Smart Driving Weekly Report

Smart Driving Competition in Nanjing Kicks Off on April 20; Tesla Makes Its Debut

The second China Smart Driving Competition – City NOA Champion Challenge will take place in Nanjing on Sunday, April 20. This event will feature owners from ten brands, including Tesla, NIO, Xiaopeng, Li Auto, Xiaomi, Zhiji, Avita, Wenjie, Zhijie, and Tengshi, competing with a total of 11 vehicle models. The competition route spans 18 kilometers and is expected to take approximately 1 hour and 30 minutes. Tesla will be making its first appearance in this competition, where the top brands will clash in an urban smart driving showdown. Tune in to the live stream on the First Electric video channel, Douyin, or Weibo starting at 1:30 PM on April 20 to see who will emerge as the champion in Nanjing.

The competition will evaluate participants on three main categories and six subcategories: NOA scenarios, NOA safety behavior (including evasion behavior, ghost behavior, traffic violations, and traffic accidents), and NOA efficiency behavior. There are also two bonus criteria: commuting time and whether there is zero takeover. For detailed records of the competition, you can refer to the Judging Record Manual.

Awards include:

1. Champion Award: The participant with the highest overall score will become the city’s new smart driving champion, receiving a “City NOA Champion” trophy and a cash prize of 3,000 yuan.
2. Participant Cash Awards: Cash rewards will be given based on the overall scores, calculated as the score multiplied by 10. For instance, a score of 79.34 would yield a reward of 793.4 yuan.
3. Judging Participant Subsidy: All judges participating in the competition will receive a cash subsidy of 500 yuan.

Note: Currently, NOA smart driving only serves as an assisted driving feature. Participants must maintain a clear understanding and approach to safety while driving.

Ministry of Industry and Information Technology Urges Car Manufacturers to Ensure Thorough Testing

On April 16, the Equipment Industry Department of the Ministry of Industry and Information Technology held a meeting to discuss the access and software online upgrade management for intelligent connected vehicles. The meeting emphasized that automotive manufacturers must fully understand relevant notification requirements, conduct extensive tests to verify driving assistance systems, clarify functional boundaries, and ensure safety measures. They should avoid exaggeration and misleading promotions, strictly adhere to disclosure obligations, and take responsibility for production consistency and quality safety to enhance the safety level of intelligent connected vehicle products.

Smart driving is categorized into six levels from L0 to L5, with current technology in China transitioning from L2 to L3. L2 is driver-dominated while L3 and above are system-dominated. In 2024, it is projected that 13.094 million new vehicles with L2 or higher functions will be insured in China, achieving a penetration rate of 57.3%. Although L3 automated driving is being piloted in certain regions, widespread implementation is still restricted by regulations. Some models support urban NOA functions, but coverage is limited. Everbright Securities estimates that by 2025, the penetration rate of domestic L2+ urban smart driving will approach 10%, with significant growth expected from 2026 onwards. Despite ongoing discussions about smart driving safety, manufacturers must uphold their responsibilities and avoid overstating product safety capabilities or engaging in excessive marketing.

Traffic Police Warn Against Misleading Advertising of Autonomous Driving

On April 17, the Traffic Safety Research Center of the Ministry of Public Security issued a reminder to drivers to exercise caution when using intelligent connected vehicle assistance functions. The article highlighted that some drivers mistakenly equate assistance features with autonomy, leading them to engage in dangerous behaviors such as using mobile phones or sleeping while the assistance feature is active, which poses severe risks to road safety. It stressed that current mainstream production vehicles are still categorized within levels 0-2, falling under “human-machine co-driving.” Drivers must maintain hand control of the steering wheel at all times while monitoring road conditions. They should also understand how to activate or deactivate assistance features and refrain from using them in adverse weather or emergency situations.

Furthermore, it was emphasized that companies should enhance training for new users, standardize marketing practices, and avoid false or exaggerated claims about assistance features to prevent misuse. Misleading advertisements, distracted driving, and substandard products can lead to legal repercussions. Under Article 28 of the Advertising Law of the People’s Republic of China, companies that falsely exaggerate assistance features (e.g., describing level 2 assistance as “autonomous driving”) could face fines of five to ten times the advertising costs, and serious violations could result in business license revocation. If misleading advertising results in severe consequences, such as traffic accidents causing injury or death, those responsible may face imprisonment for up to two years or detention, along with potential fines.

Xiaopeng Motors Delays XOS 5.6.0 Update Initially Scheduled for April

On April 17, Xiaopeng Motors announced that the rollout of the XOS 5.6.0 system update, originally planned for April 2025, will be postponed due to adjustments in version scheduling. The official OTA account of Xiaopeng Motors stated that the update will be pushed to users as soon as the version is ready. Previously, the company had mentioned the XOS 5.6.0 update in its 2025 OTA plan.

Xiaopeng Plans to Achieve L3 Smart Driving Production by Year-End

On April 15, Xiaopeng Motors announced significant advancements in its smart driving technology during the “Global Love Night and Launch of the First Global Flagship Xiaopeng X9.” CEO He Xiaopeng revealed plans to achieve mass production of L3 smart driving capabilities in China by the end of 2025. In 2024, Xiaopeng will become the first Chinese company to implement end-to-end large model production in vehicles, enhancing the model’s generalization ability and long-tail scenario handling through reinforcement learning, which is expected to improve model accuracy by more than double. The Xiaopeng Turing AI driving chip is projected to begin production in mainland China in the second quarter of 2025. He Xiaopeng further stated that Xiaopeng aims to be the first Chinese company to deploy top-tier smart driving capabilities globally, with AI R&D investment rising from 3.5 billion yuan in 2024 to 4.5 billion yuan in 2025. He believes that in the next decade, competition in the automotive market will intensify, with only five to seven new energy vehicle companies likely to survive.

Xiaopeng also announced that its Kunpeng Super Electric System will enter mass production in mainland China in the fourth quarter of 2025. This system comprises the “Kun” Super Range Extender System and “Peng” Pure Electric System, with the former claiming a pure electric range of 430 km, nearly doubling the range of existing range-extended models. The first range-extended vehicle will be a large SUV, code-named G01, developed based on the G9 model, aimed at the 200,000 yuan price segment, with mass production planned for the second half of 2025.

Landwind FREE Model to Feature Comprehensive Huawei Technology

On April 12, CEO Lu Fang of Landwind Automotive Technology Co., Ltd. confirmed that the new Landwind FREE model is a significant redesign that integrates advanced technologies from Dongfeng and Landwind, featuring “full-blood Huawei” technology aimed at creating a popular SUV. Photos of the first SUV jointly launched by Dongfeng and Huawei have been revealed, highlighting features like “Leading Beyond Generations,” “Four Years in the Making,” and “Huawei’s Strongest Smart Driving.” Landwind announced that the new model will incorporate advanced Huawei smart driving technologies, including lidar, and showcase a new front design, wheels, and cabin configurations, all marked with a red “Huawei ADS” label.

Landwind Reveals World’s First L3 Intelligent Architecture

On April 16, Landwind held a technology launch event for its “Tianyuan Intelligent Architecture,” the world’s first L3 intelligent architecture. This architecture combines the Qingyun L3 Intelligent Safety Driving Platform and the Kunpeng L3 Intelligent Safety Driving System, upgrading from the mechanical layer to the intelligent driving layer to provide a hardware and algorithm foundation for L3 smart driving. The Qingyun platform integrates steer-by-wire, brake-by-wire, and fully active suspension, supporting comprehensive vehicle intelligence and L3 autonomous driving. The Kunpeng system includes a full-domain real-time situation model and a multimodal large model that integrates visual and language actions, deploying an end-to-end large model with independent safety strategy modules and equipping 48 sensors with over 1000 TOPS computing power to support 300-meter detection distances and high-precision recognition. Lu Fang stated that the company has collaborated with multiple insurance firms, including Ping An, and aims to launch smart driving insurance by June 2025. The first model equipped with the Tianyuan architecture, code-named Taishan, is positioned as a full-size luxury six-seater SUV and will be unveiled within the year, featuring the Qingyun L3 platform and Kunpeng L3 intelligent driving.

Volkswagen Group to Showcase Self-Developed Advanced Driving Assistance System

On April 17, Volkswagen Group announced that it will debut an advanced driving assistance system designed specifically for the Chinese market at the Shanghai Auto Show. Developed by a Chinese team, the system aims to address the complex and changing road conditions in China. The group’s software company CARIAD, in collaboration with Horizon, is developing advanced driving assistance and autonomous driving technologies tailored for the market. This AI-powered system features self-learning capabilities, providing a highly human-like and multi-scenario driving assistance experience with a focus on safety. The system will initially be implemented in a new Volkswagen model aimed at the Chinese market and is expected to be applied to new intelligent connected vehicles based on the CMP platform starting next year. Volkswagen Group plans to launch over 20 electrified models by 2027 and approximately 30 pure electric models by 2030.

Aion and Didi Unveil First L4 Autonomous Driving Model Set for Mass Production by 2025

On April 13, GAC Aion and Didi showcased their first L4 autonomous driving model at the GAC Intelligent Safety Summit in Guangzhou. This model integrates Didi’s autonomous driving hardware platform with GAC Aion’s safety redundancy platform, designed for global adaptability, with plans for mass production and delivery by the end of 2025. Didi’s next-generation autonomous driving hardware platform includes 33 sensors, such as lidar, cameras, 4D millimeter-wave radar, infrared cameras, and sound sensors, employing a multi-sensor heterogeneous detection and fusion approach to achieve 360° full-scene and full-condition perception. Additionally, Didi has developed the industry’s first mass-producible centralized vehicle computing platform with over 2000 TOPS GPU computing power and 48 CPU cores, characterized by high performance and low latency. The autonomous driving system of this model features multiple layers of redundancy in algorithms, software, and hardware to minimize safety risks in emergencies. The base chassis meets the dual five-star safety standards of China and Europe, equipped with a modular battery system that provides multiple redundancies for braking, steering, communication, and power to ensure safety in electrification and vehicle integrity.

Musk: Tesla to Achieve Fully AI-Driven Autonomous Driving

On April 15, Tesla CEO Elon Musk announced via the X platform that the company is nearing the implementation of a “universal, purely AI-driven full autonomous driving (FSD) solution.” Musk indicated that this will be the first fully autonomous driving solution relying solely on cameras and Tesla’s in-house AI chips, powered by Tesla-developed AI software. This statement aligns with Tesla’s long-standing vision of achieving autonomy through vision alone. While the ambitious approach has garnered praise for its vision, it has also faced criticism over safety concerns. Musk’s comments come amid intensifying competition in the autonomous driving field, with companies like Waymo and Cruise pursuing different technological avenues for developing autonomous vehicles. Tesla’s advancements may escalate competition in the autonomous driving sector and spark further discussions regarding the reliability of purely vision-based autonomous driving technologies.

Li Auto Releases Star Ring OS Technical White Paper and Announces Open Source Plans

On April 15, Li Auto published the “Li Auto Star Ring OS Technical White Paper,” which outlines the definition, development trends, and innovative case studies of the automotive operating system. CEO Li Xiang expressed confidence that from day one, Star Ring OS would evolve into a “spatial intelligent operating system,” pledging to avoid commercial barriers and continuously invest in R&D iterations. According to the white paper, Li Auto will gradually open source various modules starting at the end of April. The specific timeline is as follows: on April 23, the Star Ring OS 0.1.0 version will be open-sourced, including safety real-time RTOS and communication bus-Lite; by June 30, the Star Ring OS 1.0.0 version will be available, encompassing the complete intelligent vehicle control and smart driving system capabilities; and after July, the complete intelligent driving system will be made open-source, including the virtualization engine. Li Xiang emphasized that open sourcing represents a new starting point for Star Ring OS, inviting all peers, developers, and partners to collaborate in refining the system and advancing the smart automotive industry.

Reports Indicate Zhi Yu Technology Secures Audi’s New Car Project and Engages with Mercedes Executives

On April 15, reports revealed that Audi will launch two new vehicles in collaboration with Zhi Yu Technology, and Mercedes executives visited Zhi Yu Technology seeking “Plan B” for smart driving solutions. Additionally, Great Wall’s second-generation Xiaolong MAX and BAIC’s Arcfox brand will also adopt Zhi Yu Technology’s smart driving solutions. Currently, Zhi Yu Technology has partnered with nine companies, including FAW, Volkswagen, SAIC-GM-Wuling, Chery, and Dongfeng, with 17 models already collaborating on smart driving, and over 30 new models in development. Zhi Yu Technology follows a purely visual smart driving route, with its “Chenghang Platform” base version (7V+32TOPS) already in mass production, achieving end-to-end urban navigation assistance for the first time. At the end of last year, Zhi Yu Technology underwent corporate changes, with Shenzhen BYD Chuangxin Materials Co., Ltd. and two investment funds managed by Hengxu Capital becoming shareholders.

SAIC Group’s Smart Driving Team Completes Integration, Zero Bundle Technology Merged into R&D Institute

On April 12, SAIC Group’s software company Zero Bundle Technology and the SAIC R&D Institute announced their integration, with both smart driving teams combining their operations to advance R&D efforts. This merger primarily involves departments with overlapping responsibilities in prior R&D work, with Zero Bundle Technology retaining two divisions focused on electronic architecture and computing platform development, maintaining a workforce of 600-700 employees. The newly integrated smart driving team has yet to determine specific projects and is currently awaiting project assignments. Previously, the R&D Institute was responsible for the Roewe and MG brands, while Zero Bundle Technology handled the Feifan brand.

If you are a frequent user of smart driving features, feel free to join the First Electric Smart Driving Alliance by contacting our operations team. If you are interested in smart driving and want to learn more, you can also join the First Electric brand smart driving group by reaching out. If you find the Smart Driving Weekly Report helpful, we welcome your likes, comments, and shares.