Autonomous Driving

Core Operating Principles and Technical Characteristics

An automotive Autonomous Driving System is an advanced intelligent system that substitutes human input for environmental monitoring, path planning, and dynamic driving tasks under designated scenarios or across all-scenario conditions. This is achieved through the synergy of on-board multi-sensor fusion perception, high-computing-power heterogeneous domain controllers, AI deep learning algorithms, and full-domain chassis by-wire technology. Its core operational logic adheres to a closed-loop "Perception-Planning-Control" workflow: the vehicle deploys multi-source heterogeneous sensors—including LiDAR, millimetre-wave radar, high-definition (HD) cameras, and ultrasonic sensors—to capture 3D environmental data round the clock; HD maps and Simultaneous Localisation and Mapping (SLAM) establish real-time spatial coordinates; AI algorithms and end-to-end large models handle trajectory prediction and decision-making for dynamic and static obstacles; and finally, steer-by-wire, brake-by-wire, and drive-by-wire systems execute steering, throttle, braking, and gear-shift inputs with precision.

In terms of technical characteristics, autonomous driving is classified from Level 0 to Level 5 based on Society of Automotive Engineers (SAE) standards. Modern advanced autonomous driving systems (such as Level 2+ to Level 4) exhibit the following core attributes:

Multi-sensor fusion and redundant architecture: Complementing camera vision with LiDAR overcomes the physical limitations of standalone sensors, whilst dual-redundancy power supplies, steering, and braking systems ensure safe handover in the event of a single-point failure. 

Data-driven and end-to-end large models: Moving beyond conventional rule-based programming, neural networks are trained on vast datasets of real-world driving data to deliver human-like decision-making and navigation through complex traffic conditions. 

Global Market Applications and Evolution

Representing the vital "next phase" of vehicle intelligence following the global electrification transition, autonomous driving has achieved commercial deployment and scale across mid-to-high-end new energy vehicles (NEVs), Robotaxi trial zones, and dedicated logistics parks. Global regulatory frameworks and traffic legislation (such as the UN WP.29 Automated Lane Keeping System regulation and national Level 3 type-approval policies) are actively accelerating its legal adoption.

In terms of technical evolution, autonomous driving is advancing rapidly towards "mapless solutions (perception-heavy, map-light), End-to-End AI model architectures, Vehicle-to-Everything (V2X) vehicle-road-cloud integration, and progression towards Level 4 high/full autonomy." Modern autonomous driving has evolved from rudimentary single-lane keeping assist into a disruptive core technology transforming personal mobility and the broader transport landscape.

Operational Guidelines and Safety Limitations 

The demarcation of liability and operational responsibility governing each autonomous driving level must be strictly adhered to. For Level 2 and Level 2+ advanced driver assistance systems (ADAS), the driver remains fully responsible for the vehicle at all times, requiring constant attention and hands on the steering wheel, with distractions, sleeping, or using electronic devices strictly prohibited. For Level 3 systems and above, the driver must be prepared to promptly resume manual control within the designated handover window whenever the system indicates operational limit exceedance or issues a takeover request. 

Under extreme inclement weather (such as torrential downpours, heavy snow, or dense fog), unlit highway stretches at night with no lead vehicle, situations where sensors are blocked by dirt, mud, or ice, or non-standard road layouts like construction diversions, the perception and decision-making performance of autonomous driving systems will degrade substantially or fail entirely. Under these conditions, engaging or relying on autonomous driving is strictly prohibited, and the driver must immediately revert to manual control.