How AI Powers Gaode’s ‘Car Cruise’ for Seamless, Hands‑Free Navigation

Overview

Gaode Map has upgraded its navigation to an AI‑native application called Car Cruise , which aims to deliver a navigation experience beyond traditional turn‑by‑turn guidance. The core module provides an “invisible” driving assistant that automatically awakens, recognizes the current route, and offers precise voice prompts, traffic warnings, and speed limits, minimizing driver interaction.

Key Differences from Traditional Navigation

1. Launch method: Traditional navigation requires manual destination entry, while Car Cruise starts automatically as soon as the driver gets in the car.

2. Intelligence level: Traditional navigation offers fixed route guidance; Car Cruise leverages AI to deliver personalized, predictive assistance.

3. Interaction mode: Traditional apps rely on touch, whereas Car Cruise emphasizes voice interaction and automatic activation to reduce driver distraction.

Product Scope

Car Cruise covers every stage of a driving journey, from default phone‑based cruising to car‑play projection and end‑of‑trip parking assistance, providing a seamless transition between scenarios.

Core Components

Smart voice broadcast & mute

– real‑time voice guidance with one‑tap mute. Driving preference settings – quick access to map style, volume, day/night mode. Vehicle‑heading switch – toggle between heading‑up and north‑up map orientation. Real‑time speedometer – displays current speed. Lane line & road name display – shows lane guidance and road names. Easy exit – a clear button to stop cruising at any time. Lane‑level navigation toggle – enables fine‑grained lane guidance in high‑precision map areas.

User Benefits

By reducing manual operations, Car Cruise significantly improves driving safety, optimizes efficiency through automatic route planning and parking assistance, and offers a personalized, frictionless experience.

Technical Architecture

The system follows a state‑driven design, integrating a cross‑platform framework with a finite‑state‑machine (FSM) that adheres to SOLID principles, achieving high cohesion, low coupling, and modularity.

Layer Overview

Application entry layer ( CarCruisePage ) – page entry, lifecycle management, and rendering.

Business orchestration layer ( CarCruiseStateController ) – core scheduler that drives the StateMachine based on CarCruiseStateConfig, exposing methods such as toDefaultState(), toProjectionState(), mapStateToProps(), and getStateView().

State machine & config layer ( StateMachine , CarCruiseStateConfig ) – registers and switches states; config defines lazy‑loaded state blueprints.

Scene state implementations ( LazyStateClasses ) – concrete classes like CarCruiseInitialState, CarCruiseDefaultState, CarCruiseProjectionState, CarCruiseEndPoiParkState that render specific UI components.

Data model ( CarCruiseModel ) – persists core data (e.g., machineStateName) with keepAlive to maintain continuity across state transitions.

UI rendering ( ViewRenderComponents ) – component set driven by state classes and finally presented by CarCruisePage.

State Machine Design

The state machine abstracts complex driving scenarios into distinct states (initial, default, projection, end‑poi‑park). Events trigger legal transitions, ensuring predictable behavior and easy extensibility.

Key States

initialState

– determines environment and jumps to the appropriate first state. defaultState – standard phone‑based cruising. projectionState – car‑play / Hicar / Ucar projection with UI and logic tailored for in‑vehicle screens. endPoiParkState – parking assistance after reaching the destination.

Car‑Play Integration

When a car‑play device is detected, the controller switches to projectionState, loading car‑specific widgets (e.g., HONOR_CAR_WIDGET_KEY_LIST) and adjusting UI size, layout, and interaction patterns for safe in‑vehicle use. Native APIs ( natives) and modules like CruiseCarplayLogic.js handle communication with the vehicle’s OS.

Data‑Driven Intelligence

Traffic lights and electronic eye warnings are fetched from multi‑road data sources, merged with real‑time positioning, and rendered on the map to give drivers advanced foresight.

Stability Measures

Lazy loading of state classes reduces initial load time.

Extensive business telemetry and APM monitoring.

Keep‑alive mechanisms ensure GPS and navigation continuity.

Future Directions

Gaode plans to deepen integration of big data and AI to improve intent recognition, state transitions, and predictive alerts, further merging the cruising experience with the broader car‑ecosystem.