed a complete model iteration approach that includes imitation learning, closed-loop RL, and rule-based planning [5]. - The high barriers to entry in the autonomous driving sector have led to generous salaries, with top talents earning starting salaries of 1 million and above [6]. - There is a notable gap in practical experience among many candidates, as they often lack the system-level experience necessary for real-world applications [7]. Group 3: Course Offerings and Structure - The article promotes a specialized course aimed at practical applications of end-to-end autonomous driving systems, highlighting the need for hands-on experience [8]. - The course covers various chapters, including an overview of end-to-end tasks, two-stage and one-stage algorithm frameworks, and the application of navigation information [13][14][15][16]. - It also addresses the integration of RL algorithms and trajectory optimization, emphasizing the importance of combining imitation learning with RL for better performance [17][18]. Group 4: Practical Experience and Knowledge Requirements - The final chapter of the course focuses on sharing production experiences, analyzing data, models, scenarios, and rules to enhance system capabilities [20]. - The course is designed for advanced learners with a foundational understanding of autonomous driving algorithms, reinforcement learning, and programming skills [21][22].

点击下方 卡片 ，关注" 自动驾驶之心 "公众号 戳我-> 领取 自动驾驶近15个 方向 学习 路线 自动驾驶已经进入下半场，行业的难点和痛点需要更多有志之士参与进来一起突破。后面我们将陆续为 大家增加圆桌访谈、实战&工业级课程、咨询等各类输出。 作为国内自动驾驶领域创作的技术平台，我们期望能够在这波激流中贡献自己的力量，成为一个真的能 给行业带来价值的平台。 众人拾柴火焰高，我们需要更多优秀的伙伴加入我们。 主要方向 包括但不限于：自动驾驶产品经理、4D标注/数据闭环、世界模型、VLA、自动驾驶大模型、强化学 习、端到端等多个方向。 岗位说明 主要面向自动驾驶培训合作（B端主要面向企业和高校、研究院所培训，C端面向较多学生、求职类人 群）、课程开发和原创文章创作。 联系我们 待遇与合作方式，欢迎添加微信wenyirumo做进一步沟通。 ...

点击下方 卡片 ，关注" 自动驾驶之心 "公众号 戳我-> 领取 自动驾驶近30个 方向 学习 路线 >>自动驾驶前沿信息获取 → 自动驾驶之心知识星球 为了解决这一问题，AlignDrive 提出了一种级联框架， 使纵向规划依赖于横向路径 ，从而实现横纵规划的紧密协同。具体而言， 模型先预测横向路径（drive path） ，然后基于动态环境信息预测沿 该路径的逐时刻1D纵向位移 。可以直观地理解为：一个模块负责"转方向盘"，另一个模块负责"踩油门和刹车"。这种设计让不同模块 专注于各自关键信息，尤其是纵向位移的预测，能够建立动态物体与自车行为之间更紧密的关联，使模型更充分地关注动态交互对象，从而提升对动态场景的交互建 模能力。 ★ 视频展示： 一、概述 近年来，端到端自主驾驶技术取得了显著进展，实现了感知和规划的联合处理。在规划阶段，现有的端到端模型通常将规划分解为并行的横向和纵向预测。虽然这种 方法有效，但存在两个主要问题：一是横向路径和速度之间的协调会变得更困难；二是静态信息的冗余编码，导致纵向规划未能充分利用行驶路径作为先验信息。这 些问题限制了模型在复杂场景中的表现。 论文作者 | Yanhao ...