Core Insights - The article discusses the rapid advancements in 3D Generative Synthesis (3DGS) technology, highlighting its applications in various fields such as 3D modeling, virtual reality, and autonomous driving simulation [2][4] - A comprehensive learning roadmap for 3DGS has been developed to assist newcomers in mastering both theoretical and practical aspects of the technology [4][6] Group 1: 3DGS Technology Overview - The core goal of new perspective synthesis in machine vision is to create 3D models from images or videos that can be processed by computers, leading to numerous applications [2] - The evolution of 3DGS technology has seen significant improvements, including static reconstruction (3DGS), dynamic reconstruction (4DGS), and surface reconstruction (2DGS) [4] - The introduction of feed-forward 3DGS has addressed the inefficiencies of per-scene optimization methods, making the technology more accessible and practical [4][14] Group 2: Course Structure and Content - The course titled "3DGS Theory and Algorithm Practical Tutorial" covers detailed explanations of 2DGS, 3DGS, and 4DGS, along with important research topics in the field [6] - The course is structured into six chapters, starting with foundational knowledge in computer graphics and progressing to advanced topics such as feed-forward 3DGS [10][11][14] - Each chapter includes practical assignments and discussions to enhance understanding and application of the concepts learned [10][12][15] Group 3: Target Audience and Prerequisites - The course is designed for individuals with a background in computer graphics, visual reconstruction, and programming, particularly in Python and PyTorch [19] - Participants are expected to have a GPU with a recommended computing power of 4090 or higher to effectively engage with the course material [19] - The course aims to benefit those seeking internships, campus recruitment, or job opportunities in the field of 3DGS [19]

Core Viewpoints - The article presents eight key viewpoints regarding the performance and evaluation of autonomous driving technologies, particularly focusing on the comparison between Horizon's HSD and Li Auto's VLA systems [3]. Group 1: Performance Evaluation - The experience with Horizon's HSD during a 1.5-hour test drive was notably better than the current production version of Li Auto's L7 VLA, although future production versions may not match the engineering version's performance [3][5]. - The evaluation of HSD's performance is limited due to the lack of comprehensive safety assessments and the variability of experiences across different locations [3][7]. - The HSD system demonstrated good vertical control, but its performance can vary significantly based on the city and driving conditions [6][7]. Group 2: Technical Comparisons - Horizon employs a VA-style end-to-end approach, while Li Auto utilizes a VLA-style end-to-end system, with the naming being a mere distinction [9][10]. - The VA-style end-to-end system is perceived to have advantages in user experience due to current limitations in computing power and bandwidth faced by the VLA approach [6][12]. - Li Auto's decision to pursue VLA for mass production is seen as a bold move, but it comes with challenges related to resource allocation and the need for higher computational requirements [11][12]. Group 3: Industry Outlook - There is a prevailing belief that many autonomous driving operators will eventually converge in capabilities, with only a few manufacturers able to survive without in-house development of autonomous driving technologies [3][11]. - The article suggests that manufacturers lacking self-research capabilities in autonomous driving may struggle to adapt to the evolving smart vehicle industry [3][11]. - The future landscape of autonomous driving will likely see a concentration of capabilities, with differentiation becoming increasingly important as the industry matures [3][11].

Core Viewpoint - The humanoid robot industry is facing significant challenges and may be entering a period of stagnation, with many companies failing to meet expectations and a lack of clear pathways to mass production [3][10]. Industry Performance - Internationally, companies like Tesla are struggling, with the Gen2 model facing overheating and durability issues, leading to a halt in production plans for this year, while Gen3 has been postponed to Q1 next year [3][4]. - Meta's AI chief and Google DeepMind's head have both indicated that true intelligence in humanoid robots is still years away, estimating a timeline of at least 5-10 years before robots can enter the home market [4]. Domestic Market Observations - The domestic market appears to be experiencing a false sense of prosperity, with many orders being reported as non-deliverable or merely framework orders that do not require immediate fulfillment [5][6]. Technological Limitations - Despite advancements in hardware, the industry has not achieved practical widespread application of robots, with AI technology not yet demonstrating the general intelligence needed for humanoid robots [8][9]. - Current AI applications in robotics are limited to specific scenarios and lack generalization capabilities, which could lead to failures in more complex environments like homes [12]. Challenges in Learning and Adaptation - Video learning, while a promising area, has not yet produced results that demonstrate the ability to generalize operations, with many companies still relying on real-world data collection rather than effective video learning techniques [15][17]. Potential Upsides - There are two uncertain factors that could influence the industry positively: 1. The performance of Tesla's Optimus Gen3, which is seen as a potential game-changer if it exceeds expectations [18][19]. 2. The possibility of hardware advancements leading to new market opportunities, as seen with companies like Yushun, which have successfully carved out niches in the entertainment sector [22][23]. Conclusion - The humanoid robot industry may be in a phase of necessary recalibration, similar to the early challenges faced by the electric vehicle sector, where technological advancements continued despite market difficulties [24].

Core Viewpoint - The article discusses the evolution of autonomous driving technology, highlighting the transition from rule-based systems to end-to-end models represented by companies like Ideal and XPeng, and currently to the world model phase represented by NIO, emphasizing the continuous presence of deep learning throughout these changes [1]. Group 1: Course Introduction - The course covers the development from modular production algorithms to end-to-end systems and now to VLA, focusing on core algorithms such as BEV perception, visual language models (VLM), diffusion models, reinforcement learning, and world models [5]. - Participants will gain a comprehensive understanding of the end-to-end technology framework and key technologies, enabling them to reproduce mainstream algorithm frameworks like diffusion models and VLA [5]. - Feedback indicates that students completing the course can achieve approximately one year of experience as end-to-end autonomous driving algorithm engineers, benefiting from the training for internships and job recruitment [5]. Group 2: Instructor Profile - The main instructor, Jason, holds a C9 undergraduate degree and a PhD from a QS top 50 university, with multiple published papers in CCF-A and CCF-B journals [6]. - He is currently an algorithm expert at a leading domestic manufacturer, engaged in the research and production of cutting-edge algorithms, with extensive experience in the development and delivery of autonomous driving perception and end-to-end algorithms [6]. Group 3: Research Guidance - The program aims to enhance practical skills and knowledge in cutting-edge topics, with a focus on helping students publish high-level papers to improve their academic prospects [8]. - The community includes over 300 instructors specializing in autonomous driving and embodied intelligence, with a high manuscript acceptance rate of 96% over the past three years [8]. Group 4: Research Process - The guidance process includes selecting research topics based on student interests, explaining key concepts, and providing essential foundational knowledge and recommended learning materials [11]. - Students will learn how to critically read literature, conduct research, and write various sections of a paper, including methods and experimental results, with continuous feedback and support throughout the process [11].

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 最近收到越来越多业内小伙伴和公司的诉求，希望自动驾驶之心能够在企业培训和求职辅导等方向上赋能。 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 感兴趣的可以添加微信wenyirumo做进一步咨询。 虽然从上半年开始，我们一直在筹办相关事宜。但众人拾柴火焰高，要推动大的行业进步，需要更多优秀的伙 伴加入我们。 现面向全球的自动驾驶领域从业者发出邀请函，自动驾驶之心期望能够和您在技术服务、培训、课程开发与科 研辅导等多个领域展开合作。 我们将提供高额的酬金与丰富的行业资源。 主要方向 ...

Core Insights - The article discusses the delivery performance of the company in October 2025, highlighting that the delivery of 31,767 units is considered low and that detailed model-specific expectations will be available later [2][3]. Delivery Performance - The delivery figure of 31,767 units for October 2025 is viewed as underwhelming, with specific model breakdowns expected to be released around November 10 [2]. - The L series orders are reported to be poor, while the MEGA model is performing normally [3]. Production Capacity Issues - The i8 model faces production capacity challenges due to a low configuration selection rate of approximately 2%, making it difficult for the supplier to ramp up production [4]. - The configuration strategy for the i8 differs significantly from the L78 model, with a smaller price gap and greater configuration differences [4]. - The i6 model will not be delivered with the Xinhua battery version in 2025, as the supplier also struggles to increase production capacity [10]. Configuration Preferences - The configuration distribution for the i8 is approximately 2% low configuration, 20% mid configuration, and 80% high configuration, which contrasts sharply with the L78 model's configuration choices [9]. - The L series models show a low configuration selection rate of about 22% for L7 and 37% for L8 [8]. Market Dynamics and Challenges - Various hypotheses are presented regarding the poor performance of L series orders, including rapid competitor iterations, economic conditions, and potential organizational issues [14][15]. - The company may have misjudged market conditions, leading to a lack of timely decisions regarding battery supply options [12][13]. Future Expectations - There are indications that the company may deliver around 100,000 units in Q4 2025, with expectations for Q1 2026 to remain positive, although the reliability of this information is uncertain [16].

Core Insights - Xiaomi has made significant strides in the autonomous driving sector since the establishment of its automotive division in September 2021, with plans to release the Xiaomi SU7 in March 2024 and the YU7 in June 2025 [2] - The company is actively engaging in advanced research, with a focus on integrating cutting-edge technologies into its autonomous driving solutions, as evidenced by a substantial number of research papers published by its automotive team [2] Research Developments - The AdaThinkDrive framework introduces a dual-mode reasoning mechanism in end-to-end autonomous driving, achieving a PDMS score of 90.3 in NAVSIM benchmark tests, surpassing the best pure vision baseline by 1.7 points [6] - EvaDrive presents an evolutionary adversarial policy optimization framework that successfully addresses trajectory generation and evaluation challenges, achieving optimal performance in both NAVSIM and Bench2Drive benchmarks [9] - MTRDrive enhances visual-language models for motion risk prediction by introducing a memory-tool synergistic reasoning framework, significantly improving generalization capabilities in autonomous driving tasks [13][14] Performance Metrics - The AdaThinkDrive framework has shown a 14% improvement in reasoning efficiency while effectively distinguishing when to apply reasoning in various driving scenarios [6] - EvaDrive achieved a PDMS score of 94.9 in NAVSIM v1, outperforming other methods like DiffusionDrive and DriveSuprim [9] - The DriveMRP-Agent demonstrated a remarkable zero-shot evaluation accuracy of 68.50% on real-world high-risk datasets, significantly improving from a baseline of 29.42% [15] Framework Innovations - ReCogDrive combines cognitive reasoning with reinforcement learning to enhance decision-making in autonomous driving, achieving a PDMS of 90.8 in NAVSIM tests [18] - The AgentThink framework integrates dynamic tool invocation with chain-of-thought reasoning, improving reasoning scores by 53.91% and answer accuracy by 33.54% in benchmark tests [22] - ORION framework effectively aligns semantic reasoning with action generation, achieving a driving score of 77.74 and a success rate of 54.62% in Bench2Drive evaluations [23] Data Generation Techniques - Dream4Drive introduces a 3D perception-guided synthetic data generation framework, significantly enhancing the performance of perception tasks with minimal synthetic sample usage [26] - The Genesis framework achieves joint generation of multi-view driving videos and LiDAR point cloud sequences, enhancing the realism and utility of autonomous driving simulation data [41] - The Uni-Gaussians method unifies camera and LiDAR simulation, demonstrating superior simulation quality in dynamic driving scenarios [42]

Group 1 - The article discusses the challenges in achieving real-time, high-fidelity, and multi-task output in autonomous driving systems, emphasizing the importance of 4D dynamic scene reconstruction [1][2] - It highlights the limitations of existing static and dynamic scene reconstruction methods, particularly their inability to handle moving objects effectively [3][4] Group 2 - The research introduces a two-phase training paradigm that first learns robust static scene priors before training the dynamic module, addressing the instability of end-to-end training [4][11] - A mixed shared architecture for the residual flow network is proposed, which allows for efficient dynamic modeling while maintaining cross-view consistency [4][14] - The method utilizes a pure visual online feed-forward framework that processes two consecutive panoramic images to output various results without offline optimization [4][18] Group 3 - The experimental results demonstrate significant improvements in novel view synthesis metrics, with the proposed method achieving a PSNR of 28.76, surpassing previous methods [13][20] - The efficiency analysis shows that the proposed method has a faster inference time of 0.21 seconds per frame, which is 38% faster than DrivingForward and 70% faster than Driv3R [18][19] - The qualitative results indicate that the proposed method effectively captures dynamic objects with clear edges and temporal consistency, outperforming existing methods in dynamic scene reconstruction [19][22]

点击下方 卡片 ，关注" 自动驾驶之心 "公众号 戳我-> 造车新势力10月成绩单出炉： 不出意外，零跑再次刷新了之前的记录，银十破七万。极氪首破六万台。小鹏、蔚来、小米也在持续刷新高。理想仍处于纯电转型的阵痛期，大致可能为L系列订 单不佳，i6/i8的产能不足。 零跑汽车：10月全系交付达70289台，同比增长超84%； 小鹏汽车：10月交付新车42013台，单月交付量创下历史新高； 理想汽车：10月交付新车31767台； 极氪科技：10月销量61636台，同比增长9.8%，环比增长20.5%，单月销量首次突破6万台； 小米汽车：10月交付新车超过40000台； 智界汽车：10月交付突破10000台； 岚图汽车：10月交付新车17218台； 北汽新能源：10月销量30542台，同比增长112%，环比增长48.7%； 智己汽车：10月销售13159台。 小米汽车：10月交付新车超过40000台 小鹏汽车：10月小鹏汽车交付新车42013台，单月交付量创下历史新高，并连续2个月超过4万台。 蔚来汽车：10月交付新车40397台，同比增长92.6%，环比增长16.25%。 星球宗旨 感知/定位/融合/部署/仿真等 ...

Core Viewpoint - The article introduces the GeoScan S1, a highly cost-effective handheld 3D laser scanner designed for industrial and educational applications, emphasizing its advanced features and capabilities for real-time 3D scene reconstruction. Group 1: Product Features - GeoScan S1 offers a lightweight design with a one-button start for efficient 3D scanning solutions, achieving centimeter-level accuracy in real-time scene reconstruction [2][10]. - The device can generate point clouds at a rate of 200,000 points per second, with a maximum measurement distance of 70 meters and 360° coverage, supporting large-scale scanning over 200,000 square meters [2][30]. - It integrates multiple sensors, including a high-precision IMU and RTK, enabling high-accuracy mapping and data synchronization [35][39]. Group 2: User Experience - The device is designed for ease of use, allowing users to export scanning results without complex setups, making it accessible for quick deployment in various environments [7][28]. - The GeoScan S1 supports offline and online rendering, enhancing the visualization of scanned data [8]. Group 3: Technical Specifications - The GeoScan S1 operates on an Ubuntu system and supports various data formats for point cloud output, including .pcd and .las [23]. - It features a compact size of 14.2cm x 9.5cm x 45cm and weighs 1.3kg without the battery, with a power input range of 13.8V to 24V [23][24]. Group 4: Market Position - The introductory price for the GeoScan S1 starts at 19,800 yuan, positioning it as one of the most affordable options in the market for handheld 3D laser scanners [10][58]. - The product has been validated through numerous projects in collaboration with academic institutions, showcasing its reliability and effectiveness in real-world applications [10][39].