City Navigate on Autopilot (CNOA)

City Navigate on Autopilot, commonly known as City NOA (Navigate on Autopilot), is an advanced driver assistance feature designed for open urban roads. It enables vehicles to autonomously navigate complex urban driving scenarios—including traffic light junctions, unprotected turns, and mixed pedestrian-vehicular traffic—under the guidance of a preset navigation route. This feature is regarded as a pivotal technological leap from highway-only assistance to full-domain point-to-point semi-autonomous driving.

Core Differences from Highway NOA

The primary difference between City NOA and Highway NOA lies in the complexity of the Operational Design Domain (ODD). Highway environments feature relatively controlled road layouts and straightforward traffic rules, whereas urban roads present dense junctions, traffic lights, pedestrians, non-motorised transport, temporary roadworks, and sudden road hazards. This places exponentially higher demands on sensor perception accuracy, decision-making algorithms, and computing power.

Technical Architecture and Evolution Trends

City NOA operates on a closed-loop "Perception-Decision-Execution" framework, with its technical roadmap undergoing significant evolution in recent years:

• From "HD Map Reliance" to "Mapless/Map-Light" Solutions: Early setups relied heavily on centimetre-level high-definition (HD) maps to supply prior road data. However, constrained by limited HD map coverage, steep updating costs, and regulatory hurdles, the industry is swiftly shifting towards "heavy perception, light map" or "mapless" approaches, achieving urban navigation by bolstering the real-time perception capabilities of on-board sensors (LiDAR, cameras, and millimetre-wave radars).

• End-to-End AI Models as the Core Driver: Traditional modular software architectures are being superseded by unified, end-to-end AI foundation models. By outputting driving trajectories directly through a single integrated model, this architecture minimises information loss between modules, delivers more natural, human-like driving behaviour, and significantly enhances the capability to handle complex junctions and corner-case scenarios.

Industry Status and Market Landscape

City NOA is transitioning from an early-adopter feature into mass-market adoption. Between January and November 2025, sales of passenger vehicles equipped with City NOA in China reached 3.129 million units, achieving a 15.1% market penetration rate, with domestic Chinese brands accounting for 81.1% of these sales.

In terms of deployment pathways, the market is powered by a dual-track strategy: in-house OEM development and third-party supplier collaborations. Momenta and Huawei together command approximately 80% of the third-party supplier market. At the same time, legacy premium marques like Mercedes-Benz and BMW are accelerating their rollout of City NOA in the Chinese market, competing through collaborations with local technology partners as well as proprietary in-house solutions.

Safety Boundaries and Regulatory Requirements

City NOA remains a Level 2+ (L2+) driver assistance system and is not fully autonomous driving. The driver remains the primary party responsible for vehicle safety at all times, and must keep their hands on the steering wheel, stay fully alert to road conditions, and be ready to take over manual control at any moment. System performance may degrade significantly in adverse conditions such as heavy downpours, thick fog, or sensor obstruction; blindly relying on this feature in severe weather or on roads with missing lane markings is strictly prohibited.