Skip to content

unitree_setup.md

Latest commit

 

History

History
113 lines (81 loc) · 4.07 KB

unitree_setup.md

File metadata and controls

113 lines (81 loc) · 4.07 KB

Setup for Unitree Robots

RoboJuDo has two environments for unitree robots:

  • UnitreeEnv: Based on unitree_sdk2py.
    • Support both UnitreeH1 and UnitreeG1.
    • May endure performance issues on UnitreeG1 due to limited computing power.
  • UnitreeCppEnv: Based on UnitreeCpp
    • Support UnitreeG1.
    • Can be deployed onboard UnitreeG1 pc2. It is much faster and stable.

🛠️SDK Installation

Choose the above Env&SDK options you want to use.

UnitreeEnv

  1. Follow the instuction to install the official SDK: unitree_sdk2py
  2. verify with 
    python -c "from robojudo.environment import UnitreeEnv"
    if any error, check installation of unitree_sdk2py

UnitreeCppEnv

As the setup of UnitreeCpp :

  1. Install the Unitree official C++ SDK: unitree_sdk2

    Note: It is recommended to use the default installation path.

  2. and then install the unitree_cpp package: For simple setup, unitree_cpp is included in RoboJuDo, just run: 
    python submodule_install.py unitree_cpp

    ⭐ make sure unitree_sdk2 is installed first

  3. verify with 
    python -c "from robojudo.environment import UnitreeCppEnv"
    if any error, check installation of unitree_sdk2 and then unitree_cpp.

🤖Deploy Guide

We provide two deployment options:

  1. Deploy the policy on the real robot.
  2. Deploy the policy on your workstation and control the robot via a wired Ethernet connection.

Deploy on Unitree Robot

Run the policy directly on the robot.

Setup

  1. Clone our repository and setup the environment on the robot’s onboard computer: see Basic Setup.
  2. check 🛠️SDK Installation for Unitree SDK setup.

Since the G1 has limited computing resources, you need to run UnitreeCppEnv.

Usually, the robot's network interface is eth0 (for G1). You don't need to modify the config. If you find it doesn't work. see network configuration for help

Deploy from Your Computer

Run the policy on your computer and control the robot via Ethernet.

Both UnitreeEnv and UnitreeCppEnv are supported.

After setup install, connect your robot via Ethernet, then check network configuration.

Network Configuration

Refer to official guide to connect and find the robot's network interface.

Then edit config to update the env_type and net_if accordingly:

For example:

Option 1: Open g1_cfg.py and modify the g1_real config.

class g1_real(g1):
    env: G1RealEnvCfg = G1RealEnvCfg(
        env_type="UnitreeEnv",  # For unitree_sdk2py
        # env_type="UnitreeCppEnv", # For unitree_cpp, check README for more details
        unitree=UnitreeEnvCfg.UnitreeCfg(
            net_if="eth0",  # note: change to your network interface
            robot="g1",
            msg_type="hg",
        ),
    )

Option 2: Go to g1_real_env_cfg.py and update the global config for g1.

class G1RealEnvCfg(G1EnvCfg, UnitreeEnvCfg):
    env_type: str = UnitreeEnvCfg.model_fields["env_type"].default
    # ====== ENV CONFIGURATION ======
    unitree: UnitreeEnvCfg.UnitreeCfg = UnitreeEnvCfg.UnitreeCfg(
        net_if="eth0", # EDIT HERE
        robot="g1",
        msg_type="hg",
    )