High‑Definition Map Data Distribution Engine: Architecture, Models, and Applications for ADAS

High‑Definition (HD) maps provide centimeter‑level absolute coordinate accuracy and rich static traffic information, far surpassing ordinary navigation maps that are limited to about 10 m precision and basic road geometry. HD maps describe not only road links but also individual lanes, lane markings, landmarks, and traffic signs, enabling machine‑readable perception for autonomous vehicles.

Advanced Driver Assistance Systems (ADAS) require precise front‑road network data for decision‑making. While ordinary digital maps serve navigation, HD maps support multiple ADAS functions, necessitating a High‑Precision Data Distribution Engine (AHP) to deliver map data to vehicle‑mounted ADAS modules. The ADAS Interface Specification (ADASIS) defines the "ADAS electronic horizon" concept, which is transmitted over in‑vehicle Ethernet.

The distribution engine is organized into three layers: the engine layer (loading, parsing, and organizing HD map data), the protocol layer (assembling messages for transmission), and the adaptation layer (interfacing with vehicle systems and delivering data to ADAS applications). Core components include AHP (the engine), AHR (reconstructor), and the ADASIS V3 protocol.

Road‑network modeling uses a three‑tier abstraction: a real‑world model, a high‑precision road‑network model, and a tree‑structured representation expressed by Path and Offset . Paths group links, and offsets (centimeter units) locate attributes along a path. Attribute models are classified as Spot (valid at a single offset), Step (valid from an offset to the next), and Linear (linearly interpolated between offsets), enabling representation of speed limits, traffic lights, and weather‑dependent restrictions.

Synchronization between AHP and AHR relies on pathControl messages to add, delete, or retain paths, and profileControl messages to update attribute data. Interaction mechanisms include broadcast, request/response, and publish/subscribe modes, with the current implementation using a request/response pattern where AHP sends ADAS messages and AHR can request or acknowledge data.

Integration can occur in three typical forms: (1) a map‑box solution that combines HD map, localization, OTA, and base software/hardware in a single ECU; (2) integration within the Infotainment Head Unit (IHU) to reduce hardware cost; and (3) embedding the engine in a domain controller to minimize cross‑domain network bandwidth and suit OEMs pursuing full‑stack solutions. Each form offers distinct advantages in task allocation, safety isolation, and scalability.

Quality assurance employs unit testing, functional testing, and specialized quality‑inspection tools, complemented by visualization utilities that display map data and engine status.

Typical application scenarios include high‑precision lateral and longitudinal positioning, highway autonomous driving (HWP) with lane‑type and lane‑line verification, and navigation‑based cruise control that checks road class, continuous route identifiers, and weather conditions before activation.

Future work aims to further integrate AHP V2/V3 architectures, enhance data‑closed‑loop capabilities, and expand data provision and recovery mechanisms.