作者 | 历不白@知乎 编辑 | 自动驾驶之心 原文链接： https://www.zhihu.com/question/1899822735284244767/answer/1989321465271706827 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 本文只做学术分享，如有侵权，联系删文 中国市场太卷了， 智驾没有芯片根本没有议价权。 我们回顾历史来说明一下。历史虽然不能说明一切，但是历史却是现实的一面镜子。 在 2004 年至 2010 年间。全球视频监控市场正经历从模拟信号向数字和网络监控。 当时的行业话语权掌握在德州仪器（Texas Instruments, TI）和安霸（Ambarella）等老牌半导体巨头手中 。 TI 作为通用 DSP（数字信号处理）领域的霸主，其方案如经典的 DM365、DM368 系列芯片，本质上是通用的计算引擎。 这意味着下游的安防器材厂商不仅要购买昂贵的芯片，还需要配备庞大的软件团队，在底层的 DSP 上进行极具挑战性的视 频编解码开发和图像算法调优。 对于 ...

Group 1: Automotive Manufacturers - Key manufacturers mentioned include Kia, Geely, Volvo, Jaguar Land Rover, and BYD [3][4][5][11][12] - Notable electric vehicle manufacturers include Rivian, VinFast, and NIO [14][11] - The article highlights the growing presence of Chinese automotive brands in the global market [11][12] Group 2: ADAS and Autonomous Driving Technology - Tier 1 suppliers for Advanced Driver Assistance Systems (ADAS) include Desay SV, Huawei, and Baidu Apollo [16][17] - High-precision positioning suppliers mentioned are Qianxun S, Beidou Star, and Huace Navigation [18][19] - Key players in the LiDAR market include Luminar, Innoviz, and Hesai Technology [21] Group 3: Vehicle Connectivity and Cloud Services - Major cloud service providers for automotive applications include AWS, Microsoft, and Huawei [31][32] - Companies involved in data integration and vehicle-to-everything (V2X) communication include Freetech and MAXIEYE [32][33] Group 4: In-Vehicle Technology and User Experience - Tier 1 suppliers for cockpit technology include Bosch, Aptiv, and Visteon [50][54] - The article discusses the importance of user interface and interaction technologies, highlighting companies like iFlytek and Tencent [61] Group 5: Electric and Hybrid Vehicle Components - Key suppliers for electric vehicle components include Bosch, AUMOVIO, and Magna [26][40] - The article emphasizes the role of battery technology and management systems in the growth of electric vehicles [40][41] Group 6: Commercial Vehicle Technologies - Companies providing ADAS for commercial vehicles include Eastsoft and ZF [48] - The article notes the increasing demand for smart logistics and fleet management solutions [48][49]

以下文章来源于红色星际 ，作者红色星际科技 红色星际 . 让更多人，更深入地了解自动驾驶行业！ 作者 | 红色星际科技 来源 | 红色星际 原文链接： 某头部具身公司创始团队的"裂痕" 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 本文只做学术分享，如有侵权，联系删文 25年对于某头部具身公司来说，可谓是喜忧参半。 一方面，随着资本的追捧，估值不断飙升破百亿，而且订单也不断增长；另一方面，创始人之间也逐渐出现分歧，围绕着是卷量产商业化还是前沿技 术，产生了分歧。 该头部具身公司内部创始人之间逐渐出现了两派：量产派和学术派。 量产派，主要是来自于智驾行业背景的。这部分人相信"沿途下蛋"的发展路径，专注于商业化，聚焦于目前能出货的场景和需求，投入资源把量产交 付做好，然后把出货量做起来。 学术派，主要是来自于高校教师背景的，也就是业界的学术大牛。这部分人带有一定的学术思维，相信应该"直奔珠峰"，喜欢探索具身技术的上限， 认为打造高泛化的模型才是最重要的。 两派人马围绕着该做简单场景还是高难度场景，以及该把资 ...

Core Insights - The article discusses the recent promotion of NIO's NWM 2.0, highlighting its positive reception and the potential of world models in intelligent driving [1] - It emphasizes that the true limit of intelligent driving lies in world models, which utilize video as a core component to understand spatiotemporal and physical laws, enabling machines to comprehend environments like humans do [1] Group 1: World Model Concept - World models address spatiotemporal cognition, while language models focus on conceptual cognition, with the former being more effective in modeling the real world's four-dimensional space-time [1] - The article mentions that many AI giants are developing general world models, including projects like Li Feifei's Marble, Yann LeCun's V-JEPA 2, and DeepMind's Genie 3 [1] Group 2: Challenges in Understanding World Models - The definition of world models remains vague, leading to confusion among newcomers in the field, who often spend significant time navigating challenges without clear guidance [1] - The article notes that understanding world models and completing tasks like data generation and closed-loop simulation can be particularly difficult for beginners [1] Group 3: Course Overview - A course is being offered to help individuals understand the world model domain in autonomous driving, featuring insights from industry algorithm experts [2][6] - The course will cover various aspects of world models, including their historical development, application cases, and different schools of thought within the field [6][10] Group 4: Course Structure - The course consists of six chapters, starting with an introduction to world models and their connection to end-to-end autonomous driving [6] - Subsequent chapters will delve into background knowledge, discussions on general world models, video generation-based models, OCC generation models, and industry applications [6][8][9][10] Group 5: Expected Outcomes - The course aims to equip participants with the skills to reach a level comparable to a world model autonomous driving algorithm engineer within a year [14] - Participants will gain a deeper understanding of key technologies such as BEV perception, multimodal large models, and generative models, enabling them to apply their knowledge in practical projects [14]

Core Viewpoint - The article emphasizes the importance of defining appropriate primary metrics (loss functions) in autonomous driving data loops, arguing that traditional metrics like MPI (Miles Per Intervention) are inadequate for driving problem-solving and system performance improvement [5][10][87]. Group 1: Data Loop and Metrics - The organization should be viewed as a large model where the primary metric acts as the loss function, guiding the optimization process [15][87]. - The common metric MPI is criticized for focusing on how often human intervention is needed rather than the vehicle's performance in avoiding "stupid" or "dangerous" actions [22][80]. - The article introduces two new metrics: MPS (Miles Per Stupid) and MPD (Miles Per Dangerous), which are more aligned with the actual performance of the autonomous system [10][44][80]. Group 2: Limitations of MPI - MPI is defined as total mileage divided by the number of interventions, which can mislead organizations into optimizing for fewer interventions rather than improving vehicle behavior [18][22]. - The timing of interventions often does not correlate with the actual problems occurring, leading to a misalignment in performance metrics [25][26]. - The article highlights that relying on MPI can create negative incentives, encouraging teams to avoid reporting issues rather than addressing them [26][90]. Group 3: MPS and MPD Implementation - MPS focuses on the frequency of "stupid" actions taken by the vehicle, while MPD addresses "dangerous" actions, providing a clearer picture of system performance [44][80]. - The organization can utilize triggers to define and capture these behaviors, allowing for a more precise analysis of performance [47][85]. - The metrics MPS and MPD can be used to drive self-improvement within the organization, ensuring that the focus remains on enhancing vehicle behavior rather than merely reducing human intervention [87][90]. Group 4: Examples and Case Studies - The article provides examples of how MPS and MPD can be applied in real scenarios, such as analyzing sudden braking events and their causes, which can lead to actionable insights for system improvement [49][51][66]. - It discusses the importance of understanding the context behind performance metrics, emphasizing that both improvements and deteriorations in metrics should be investigated thoroughly [59][78]. - The article concludes that effective metrics should not only reflect performance but also guide the organization towards continuous improvement and problem resolution [87][90].

Group 1 - The core viewpoint of the article highlights the competitive landscape of the autonomous driving industry, emphasizing the focus on technology, cost, and efficiency as key areas of competition this year [1] - The industry has seen a shift with many professionals transitioning to sectors like embodied AI and drones, while autonomous driving remains a mature AI field, making algorithm talents highly sought after [1][2] - Major technological directions in autonomous driving have converged this year, including end-to-end systems, VLA, world models, and reinforcement learning, with many midstream companies tackling challenges like OCC and multi-sensor fusion perception [3] Group 2 - The membership of the paid community focused on autonomous driving has officially surpassed 4,000, indicating a growing interest in the development of technology routes and job information [3] - The company expresses gratitude to its supporters and announces various benefits and discounts for the new year, encouraging continued efforts in the upcoming year [4]

Core Insights - The article discusses the development of ColaVLA, a unified visual-language-action framework for autonomous driving that enhances trajectory planning by leveraging cognitive latent reasoning and hierarchical parallel planning [4][10][50]. Group 1: Background and Challenges - Traditional autonomous driving systems separate perception, prediction, and planning into distinct modules, while recent end-to-end (E2E) systems integrate these tasks into a unified learning pipeline [3][6]. - Visual-language models (VLMs) have been increasingly integrated into autonomous driving systems to inject cross-modal prior knowledge and world knowledge, but they face three core challenges: modal mismatch, high latency from autoregressive reasoning, and inefficiencies in planner design [7][9]. Group 2: ColaVLA Framework - ColaVLA proposes a unified framework that shifts the reasoning process from explicit text-based chains to a unified latent variable space, combined with a hierarchical parallel trajectory decoder [10][18]. - The cognitive latent reasoning component efficiently completes scene understanding and decision-making through two forward propagations, extracting decision-relevant information from multimodal inputs [11][21]. - The hierarchical parallel planner generates multi-scale trajectories in a single forward pass, maintaining causal structure and significantly reducing reasoning latency [12][28]. Group 3: Experimental Results - ColaVLA achieved state-of-the-art performance on the nuScenes benchmark, with the lowest average L2 error of 0.30 meters and a collision rate of 0.23%, outperforming existing action-based methods [37][38]. - In closed-loop evaluations, ColaVLA reached a NeuroNCAP score of 3.48, significantly improving safety metrics by reducing average collision rates from 65.1% to 36.8% [39][40]. - The framework demonstrated over five times the reasoning speed compared to traditional text-based autoregressive models, showcasing its efficiency and robustness [40][41].

Core Insights - The article discusses the vision of world models in understanding and transforming the physical world, emphasizing the role of continuous technological breakthroughs in generative AI for autonomous driving [2] - It highlights the ongoing exploration of world models in the autonomous driving sector, particularly in video generation and OCC generation [2][3] - The article addresses the challenges faced by newcomers in grasping the concept of world models and the complexities involved in data generation and closed-loop simulation [4][5] Summary by Sections Introduction to World Models - The first chapter provides an overview of world models and their connection to end-to-end autonomous driving, detailing the historical development and current applications [12] - It categorizes different types of world models, including purely simulated models and those that integrate planning and sensory input generation [12] Background Knowledge - The second chapter covers foundational knowledge related to world models, including scene representation and technologies like Transformer and BEV perception [13] - This chapter is crucial for understanding the technical vocabulary frequently encountered in job interviews related to world models [13] General World Model Discussion - The third chapter focuses on general world models and recent advancements in autonomous driving, discussing notable models such as Marble, Genie 3, and VLA+ algorithms [14] Video Generation-Based World Models - The fourth chapter delves into video generation algorithms, highlighting significant works like GAIA-1 & GAIA-2 and recent advancements in the field [15] OCC-Based World Models - The fifth chapter centers on OCC generation algorithms, discussing three major papers and a practical project that extends to vehicle trajectory planning [16] World Model Job Topics - The sixth chapter shares practical insights from industry experience, addressing the application of world models in the industry, common pain points, and interview preparation [17] Course Overview - The course aims to provide a comprehensive understanding of end-to-end autonomous driving, with a focus on world models, and is designed for individuals looking to enter the autonomous driving industry [17][20] - It includes detailed discussions on key technologies and methodologies, ensuring participants can apply their knowledge in real-world projects [20] Course Schedule - The course is set to begin on January 1, with a duration of approximately two and a half months, featuring offline video lectures and online Q&A sessions [21][22]

Core Viewpoint - The article emphasizes the establishment of a comprehensive community for autonomous driving knowledge, aiming to facilitate learning, sharing, and collaboration among industry professionals and newcomers in the field [22][23]. Group 1: Community and Learning Resources - The "Autonomous Driving Heart Knowledge Planet" has been created to provide a platform for technical exchange, academic discussions, and engineering problem-solving, with members from renowned universities and leading companies in the autonomous driving sector [22][23]. - The community has over 4,000 members and aims to grow to nearly 10,000 in the next two years, offering a rich environment for both beginners and advanced learners [8][10]. - Various learning resources, including video tutorials, articles, and structured learning paths, are available to help members quickly access information and enhance their skills in autonomous driving [10][16]. Group 2: Technical Insights and Developments - Recent updates include insights from industry leaders on topics such as Waymo's latest base model, advancements in self-driving technology, and discussions on data loops and training cycles [7][10]. - The community has compiled over 40 technical routes covering various aspects of autonomous driving, including VLA benchmarks, multi-modal models, and data annotation practices [10][23]. - Members can engage with industry experts to discuss trends, technological advancements, and challenges in mass production of autonomous vehicles [11][26]. Group 3: Job Opportunities and Career Development - The community provides job recommendations and internal referrals to help members connect with potential employers in the autonomous driving industry [16][26]. - Regular discussions on career paths, research directions, and practical applications in the field are facilitated to support members in their professional growth [25][96]. - The platform encourages collaboration and networking among members, fostering a supportive environment for career advancement [20][26].

Core Insights - The article discusses two core papers from Horizon Robotics: DiffusionDrive and ResAD, focusing on their contributions to end-to-end autonomous driving solutions [2][3]. DiffusionDrive - The overall architecture of DiffusionDrive consists of three parts: perception information, navigation information, and trajectory generation [6]. - Perception information includes dynamic/static obstacles, traffic lights, map elements, and drivable areas, emphasizing the need to convey perception tasks to planning tasks in an end-to-end manner [6]. - Navigation information is crucial for avoiding incorrect routes, especially in complex urban environments like Shanghai, where navigation challenges are significant [7]. - The core concept of trajectory generation is "Truncated Diffusion," which leverages fixed patterns in human driving behavior to reduce training convergence difficulty and inference noise [8][10]. - The article outlines a method for trajectory generation using K-Means clustering to describe common human driving behaviors, which simplifies the training process [9]. ResAD - ResAD introduces a residual design that predicts the difference between future trajectories and inertial extrapolated trajectories, rather than generating future trajectories directly [12]. - The residual regularization helps manage the increasing residuals over time, ensuring that the model focuses on the true diversity of driving behaviors [13][14]. - The design allows for different noise perturbations in the trajectory generation process, adjusting learning difficulty based on the direction of motion [15]. - ResAD also features a trajectory ranker that utilizes a transformer model to predict metric scores based on top-k trajectory predictions and environmental information [16]. Conclusion - Both papers from Horizon Robotics provide valuable insights and methodologies for enhancing autonomous driving systems, encouraging further exploration and development in the field [18].