Agent Network · ANet4D ANet4D

First steps

  • Overview
    • Positioning
    • Current version
    • Capabilities and delivery status
    • Documentation map
  • Installation and requirements
    • Server
    • Browser
    • Ports
    • Directories
    • Environment variables
    • Run profiles
    • Optional data: the six UrbanScene3D cities
    • Access from another machine
    • Environment diagnostics
    • Building this documentation
  • Quick start
    • Prerequisites
    • Get the source: git clone
    • Install dependencies: make setup
      • Warnings in the output
    • Generate the synthetic demo city: make demo-world
    • Start the shared show: make run
    • Confirm the service is ready: make status
    • Open it in a browser
    • Stop: make stop
    • Start the sandbox service: AWR_PROFILE=sandbox make run
    • Query the sandbox interface
    • Drive a session with the Python SDK
    • Stop the sandbox service
    • Troubleshooting
      • Startup fails with a busy port
      • sim-core is repeatedly reported as hung
      • Adding a drone returns 110 PARAM_OUT_OF_RANGE
      • Adding a drone returns 525 NEST_INCOMPATIBLE
      • Creating a session returns 409 SANDBOX_EXISTS
      • httpx.ReadTimeout while queueing
      • Messages about the six cities failing to build
    • What you have after this step, and what you do not

User guide

  • Shared show
    • What runs
    • Available interactions
    • The other six worlds: each city's own simulated monitoring demonstration
    • Viewer seats and queueing
    • Rendering and quality
  • Sandbox
    • Opening a sandbox
    • Scenario templates
    • Interface
    • Session limits
    • Clock and rate
    • Manual control
    • Snapshots, reset and world switching
    • Lifetime, renewal and power saving
    • Capacity
  • Targets and perception
    • Classes
    • Discovery criterion
    • Sensors
    • Detection, recognition and identification
    • Perception quantification and field views
    • Related endpoints
  • Tasks, nests and relay
    • Task types
    • Allocation preview
    • GSD scan planning
    • Nests
    • Allocators
    • 7x24 relay monitoring
      • Trigger and stations
      • Early dispatch and handover
      • Statistics and sustainability assessment
  • Phones and tablets
    • Device classes
    • Tablet: landscape commander
    • Phone: five tabs
    • Touch gestures
    • Quality
    • Full screen and home-screen launch
  • Internal identity entry
    • Location
    • Purpose
    • Security conventions
  • ANet4D Python SDK quickstart
    • 0. Measured runs
    • 1. Prerequisites
    • 2. Install the SDK
    • 3. Write an agent
    • 4. Run the example agent
      • 4.1 Platform side: start the environment and wait for agents
      • 4.2 Researcher side: connect the agents
      • 4.3 In-process run
    • 5. Switch the physics tier and the observation level
    • 6. Baseline policies and the reference detector
    • 7. Package the agent as an image
    • 8. Build the wheel and the sdist from source
    • 9. Troubleshooting
    • 10. After this guide

Open interface

  • Open interface
    • Authentication
    • Error format
    • Idempotency and rate limits
    • Realtime protocol
    • Sandbox REST v1 endpoints
    • Local runtime API endpoints
  • Python SDK
    • Installation
    • Minimal example
    • Object model
    • Call reference
    • Error handling
    • Queueing, renewal and power saving
    • Manual control example
    • Example programs

System and data

  • Architecture and coordinate frames
    • Overall structure
    • Data flow of one interaction
    • Clocks
    • Coordinate frames
    • Contracts
  • Worlds and data
    • World Package v1
    • The seven worlds
    • ANet Synthetic City
    • UrbanScene3D
    • Future data sources
  • Deployment
    • Topology
    • Resource limits
    • Deployment procedure
    • Constraints
    • Operations tools
    • Publishing this documentation site

All the rest

  • Versions and roadmap
    • Released
      • V0.1 (D1): a World Runtime with every layer in place
      • V0.2-demo (D2 and D2.1): interactive sandbox and open interface
      • V0.3 (D3): Event Lab
      • V0.3.2: ANet4D
      • V0.3.3: Event Lab samples
    • D3 design record: event simulation and active observation benchmark
      • Design highlights
      • Phases
    • Forward direction: D4 collaborative experimentation platform
    • Original version themes (design baseline)
  • Acceptance summary
    • Results
    • Fixes confirmed
    • Failed items and root causes
    • Test conditions
    • Full reports
  • Frequently asked questions
    • Naming
    • Runtime environment
    • Worlds and data
    • Shared show and sandbox
    • Interface and SDK
    • Feedback
  • Citation and license
    • Citing ANet4D
    • Citing UrbanScene3D
    • ANet research
    • License
    • Marks
    • Acknowledgments
Agent Network ANet4D
  • Search


© Copyright 2026, Agent Network Research. Last updated on 2026-10-26.

Built with Sphinx using a theme provided by Read the Docs.
ANet4D en · latest
Languages
zh-cn
en
Versions
latest
Project
GitHub
Live demo
Releases

Hosted at docs.agentnetwork.org.cn/anet4d