Autonomous Planetary Rover Training Platform
RAVEN is a high-fidelity simulation and reinforcement learning platform for training autonomous rovers to navigate extraterrestrial and off-road terrain. Built by Arca.
The backend is now wired to default to isaac mode. No machine-specific Isaac Sim path is hardcoded in the repo. Set ISAAC_SIM_PATH at runtime on the machine that is running Isaac Sim.
The current Isaac integration is startup and health-check wiring, not full live telemetry. Until a real Isaac state provider is registered, the app should remain idle instead of showing fake rover data.
cd bridge
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install -r requirements.txt
python server.pyImportant:
- Use Python
3.12for the bridge environment. Python3.14currently breakspydantic-coreduring install on Windows. - The bridge defaults to
isaacmode. - Set
ISAAC_SIM_PATHon the Windows machine before starting the backend if Isaac Sim is not installed in the default location. - Use
RAVEN_SIM_MODE=mockonly if you explicitly want the legacy mock backend.
cd app
npm install
npm run tauri devFor the Windows machine test flow, see docs/windows-test-flow.md.
See instructions.md for the full project specification and development roadmap.