Spaces:
Sleeping
Download README.md from Ducdata1808/AGV_simulator: direct link, hf CLI and curl.
- Browser
- Download file 4.66 kB
-
https://huggingface.co/spaces/Ducdata1808/AGV_simulator/resolve/main/README.md
- Command line
-
hf download hf://spaces/Ducdata1808/AGV_simulator/README.md
-
curl -L -o README.md https://huggingface.co/spaces/Ducdata1808/AGV_simulator/resolve/main/README.md
title: AGV Simulator
emoji: π€
colorFrom: blue
colorTo: indigo
sdk: docker
app_port: 7860
pinned: false
VDA 5050 Robot Simulator
Python-based simulation environment for robots with VDA 5050 protocol (specifically targeting version 2.0.0). It allows simulating multiple AGVs (Automated Guided Vehicles), sending them orders, and visualizing their movement and status.
Demo Video
Features
- VDA 5050 v2.0.0 Compliance: Simulates key aspects of the VDA 5050 protocol including:
- Connection states (
ONLINE,OFFLINE,CONNECTION_BROKEN) - Vehicle State (
State,AgvPosition,ActionState, etc.) - Order processing (
Order,Node,Edge) - Instant Actions (
InstantActions,Action) - Visualization messages (
Visualization)
- Connection states (
- Multi-Robot Simulation: Can simulate multiple AGVs concurrently, configured via kα»ch bαΊ£n JSON.
- AI Fleet Coordination Benchmarking: Evaluates LLMs (Gemini, Llama, Qwen) in generating optimal conflict-free routes and actions under VDA 5050 constraints.
- Web UI Dashboard: Interactive frontend displaying real-time simulation runs, layout configurations, obstacle avoidance, and remaining battery levels.
- CI/CD Integration: Automated unit tests and Docker image build verification via GitHub Actions.
Directory Structure
vda5050-robot-simulator/
βββ .github/workflows/ # CI configuration (automated test & docker build)
βββ benchmark/ # Core evaluation engine (runner, metrics logger)
βββ factory_layouts/ # Warehouse maps (nodes, edges, walls, charging zones)
βββ frontend/ # Web UI visualizer dashboard (index.html)
βββ legacy_mqtt/ # Legacy visualizer and MQTT runner scripts
βββ models/ # API adapters (Gemini, Qwen Cloud, Groq, HF, OpenRouter)
βββ protocol/ # VDA 5050 protocol data structures (Python dataclasses)
βββ scenarios/ # JSON scenarios (Level 1 Basic -> Level 4 Expert)
βββ scripts/ # API connection tests and helper utilities
βββ tests/ # Automated unit test suite
βββ app.py # FastAPI backend server
βββ config.py / config.toml # Base simulator configuration files
βββ main.py # Core AGV state machine and simulator logic
βββ run_benchmark.py # CLI benchmark execution tool
βββ requirements.txt # Python dependency list
Setup
Clone & Virtual Environment:
git clone git@github.com:datascience180806/AGV_research.git cd vda5050-robot-simulator python -m venv venv source venv/bin/activate # On Windows: .\venv\Scripts\activate pip install -r requirements.txtAPI Credentials: Create a
.envfile in the root directory:# API Keys for Model Adapters (Fill in what you use) GEMINI_API_KEY=your_gemini_api_key GROQ_API_KEY=your_groq_api_key QWEN_API_KEY=your_alibaba_qwen_api_key OPENROUTER_API_KEY=your_openrouter_api_key HF_TOKEN=your_huggingface_token
Web Visualizer & Benchmark Tool
This project features a Web-based interactive visualizer dashboard and benchmark suite that runs without needing a local MQTT broker.
Start the Backend API Server:
python -m uvicorn app:app --host 127.0.0.1 --port 8000This runs the FastAPI server at
http://127.0.0.1:8000. You can access the API documentation athttp://127.0.0.1:8000/docs.Start the Frontend Client: Run a local static web server in the root directory:
python -m http.server 3000Then, open your browser and navigate to
http://localhost:3000/frontend/.Run Benchmarks via CLI: You can also run scenarios via command line:
# Run a single scenario with Gemini 2.5 Flash python run_benchmark.py --scenario scenarios/level_1_basic/scenario_001.json --model gemini-2.5-flash # Run a scenario using Qwen Max Cloud python run_benchmark.py --scenario scenarios/level_1_basic/scenario_001.json --model qwen-max
Running Tests
To run the unit tests locally:
python -m unittest discover -s tests -p "test_*.py"
Docker Deployment
To build and run the backend using Docker:
# Build the Docker image
docker build -t vda5050-robot-simulator .
# Run the container
docker run -p 7860:7860 vda5050-robot-simulator
The FastAPI backend will then be exposed at http://localhost:7860.
