Core Viewpoint - The article highlights the strategic collaboration between Youjia Innovation and Shenzhen Postal Group, along with Eastern Public Transport, focusing on the integration of logistics and public transport to enhance resource utilization and operational efficiency [2][3]. Group 1: Strategic Collaboration - Youjia Innovation has entered a strategic partnership with Shenzhen Postal and Eastern Public Transport, aiming for a systematic layout in the "logistics + public transport" sectors [2]. - The collaboration emphasizes leveraging each party's core resources and channel advantages to create a long-term, mutually beneficial partnership [2][3]. - The partnership will focus on four main areas: integration of unmanned logistics vehicles with public transport, composite utilization of bus station spaces, cross-scenario business expansion, and collaborative innovation in labor utilization [3]. Group 2: Technological Advancements - Youjia Innovation is utilizing its L1-L4 autonomous driving technology and production experience to develop a smarter, safer, and more efficient delivery model [3]. - The collaboration aims to innovate the urban delivery system and public transport network integration, reducing redundancy costs in warehousing and transportation [3]. Group 3: Recent Developments - This partnership marks the third announced application scenario for Youjia Innovation's L4 business in a short period [4]. - Previous collaborations include a partnership with Chongqing Recycled Resources Group to deploy new energy heavy-duty trucks and intelligent scheduling systems in mining areas [4]. - Youjia Innovation also secured a project for L4 unmanned minibuses at Ezhou Huahu International Airport, marking its first commercial project in airport autonomous driving [4][5].

Core Insights - The recognition of Peng Jun, CEO of Pony.ai, as one of the "100 Most Influential People in AI" by TIME magazine highlights the growing importance of autonomous driving in the AI sector and reflects the global strategic emphasis on this technology [1][2][3] Group 1: Industry Developments - 2025 is referred to as the "Year of Robotaxi Mass Production," with Pony.ai's seventh-generation autonomous Robotaxi achieving mass production and public road testing in cities like Guangzhou and Shenzhen [2][3] - Pony.ai is actively expanding its global footprint, with operations and testing in regions such as Asia (South Korea, Singapore), the Middle East (Saudi Arabia, UAE), and Europe (Luxembourg) [2][3] - The company has produced over 200 units of its seventh-generation Robotaxi, aiming for a fleet size of 1,000 vehicles, which signifies a shift from technical exploration to practical application [3][11] Group 2: Technological Advancements - The seventh-generation Robotaxi utilizes 100% automotive-grade components, achieving a 70% reduction in the total cost of the autonomous driving suite, and has reached operational profitability per vehicle [3][11] - The vehicle features advanced safety architectures, including over 20 safety redundancies and more than 1,000 detection designs, differentiating it from L2 assisted driving systems [5][11] - The company emphasizes the importance of technological innovation and cost optimization, leveraging large-scale production to drive down costs in key components like LiDAR and Orin X chips [4][12] Group 3: Strategic Vision and Challenges - The company aims to achieve breakeven by 2028-2029, with a focus on enhancing self-sufficiency and solidifying technological barriers [4][5] - Regulatory frameworks are seen as essential for the healthy development of the industry, with Pony.ai participating in the formulation of standards for intelligent connected vehicles [6][9] - The company believes that the mass adoption of Robotaxi will take at least another decade, requiring ongoing efforts to address production capacity, regulatory improvements, and user trust [6][10] Group 4: Competitive Landscape - Pony.ai is positioned as a leader in the autonomous driving sector, with a unique advantage stemming from China's dense population and diverse driving conditions, which provide rich testing scenarios [12][13] - The company is leveraging government policies that support the development of autonomous driving technologies, which have been elevated to a national strategic level [13] - Internationally, Pony.ai is collaborating with regulatory bodies to facilitate the understanding and implementation of autonomous driving technologies in various markets [13]

Group 1 - Waymo announced its expansion into Denver and Seattle, furthering its presence in the U.S. market [2] - BYD's overseas sales are expected to double this year, with significant historical milestones in its international business since 1998 [2] - Rivian launched a custom version of its R1S model, inspired by 1980s Miami design, featuring a power output of 850 hp and a range of 597 km [2] Group 2 - In July 2025, China's automobile exports reached 694,000 units, a month-on-month increase of 12.1% and a year-on-year increase of 25.6%, with export value at $11.84 billion [4] - From January to July 2025, total automobile exports amounted to 4.166 million units, reflecting a year-on-year growth of 19.6% [4]

Company Overview - WeRide is a global leader and first mover in the autonomous driving industry, being the first publicly traded robotaxi company [1] - The company has tested or operated its autonomous vehicles in over 30 cities across 10 countries [1] - WeRide is the only technology company to have received autonomous driving permits in six markets: China, France, Saudi Arabia, Singapore, the UAE, and the US [1] - The WeRide One platform offers autonomous driving products and services ranging from Level 2 to Level 4, catering to mobility, logistics, and sanitation industries [1] - WeRide was recognized in Fortune Magazine's 2024 "The Future 50" list [1] Upcoming Investor Conferences - WeRide management will participate in several investor conferences, including: - Goldman Sachs Asia Leaders Conference in Hong Kong on September 3, 2025 [3] - BofA Asia Pacific Conference in Hong Kong on September 8, 2025 [3] - CLSA 32nd Investors' Forum 2025 in Hong Kong on September 9, 2025 [3] - Goldman Sachs Communacopia + Technology Conference 2025 in San Francisco on September 11, 2025 [3] - BofA Future Car Conference, virtual, on September 24, 2025 [3] - Morgan Stanley Asia BEST Corporate Day in New York on September 24, 2025 [3] - Evercore ISI Autonomous, ADAS, AI Forum in New York on September 30, 2025 [3]

Group 1 - The event "Smart City and Culinary Tourism Integration" in Shunde marks the launch of an innovative model combining food and autonomous driving, enhancing the integration of technology and cultural tourism [1][2] - The newly opened "Shunde Culinary Tourism Autonomous Driving Line" connects over 20 cultural tourism landmarks, covering a distance of 7.2 kilometers, catering to the younger generation's demand for smart travel and cultural experiences [1][2] - The initiative aims to promote the deep integration of technology with culture and tourism, encouraging more visitors and businesses to engage with the Shunde area [1][2] Group 2 - The event is a significant step towards building an "Intelligent Travel + Cultural Tourism" ecosystem in Lecong, with the service already covering the core area of Foshan New City [2] - The autonomous driving service is designed to make exploring Shunde's culinary attractions more convenient and engaging, appealing to both residents and tourists [2] - Future plans include creating a "Mobile Culinary Tourism Museum," launching a "Night Tour Line," and ensuring seamless connections with public transport to enhance the travel experience [2][3] Group 3 - The company aims to empower cultural tourism through technology, establishing a unique travel brand for Foshan that offers rich, deep, and intelligent experiences for visitors [3]

Legal & Compliance - Nvidia 将面临一位工程师泄露自动驾驶商业机密案件的审判，该工程师被指控从前雇主处窃取了这些机密 [1]

Core Insights - The article emphasizes the importance of accurately predicting the motion of road agents for the safety of autonomous driving, introducing a reward-driven intent reasoning mechanism to enhance trajectory prediction reliability and interpretability [3][5][10]. Summary by Sections Introduction - Trajectory prediction is a critical component of advanced autonomous driving systems, linking upstream perception with downstream planning modules. Current data-driven models often lack sufficient consideration of driving behavior, limiting their interpretability and reliability [5][10]. Methodology - The proposed method adopts a "reasoning first, then predict" strategy, where intent reasoning provides prior guidance for accurate and reliable multimodal motion prediction. The framework is structured as a Markov Decision Process (MDP) to model agent behavior [8][10][12]. - A reward-driven intent reasoning mechanism is introduced, utilizing Maximum Entropy Inverse Reinforcement Learning (MaxEnt IRL) to learn agent-specific reward distributions from demonstrations and relevant driving environments [8][9][10]. - A new query-centered IRL framework, QIRL, is developed to efficiently aggregate contextual features into a structured representation, enhancing the overall prediction performance [9][10][18]. Experiments and Results - The proposed method, referred to as FiM, is evaluated on large-scale public datasets such as Argoverse and nuScenes, demonstrating competitive performance against state-of-the-art models [28][30][32]. - In the Argoverse 1 dataset, FiM achieved a minimum average displacement error (minADE) of 0.8296 and a minimum final displacement error (minFDE) of 1.2048, outperforming several leading models [32][33]. - The results indicate that the intent reasoning module significantly enhances prediction confidence and reliability, confirming the effectiveness of the proposed framework in addressing complex motion prediction challenges [34][36]. Conclusion - The work redefines the trajectory prediction task from a planning perspective, highlighting the critical role of intent reasoning in motion prediction. The proposed framework establishes a promising baseline for future research in trajectory prediction [47].

Summary of Horizon Robotics Conference Call Company Overview - **Company**: Horizon Robotics (9660.HK) - **Industry**: Autonomous Driving Technology Key Financial Highlights - **1H25 Revenue**: Rmb1.6 billion, representing a **68% YoY increase** and **8% HoH increase**, exceeding estimates by **6%** and **9%** respectively [1][3] - **Gross Margin**: 65.1% in 1H25, consistent with expectations [3] - **Operating Loss**: Rmb1.855 billion in 1H25, higher than the estimated loss of Rmb1.412 billion due to increased cloud service fees [3][7] - **Net Loss**: Rmb5.233 billion in 1H25, significantly worse than the expected loss of Rmb1.4 billion [3][7] Product Development and Market Strategy - **Mass Production Plans**: The company plans to start mass production of the Horizon Robotics SuperDrive (HSD) on the Journey 6P platform in **2H25**, targeting urban NOA features in vehicles [1][2] - **Product Mix Upgrade**: Anticipated ramp-up of HSD and J6P platform in **2026**, with expectations for higher dollar content per vehicle due to increased ASP [2] - **Urban NOA Penetration**: The HSD solution is expected to penetrate lower-priced car models, enhancing the product mix towards higher-end integrated solutions [1][2] Market Performance and Future Outlook - **Highway NOA Shipments**: Shipments of highway NOA-capable products reached **0.98 million** in 1H25, accounting for **50%** of total shipments, contributing to improved blended ASP [1] - **Revenue Projections**: Revenue estimates for 2025E revised up to Rmb3.605 billion, reflecting a **1% increase** from previous estimates [8][16] - **Target Price**: The 12-month target price is set at **HK$14.00**, indicating a potential upside of **76.3%** from the current price of **HK$7.94** [16] Risks and Challenges - **Competitive Landscape**: Risks include increased competition and pricing pressure in the auto supply chain amid slow demand [15] - **Product Mix Transition**: Potential delays in the transition towards advanced driver-assistance systems (ADAS) could impact growth [15] - **Supply Chain Vulnerabilities**: Geopolitical tensions may pose supply chain risks [15] Conclusion - Horizon Robotics is positioned for growth with strong revenue increases driven by product innovation and market expansion. However, the company faces significant challenges, including operational losses and external market pressures. The outlook remains positive with a maintained "Buy" rating based on anticipated product advancements and market penetration strategies.

Group 1 - The event showcased various technological innovations aimed at enhancing the experience of the upcoming sports event, including autonomous vehicles and smart assistive devices [1][2] - The Hong Kong University of Science and Technology introduced micro-nano cooling technology that can reduce surface temperatures by at least 15 degrees Celsius, and indoor navigation technology has been implemented for audience convenience [2] - Technologies such as L4 autonomous driving for athlete and audience transport, smart inspection robots, and AI-driven sports systems were highlighted as part of the event's operational support [2][3] Group 2 - Health technology plays a crucial role in the event, with exoskeleton devices designed for athlete rehabilitation and wellness solutions like ultrasonic medicine baths and sleep aid sofas being presented [3] - The integration of innovative technologies aims to create a superior competition environment for athletes and a more convenient viewing experience for spectators [3]

Core Insights - The article introduces a novel "perception-in-plan" paradigm for end-to-end autonomous driving, implemented in the VeteranAD framework, which integrates perception directly into the planning process, enhancing the effectiveness of planning optimization [5][39]. - VeteranAD demonstrates superior performance on challenging benchmarks, NAVSIM and Bench2Drive, showcasing the benefits of tightly coupling perception and planning for improved accuracy and safety in autonomous driving [12][39]. Summary by Sections Introduction - The article discusses significant advancements in end-to-end autonomous driving, emphasizing the need to unify multiple tasks within a single framework to prevent information loss across stages [2][3]. Proposed Framework - VeteranAD framework is designed to embed perception into planning, allowing the perception module to operate more effectively in alignment with planning needs [5][6]. - The framework consists of two core modules: Planning-Aware Holistic Perception and Localized Autoregressive Trajectory Planning, which work together to enhance the performance of end-to-end planning tasks [12][39]. Core Modules - **Planning-Aware Holistic Perception**: This module interacts across three dimensions—image features, BEV features, and surrounding traffic features—to achieve a comprehensive understanding of traffic elements [6]. - **Localized Autoregressive Trajectory Planning**: This module generates future trajectories in an autoregressive manner, progressively refining the planned trajectory based on perception results [6][16]. Experimental Results - VeteranAD achieved a PDM Score of 90.2 on the NAVSIM navtest dataset, outperforming previous learning methods and demonstrating its effectiveness in end-to-end planning [21]. - In open-loop evaluations, VeteranAD recorded an average L2 error of 0.60, surpassing all baseline methods, while maintaining competitive performance in closed-loop evaluations [25][33]. Ablation Studies - Ablation studies indicate that the use of guiding points from anchored trajectories is crucial for accurate planning, as removing these points significantly degrades performance [26]. - The combination of both core modules results in enhanced performance, highlighting their complementary nature [26]. Conclusion - The article concludes that the "perception-in-plan" design significantly improves end-to-end planning accuracy and safety, paving the way for future research in more efficient and reliable autonomous driving systems [39].