High‑Definition Map Data Distribution Engine: Concepts, Architecture, and Applications
The High‑Definition Map Data Distribution Engine (AHP) delivers centimeter‑accurate, machine‑oriented road‑network and attribute data to ADAS and autonomous‑driving modules via the ADASIS electronic horizon, using layered architecture, path/offset models, attribute interpolation, synchronized control messages, and flexible integration, testing, and future‑ready extensions.
High‑Definition (HD) maps differ from ordinary navigation maps in that they provide centimeter‑level absolute positioning accuracy and a wealth of static traffic‑related information, such as lane geometry, lane markings, landmarks, and traffic signs. While ordinary maps serve drivers, HD maps are machine‑oriented and support Advanced Driver Assistance Systems (ADAS) and autonomous driving.
Precision is a key distinction: typical navigation maps have ~10 m accuracy, whereas HD maps require 10‑20 cm accuracy to reliably locate a vehicle within a specific lane.
ADAS applications need forward road‑network and attribute data for decision‑making. The HD‑map data is therefore distributed via a High‑Precision Data Distribution Engine (referred to as AHP – ADAS Horizon Provider) that bridges HD‑map data and ADAS modules.
The ADAS Interface Specification (ADASIS) defines the concept of an “ADAS electronic horizon”, which represents the vehicle’s forward road‑network and its attributes. This horizon is expressed through a hierarchical tree of possible drivable paths, with geometric and attribute models serialized over the vehicle’s Ethernet.
The distribution engine consists of four main components: AHP (data provider), AHR (data reconstructor), the ADASIS V3 protocol, and the ADAS applications that consume the data.
Architecture layers include:
Engine layer – loading, parsing, and organizing HD‑map data.
Protocol layer – packaging data into protocol messages for transmission.
Adaptation layer – interfacing with vehicle systems and delivering data to ADAS modules.
Modeling of the road network is performed at three abstraction levels: a real‑world model, a digitized map model, and a vehicle‑centric model expressed as Paths and Offsets. Paths are collections of links (road segments) and can be represented either as simple independent Paths or as an optimized tree that reduces redundancy.
Attributes are attached to Paths using three interpolation types:
Spot – valid only at a single Offset (e.g., a traffic light).
Step – valid from an Offset up to the next Offset (e.g., speed limits).
Linear – linearly interpolated between two Offsets.
Vehicle position is expressed by Path and Offset, possibly spanning multiple Paths, and includes a timestamp indicating when sensor data was received.
Synchronization between AHP and AHR is achieved via pathControl messages (add, delete, or keep Paths) and profileControl messages for attribute updates.
Interaction mechanisms between the distribution engine and receivers include broadcast, request/response, and publish/subscribe patterns, with the current implementation using a request/response model.
Integration scenarios for the distribution engine are:
Embedded in a map box (HD‑map ECU) with dedicated hardware and software stacks.
Integrated within the Infotainment Head Unit (IHU) to reduce hardware cost.
Integrated inside a domain controller to minimize cross‑domain network bandwidth.
Quality assurance employs unit testing, functional testing, quality‑inspection tools, and visualization utilities.
Typical application examples include high‑precision localization (combining HD‑map data with sensor data), highway autonomous driving (HWP) activation based on map‑derived lane geometry and curvature, and navigation‑based cruise control that leverages map attributes such as road class, junctions, and weather‑dependent speed limits.
Future evolution aims to further merge AHP V2 and V3 architectures, enhance data‑closed‑loop capabilities, and expand data provision and recovery functions.
Amap Tech
Official Amap technology account showcasing all of Amap's technical innovations.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.