Booster Deploy

An easy-to-use deployment framework that enables seamlessly running the same policy code in both simulation and on real robots.

Booster Deploy is a lightweight deployment framework that supports running control policies on Booster robots (sim2real), MuJoCo (sim2sim), and Webots (internal sim2sim). The system adopts many well-established designs from IsaacLab to provide modular abstractions, allowing unified policy execution across both simulated and real robotic platforms.

Add and list tasks: 1. Create a subfolder under tasks/ for your task. 2. Implement a Policy/PolicyCfg and provide a ControllerCfg referencing the policy. 3. Place policy checkpoints under models/ and reference the path in the config. 4. Register your ControllerCfg config in the task registry (see existing tasks for the registration pattern). 5. Check all available tasks: bash python3 scripts/deploy.py --list

- Download and install BoosterAssets: - Clone the boosterassets which contains Booster robot models and resources. - Install boosterassets python helper following the instructions in the repository.

- Install Python dependencies on local machine: pip install -r requirements.txt

- Launch the task in mujoco: bash python scripts/deploy.py --task <TASKNAME> --mujoco

IMPORTANT: Make sure to install Booster Firmware >= v1.4 on the robot before proceeding.

NOTE: If you plan to deploy on the T1 Standard Edition robot, you need to choose to deploy on the motion board rather than the perception board.

- After you finish testing your task with Sim2Sim locally, copy the project to the robot.

- Install Booster Robotic SDK on robot: - Clone the latest Booster Robotics SDK repository into the robot. - Follow the build instructions in the SDK repository. - Important: Make sure to build and install the Python bindings: bash cd boosterroboticssdk mkdir build && cd build cmake .. -DBUILDPYTHONBINDING=ON make -j$(nproc) sudo make install

- Install Python dependencies on the robot: pip install -r requirements.txt