Workflow
世界模型+闭环强化学习
icon
Search documents
蔚来世界模型新版加速覆盖 马麟:Cedar系列1.4.0已送备即将推送
Feng Huang Wang· 2026-01-28 10:31
Core Viewpoint - NIO is advancing its smart driving technology with the release of the Cedar and Cedar S version 1.4.0, indicating a commitment to iterative upgrades across different vehicle platforms [1] Group 1: Smart Driving Technology Updates - NIO's smart driving R&D head announced that the Cedar and Cedar S version 1.4.0 has been submitted for approval, showcasing the company's focus on upgrading smart driving capabilities [1] - The new version of the "World Model" has been launched, covering over 460,000 vehicles equipped with the Banyan smart system, indicating a significant scale of deployment [1] - The upgrade features a "World Model + Closed-loop Reinforcement Learning" architecture, enhancing lane selection, dynamic game theory, and vehicle control to a more human-like level [1] Group 2: New Functionalities and Hardware - The upgrade introduces a new "Urban Navigation Battery Swap" function, which can automatically connect to over 2,000 battery swap stations nationwide [1] - NIO emphasizes its high-performance hardware, such as the Banyan platform with a computing power of 1016 TOPS, which supports ongoing software upgrades [1] - The Cedar series vehicles, equipped with the self-developed 5nm driving chip NX9031, are progressively receiving the latest technological enhancements [1]
蔚来世界模型NWM新版推送:新增城区领航换电功能
Feng Huang Wang· 2026-01-28 02:55
Core Insights - NIO has launched a new version of its "NIO World Model (NWM)" on January 28, which covers over 460,000 vehicles equipped with the "Banyan" system, with updates for "Cedar" and "Cedar S" models coming soon [1] Group 1: System Upgrade Features - The upgrade utilizes a "world model + closed-loop reinforcement learning" paradigm aimed at enhancing driving comfort and efficiency [1] - Key improvements include enhanced lane selection accuracy, strong decision-making capabilities, and stable vehicle control, providing a balanced experience of safety and efficiency [1] - The system has significantly improved the ability to select lanes, avoid traffic, and dynamically change lanes, along with better performance in navigating intersections and overall road efficiency [1] Group 2: New Functionalities - A new "urban navigation battery swap" feature has been introduced, supporting over 2,000 second-generation and above battery swap stations nationwide, facilitating the transition from navigation driving to automatic battery swapping in urban scenarios [1] - NIO's high computing power hardware, such as the Banyan platform with a computing power of 1016 TOPS, supports this model upgrade [1] Group 3: Future Plans - NIO plans to increase cloud training computing power investments by 2026 to accelerate the iteration speed of its intelligent driving assistance features [2]