根据即将于4月1日起实施的《北京市自动驾驶汽车条例》规定，将支持自动驾驶汽车用于个人乘用车、城市公共汽电车、出租车、城市运行保障等出行服务。这也就意味着，L3级自动驾驶汽车已经具备合法上路的法规依据，并将覆盖到私家车领域。在这样的大背景下，自动驾驶也成为今年全国“两会”上代表委员们热议的话题，小米CEO雷军、小鹏汽车董事长兼CEO何小鹏等全国人大代表都提出了加速自动驾驶技术落地以及完善自动驾驶法规与标准等建议。

　　何为自动驾驶？

　　如今，汽车产业已迈入智能化高级阶段，相关波士顿咨询预测，2025年中国自动驾驶市场规模将达4500亿元，已成汽车行业、科技行业竞逐的核心领域。那么，自动驾驶如何定义？和如今逐步普及的智能驾驶有何不同？根据2022年实施的《汽车驾驶自动化分级》（GB/T 40429-2021）规定，将驾驶自动化分为了6级。其中，L0至L2级为驾驶辅助，现阶段多家车企提出的“智驾平权”，都属于L2级范畴，驾驶员需全程监控驾驶。从此次北京放开的L3级开始，就进入了自动驾驶的范畴，属于有条件自动驾驶，紧急情况下需要驾驶员接管车辆；而L4级和L5级则分别代表高度自动驾驶和完全自动驾驶，车辆几乎可以在所有场景下自主运行。

　　目前，自动驾驶领域主要有激光雷达方案和纯视觉方案两种技术路线。其中，激光雷达方案是通过发射激光束并接收反射信号，生成高精度的三维环境地图，并与其他传感器（如摄像头、雷达）结合使用，形成多传感器融合的感知系统。优势在于可靠性、高精度感知和全天候性能；缺点也显而易见，就是成本高昂。另外，激光雷达通常体积较大、功耗高，对于产品设计和能源管理要求很高。相比之下，纯视觉方案主要依赖摄像头作为核心传感器，通过计算机视觉算法对摄像头捕捉到的图像进行处理，识别道路上的车辆、行人、交通标志、信号灯等物体，并基于这些信息做出驾驶决策。核心在于深度学习和神经网络，通过大量的数据训练，逐步提升系统对复杂环境的理解能力。优势在于成本较低、技术成熟和数据丰富，缺点在于受环境影响较大，精度有限，对于计算资源需求高。

　　自动驾驶迎来“曙光”

　　近年来，我国在自动驾驶领域的政策支持力度不断加大。从国家层面的战略规划到地方政府的实施细则，自动驾驶的法规环境在逐步完善。早在2015年，工信部批准上海国际汽车城建立了中国首个国家智能网联汽车（上海）试点示范区” 标志着我国自动驾驶道路试点的正式启动。随后，《智能网联汽车道路测试管理规范（试行）》、《关于开展智能网联汽车准入和上路通行试点工作的通知》、《自动驾驶汽车运输安全服务指南（试行）》、《关于开展智能网联汽车“车路云一体化”应用试点工作的通知》等一系列政策的相继发布，为自动驾驶行业的发展提供了明确、广阔的市场前景。

　　如今，在国家政策的引导下，地方层面已经开始针对自动驾驶展开立法尝试。作为国内首部针对自动驾驶领域的地方性法规，即将实施的《北京市自动驾驶汽车条例》，明确鼓励支持自动驾驶汽车技术创新和产业发展，对于驾驶行业发展具有意义，不仅为L3级自动驾驶汽车的合法上路提供了法规依据，还推动了自动驾驶技术的商业化落地、基础设施建设、数据安全管理等多个方面的发展。

　　按照《北京市自动驾驶汽车条例》要求，对于自动驾驶创新应用活动进行了全环节规范，包括道路测试和示范活动的办理流程、安全评估制度以及申请开展道路应用试点活动的条件、程序。同时，在安全保障方面也作出了明确规定，如安全员、平台安全监控人员的岗位职责，网络安全和数据安全管理等。此外，还鼓励保险机构开发符合自动驾驶汽车特点的专属保险产品，维护相关企业、驾驶员和行人的权益。

　　除了国家政策和地方法律法规的扶持，消费者接受度的逐步提升也为自动驾驶未来的商业化奠定了重要基础。事实上，随着自动驾驶技术的不断发展，以及全国各地智能网联汽车试点工作的推进，消费者对于自动驾驶的信任度有所增加，尤其是在一些特定场景下，如高速公路、封闭园区等，自动驾驶的便利性和安全性已经得到了初步验证。据《汽车智能化发展报告（2024）智驾篇》报告显示，智能化已经成为中国消费者购车最重要的考量因素之一，有90%的消费者愿意为高阶智能驾驶服务额外付费。另外，2025年2月，比亚迪的“开创全民智驾时代”，不仅降低了智驾门槛，同时也让更多消费者了解到什么是智驾，为下一阶段的自动驾驶普及埋下伏笔。

　　规模化量产仍面临多重挑战

　　虽然现阶段国家政策和市场层面为自动驾驶的发展提供了有力支持，但要实现规模化量产落地，依然面临着诸多挑战。

　　1、法律法规滞后事故权责难划清

　　众所周知的，便是相关法律法规滞后的问题。一方面，自动驾驶汽车的合法身份尚未明确，只是在北京等试点城市得到了承认，目前还没有全国性的法律支撑；另一方面，自动驾驶目前仍没有统一的法律框架规范。一旦发生交通事故，由于涉及驾驶员、汽车制造商、软件供应商等多个主体，难以明确责任划定。此前，深圳、广州、长沙等地先后发布了针对自动驾驶车辆的地方性法规，但均未对用户最关心的事故责任界定作出明确规定。而即将实施的《北京市自动驾驶汽车条例》同样无法解决这个问题，只是规定了“自动驾驶汽车上路通行期间，违反道路交通安全法律法规或者发生交通事故的，由公安机关交通管理部门按照国家有关规定调查和处理。”对此，有专家表示，交通事故的责任认定属于国家立法事权，北京等地方立法无权作出规定。车质网曾针对“L3级自动驾驶出事故，谁来负责？”这个话题进行过相关调查，超8成的网友认为应该是车企承担责任。参考德国现行的自动驾驶法规来看，允许L3级自动驾驶上路，明确系统运行时，责任暂时转移给车辆制造商，但驾驶员仍需具有接管能力。法规一般会规定系统要求驾驶员接管的时间，比如10秒内必须接管，否则驾驶员可能承担部分责任。

　　除了法规滞后外，如何判定车辆出现事故时是否处于自动驾驶状态，也是一个难点。此前，理想L9(配置|询价)在2022年上市时，车头和车尾共配备了5个蓝绿色的标志灯（ADS标志灯），作用就是向外界传递车辆是否处于智能驾驶状态。后续，比亚迪和小鹏在推出的新车型中也加入了类似的标志灯。不过，由于目前没有强制性国标要求，大部分具有高阶智驾功能的车辆都没有配备。

　　2、技术瓶颈感知与决策能力不足

　　尽管自动驾驶技术取得了显著进展，但其在极端天气和复杂路况下的应用仍面临重大挑战。现有传感器系统，包括激光雷达、视觉摄像头和毫米波雷达等核心感知设备，在应对浓雾、暴雨等恶劣气象条件时，其感知精度和可靠性尚未达到商业化部署的要求。更为关键的是，自动驾驶算法在面对突发交通事件（如交通事故、道路施工等动态场景）时，决策的稳定性和准确性依然不足，对于自动驾驶车辆的安全性和可靠性产生直接影响。根据车质网投诉数据显示，自2020年以来，“驾驶系统故障”投诉量始终处于较高水平，其中2023年和2024年均超过千宗，反映出当前智驾功能仍存在局限性。

　　据来自深圳的刘先生向车质网反馈，其购买的某国产新势力品牌SUV，在行驶过程中开启了NOA辅助驾驶，在前方道路无其他障碍物的情况下，系统误判断道路情况，导致车辆失控撞上绿化带。

　　另外，此前首批测试特斯拉的车主被爆出现频繁违章，如压实线变道、闯红灯、走公交车道等，反映出系统感知、决策和控制模块存在不足。特别是在国内复杂的交通环境中，传感器数据可能会受到干扰或遮挡，进而导致决策失误。尤其是特斯拉这种走纯视觉路线的智驾方案，受外界环境影响更为明显。更重要的是，现阶段国内自动驾驶都是依靠单车智能，对于车-车、车-路以及车-人的协同水平远未达到要求。

　　3、高昂成本限制商业化进程

　　除了法规和技术层面的挑战外，高昂的成本也是自动驾驶所面临的待解难题之一。具备自动驾驶能力的车辆往往需要搭载高精度传感器，同时为了决策和执行的准确性，还需要运行先进、复杂的算法，这就需要强大的算力支撑，令单车成本远高于传统车辆。以特斯拉焕新Model Y(配置|询价)为例，选装FSD智能辅助驾驶功能需要花费6.4万元，接近入门车型售价的1/4。当然，也有车企不对高阶智能驾驶功能收费，比如理想，但旗下搭载AD Max系统的车型售价普遍在30万以上。此外，大规模测试和验证所需的复杂场景模拟进一步增加了研发和运营成本，限制了企业普及的能力。

　　4、存在个人隐私泄露风险

　　自动驾驶车辆需要依赖大量的传感器（如摄像头、激光雷达、GPS等）来感知周围环境，这些设备会持续收集包括车辆位置、行驶轨迹、车内乘客行为等敏感信息。车内摄像头和麦克风可能记录乘客的对话、面部表情甚至生物特征数据（如疲劳状态），这些数据的收集可能超出用户预期。一旦被黑客攻击，数据泄露，不仅侵犯个人隐私，甚至还可能被用于情报搜集或网络攻击，危及国家安全。

　　自动驾驶真正落地需多方努力

　　不可否认的是，完全自动驾驶是未来的发展趋势，但在当前阶段，如何加速自动驾驶量产商用进程仍需要多方共同努力。

　　首先是国家层面，应尽快出台统一的自动驾驶汽车法律体系，明确自动驾驶汽车的上路规则、事故责任认定标准和赔付主体，并完善网络安全、数据安全等配套法律。同时，通过制定国家强制标准，推动汽车企业在新产品上标配ADS标志灯，为事故后期责任认定提供有力的证据。在此，也呼吁工信部在《道路机动车辆生产企业及产品准入许可管理办法》中，将配备ADS标志灯列为新车准入条件，作为车辆上市前必须通过的安全检测项目之一。

　　其次是企业层面。自动驾驶技术的快速发展不仅依赖于技术的突破，还需要通过降低成本、创新商业模式来实现商用落地。对此，企业应继续加大研发投入，推动传感器（如激光雷达、摄像头）、智驾芯片等硬件设备实现规模化生产，进一步降低成本。另外，对于自动驾驶系统进行大量的AI和仿真训练，优化自动驾驶系统在复杂场景（如极端天气、突发交通事件）中的表现。

　　再有就是自动驾驶保险。当前，虽然市面上有一些智驾险种，但基本上都是汽车企业在主导，只针对本品牌，带有一定的局限性。保险公司应积极开发适用于自动驾驶汽车专属保险，包括交强险、商业险、三者险等，降低自动驾驶汽车推广门槛，维护驾驶人、乘客和行人的权益。

　　而对于消费者而言，要先明确智能驾驶、自动驾驶和无人驾驶的区别和适用场景，通过试乘试驾体验，提升对于自动驾驶技术的认知，根据自身实际需求进行理性消费。

　　总结

　　业内普遍认为，2025年或将成为L3级自动驾驶规模化元年，而北京市率先实施的《北京市自动驾驶汽车条例》无疑就是自动驾驶发展进程中一个新的“拐点”，为其他城市和地区自动驾驶落地提供可借鉴的样板。可以预见的是，随着L3级私家车实现合法上路，行业即将迎来新一轮的洗牌。在政策铺路、消费者接受度提升的大背景下，各大车企之间的竞争也将进入全新阶段。谁能将技术领先转化为商业成功，谁就是留在牌桌上的那一位。

According to the Regulations of autonomous vehicle in Beijing, which will be implemented on April 1, autonomous vehicle will be supported for personal passenger vehicles, urban buses and trolley buses, taxis, urban operation guarantee and other travel services. This means that L3 autonomous vehicle already have the legal basis for going on the road and will cover the field of private cars. Against such a backdrop, automatic driving has also become a hot topic for delegates at this year's "two sessions" of the National People's Congress. Representatives of the National People's Congress, including Lei Jun, CEO of Xiaomi, and He Xiaopeng, chairman and CEO of Xiaopeng Auto, put forward suggestions on accelerating the landing of automatic driving technology and improving automatic driving laws and standards.

What is autonomous driving?
The landing of L3 level autonomous driving is approaching, ushering in a new "turning point"

Nowadays, the automotive industry has entered an advanced stage of intelligence. Boston Consulting Group predicts that the size of China's autonomous driving market will reach 450 billion yuan by 2025, becoming a core area of competition for the automotive and technology industries. So, how is autonomous driving defined? How is it different from the increasingly popular intelligent driving nowadays? According to the "Classification of Automotive Driving Automation" (GB/T 40429-2021) implemented in 2022, driving automation is divided into 6 levels. Among them, L0 to L2 levels are driving assistance, and the "intelligent driving equality" proposed by many car companies at present belongs to the L2 level category, where drivers need to monitor their driving throughout the entire process. Starting from the L3 level that Beijing has opened up this time, it has entered the category of autonomous driving, which belongs to conditional autonomous driving and requires the driver to take over the vehicle in emergency situations; L4 and L5 levels respectively represent highly autonomous driving and fully autonomous driving, where vehicles can operate autonomously in almost all scenarios.

At present, there are two main technological routes in the field of autonomous driving: laser radar solutions and pure vision solutions. Among them, the LiDAR scheme generates high-precision 3D environmental maps by emitting laser beams and receiving reflected signals, and combines them with other sensors (such as cameras and radars) to form a multi-sensor fusion perception system. The advantages lie in reliability, high-precision perception, and all-weather performance; The disadvantage is also obvious, which is the high cost. In addition, LiDAR typically has a large volume and high power consumption, which places high demands on product design and energy management. In contrast, pure visual solutions mainly rely on cameras as core sensors, processing the images captured by the cameras through computer vision algorithms, identifying objects such as vehicles, pedestrians, traffic signs, and traffic lights on the road, and making driving decisions based on this information. The core lies in deep learning and neural networks, which gradually improve the system's ability to understand complex environments through extensive data training. The advantages lie in lower cost, mature technology, and abundant data, while the disadvantages include greater environmental impact, limited accuracy, and high demand for computing resources.

Autonomous driving welcomes' dawn '
The landing of L3 level autonomous driving is approaching, ushering in a new "turning point"

In recent years, China's policy support in the field of autonomous driving has been continuously increasing. From national strategic planning to local government implementation rules, the regulatory environment for autonomous driving is gradually improving. As early as 2015, the Ministry of Industry and Information Technology approved the establishment of China's first national intelligent connected vehicle (Shanghai) pilot demonstration zone in Shanghai International Automobile City, marking the official launch of China's autonomous driving road pilot. Subsequently, a series of policies were issued, such as the Specification for Road Test Management of Intelligent Connected Vehicles (Trial), the Notice on Carrying out Pilot Work of Access and Road Access of Intelligent Connected Vehicles, the Guide to Transport Safety Services for autonomous vehicle (Trial), and the Notice on Carrying out Pilot Work of "Car Road Cloud Integration" Application of Intelligent Connected Vehicles, providing a clear and broad market prospect for the development of the automatic driving industry.

Nowadays, under the guidance of national policies, local governments have begun legislative attempts towards autonomous driving. As the first local regulation in the field of autonomous driving in China, the upcoming autonomous vehicle Regulations on Autonomous Vehicles clearly encourages and supports the technological innovation and industrial development of autonomous vehicle, which is meaningful for the development of the driving industry. It not only provides a legal basis for the legal launch of L3 autonomous vehicle, but also promotes the commercialization of autonomous driving technology, infrastructure construction, data security management and other aspects of development.

According to the requirements of the Regulations of autonomous vehicle in Beijing, the whole process of automatic driving innovative application activities was standardized, including the handling process of road testing and demonstration activities, safety assessment system, and the conditions and procedures for applying for road application pilot activities. At the same time, clear regulations have been made in terms of security protection, such as the job responsibilities of security officers, platform security monitoring personnel, network security and data security management, etc. In addition, insurance institutions are encouraged to develop exclusive insurance products that conform to the characteristics of autonomous vehicle and protect the rights and interests of relevant enterprises, drivers and pedestrians.

In addition to the support of national policies and local laws and regulations, the gradual increase in consumer acceptance has also laid an important foundation for the commercialization of autonomous driving in the future. In fact, with the continuous development of autonomous driving technology and the promotion of pilot projects for intelligent connected vehicles across the country, consumers' trust in autonomous driving has increased, especially in certain scenarios such as highways and closed parks. The convenience and safety of autonomous driving have been preliminarily verified. According to the "Report on the Development of Automotive Intelligence (2024) Intelligent Driving", intelligence has become one of the most important considerations for Chinese consumers when buying a car, with 90% of consumers willing to pay extra for advanced intelligent driving services. In addition, in February 2025, BYD's "Creating the Era of Intelligent Driving for All" not only lowers the threshold for intelligent driving, but also allows more consumers to understand what intelligent driving is, laying the groundwork for the next stage of popularization of autonomous driving.

Large scale mass production still faces multiple challenges
The landing of L3 level autonomous driving is approaching, ushering in a new "turning point"

Although the current national policies and market level provide strong support for the development of autonomous driving, there are still many challenges to achieve large-scale production and implementation.

1. Laws and regulations lag behind, and it is difficult to distinguish the rights and responsibilities of accidents

As is well known, it is the problem of outdated laws and regulations. On the one hand, the legal identity of autonomous vehicle has not yet been clarified, but has been recognized in pilot cities such as Beijing. At present, there is no national legal support; On the other hand, there is currently no unified legal framework for autonomous driving. Once a traffic accident occurs, it is difficult to clearly define responsibilities due to the involvement of multiple parties such as drivers, car manufacturers, software suppliers, etc. Previously, local regulations for autonomous vehicles have been issued in Shenzhen, Guangzhou, Changsha and other places, but none of them have made clear provisions on the definition of accident liability that users are most concerned about. The upcoming autonomous vehicle autonomous vehicle Regulations cannot solve this problem, but only stipulates that "when autonomous vehicle are on the road, if they violate road traffic safety laws and regulations or have traffic accidents, the traffic management department of the public security organ shall investigate and deal with them in accordance with relevant national regulations." In this regard, some experts said that the responsibility for traffic accidents belongs to the national legislative authority, and local legislations such as Beijing have no right to make provisions. Who is responsible for accidents involving L3 level autonomous driving A related survey has been conducted on this topic, and over 80% of netizens believe that car companies should bear the responsibility. Referring to the current autonomous driving regulations in Germany, it is allowed for L3 level autonomous driving to be on the road, and it is clear that the responsibility is temporarily transferred to the vehicle manufacturer during system operation, but the driver still needs to have the ability to take over. Regulations generally stipulate the time required for the driver to take over, such as within 10 seconds, otherwise the driver may bear some responsibility.

In addition to outdated regulations, determining whether a vehicle is in autonomous driving mode when an accident occurs is also a challenge. Previously, when the Ideal L9 (configuration | inquiry) was launched in 2022, it was equipped with five blue-green warning lights (ADS warning lights) at the front and rear, which were used to communicate to the outside world whether the vehicle was in intelligent driving mode. Subsequently, BYD and Xiaopeng also added similar marker lights to their newly launched models. However, due to the absence of mandatory national standards, most vehicles with advanced intelligent driving functions are not equipped with them.

2. Insufficient perception and decision-making ability of technological bottlenecks

Despite significant progress in autonomous driving technology, its application in extreme weather and complex road conditions still faces significant challenges. The existing sensor systems, including laser radar, visual camera, millimeter wave radar and other core sensing equipment, have not yet met the requirements of commercial deployment in terms of sensing accuracy and reliability when dealing with severe weather conditions such as fog and rainstorm. More importantly, autonomous driving algorithms still lack stability and accuracy in decision-making when facing sudden traffic events (such as traffic accidents, road construction, and other dynamic scenarios), which directly affects the safety and reliability of autonomous vehicles. According to complaint data from the Car Quality Network, the number of complaints about "driving system failures" has remained at a high level since 2020, with over a thousand cases reported annually in 2023 and 2024, reflecting the limitations of current intelligent driving functions.

According to feedback from Mr. Liu from Shenzhen to Chezhiwang, the SUV he purchased from a certain domestic new force brand had NOA assisted driving activated during driving. In the absence of other obstacles on the road ahead, the system misjudged the road conditions, causing the vehicle to lose control and collide with the green belt.

In addition, the first batch of Tesla test car owners were found to have frequent violations, such as changing lanes by pressing solid lines, running red lights, and walking on bus lanes, reflecting deficiencies in the system's perception, decision-making, and control modules. Especially in the complex traffic environment in China, sensor data may be interfered with or obstructed, leading to decision-making errors. Especially for Tesla's pure visual driving solution, it is more susceptible to external environmental influences. More importantly, at present, domestic autonomous driving relies on single vehicle intelligence, and the level of collaboration between vehicles, vehicles, roads, and vehicles and humans is far from meeting the requirements.

3. High cost limits commercialization process

In addition to regulatory and technical challenges, high costs are also one of the unresolved difficulties faced by autonomous driving. Vehicles with autonomous driving capabilities often need to be equipped with high-precision sensors, and in order to ensure accuracy in decision-making and execution, they also need to run advanced and complex algorithms, which require strong computing power support, making the cost of a single vehicle much higher than that of traditional vehicles. Taking Tesla's refreshed Model Y (configuration | inquiry) as an example, the optional FSD intelligent assisted driving function costs 64000 yuan, which is close to 1/4 of the entry-level model price. Of course, there are also car companies that do not charge for advanced intelligent driving functions, such as Ideal, but their models equipped with AD Max system are generally priced at over 300000 yuan. In addition, the complex scenario simulations required for large-scale testing and validation further increase R&D and operational costs, limiting the ability of enterprises to popularize them.

4. There is a risk of personal privacy leakage

Autonomous vehicles rely on a large number of sensors (such as cameras, LiDAR, GPS, etc.) to perceive the surrounding environment, which continuously collect sensitive information including vehicle location, driving trajectory, and passenger behavior inside the vehicle. The in car cameras and microphones may record passengers' conversations, facial expressions, and even biometric data (such as fatigue status), and the collection of these data may exceed user expectations. Once hacked, data leakage not only violates personal privacy, but may also be used for intelligence gathering or cyber attacks, endangering national security.

The true implementation of autonomous driving requires multiple efforts

It cannot be denied that fully autonomous driving is the future development trend, but at this stage, how to accelerate the mass production and commercial process of autonomous driving still requires joint efforts from multiple parties.

First of all, at the national level, a unified legal system for autonomous vehicle should be introduced as soon as possible to clarify the road rules for autonomous vehicle, the criteria for identifying accident liability and the subject of compensation, and improve supporting laws such as network security and data security. At the same time, by formulating national mandatory standards, automobile companies are encouraged to equip their new products with ADS marker lights as standard, providing strong evidence for determining responsibility in the later stages of accidents. Here, we also call on the Ministry of Industry and Information Technology to include the installation of ADS marker lights as a new vehicle access condition in the "Management Measures for Road Motor Vehicle Production Enterprises and Product Access Permits", as one of the safety inspection items that vehicles must pass before being launched.

Next is at the enterprise level. The rapid development of autonomous driving technology not only relies on technological breakthroughs, but also requires commercial implementation through cost reduction and innovative business models. In this regard, enterprises should continue to increase research and development investment, promote the large-scale production of hardware devices such as sensors (such as LiDAR, cameras) and smart driving chips, and further reduce costs. In addition, a large number of AI and simulation training are carried out for the auto drive system to optimize the performance of the auto drive system in complex scenes (such as extreme weather, sudden traffic events).

Another thing is autonomous driving insurance. Currently, although there are some smart driving insurance products on the market, they are mostly dominated by automotive companies and only targeted at their own brands, with certain limitations. Insurance companies should actively develop exclusive insurance applicable to autonomous vehicle, including compulsory traffic insurance, commercial insurance, tripartite insurance, etc., reduce the threshold for the promotion of autonomous vehicle, and protect the rights and interests of drivers, passengers and pedestrians.

For consumers, it is necessary to first clarify the differences and applicable scenarios between intelligent driving, autonomous driving, and unmanned driving. Through test driving experiences, they can enhance their understanding of autonomous driving technology and make rational consumption based on their actual needs.

Summary

It is widely believed in the industry that 2025 may be the first year of large-scale L3 automatic driving, and the autonomous vehicle Autonomous Vehicle Regulations, which was first implemented in Beijing, is undoubtedly a new "inflection point" in the development process of automatic driving, providing a model for other cities and regions to use for reference. It can be foreseen that with the legalization of L3 level private cars on the road, the industry is about to usher in a new round of reshuffling. Against the backdrop of policy paving and increased consumer acceptance, competition among major car companies will also enter a new stage. Whoever can turn technological leadership into commercial success is the one who stays on the table.