# Wiki-GRx-Sim

**Repository Path**: FourierIntelligence/wiki-grx-sim

## Basic Information

- **Project Name**: Wiki-GRx-Sim
- **Description**: An environment used to train GRx in NVIDIA's Isaac Sim
- **Primary Language**: Python
- **License**: BSD-3-Clause
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2024-06-03
- **Last Updated**: 2024-08-30

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Wiki-GRx-Sim

<img src="./pictures/GR1T1.png" width="300" />

This repository provides an environment used to train GRx to walk on rough terrain using NVIDIA's Orbit, Isaac Sim and rsl_rl libraries from Legged Robotics @ ETH Zürich.

### Useful Links

* Orbit: https://github.com/isaac-sim/IsaacLab(name is changed to Isaac Lab)
* Isaac Sim: https://developer.nvidia.com/isaac-sim
* rsl_rl: https://github.com/leggedrobotics/rsl_rl.git

### Installation

0. Install Ubuntu 20.04:
    - The suggest version is Ubuntu 20.04, because Orbit recommends running on Ubuntu 20.04.
    - Official Website：https://releases.ubuntu.com/focal/
    - Installation Guidance：https://ubuntu.com/tutorials/install-ubuntu-desktop#1-overview

1. Install Nvidia Driver:
    - Install Nvidia driver using the Software & Updates application that comes with Ubuntu 20.04.
    - Make sure you can see the GPU information and CUDA information by using the command line `nvidia-smi` in the terminal. As shown in the example below:

    ```
    +-----------------------------------------------------------------------------+
    | NVIDIA-SMI 525.125.06   Driver Version: 525.125.06   CUDA Version: 12.0     |
    |-------------------------------+----------------------+----------------------+
    | GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
    | Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
    |                               |                      |               MIG M. |
    |===============================+======================+======================|
    |   0  NVIDIA GeForce ...  Off  | 00000000:01:00.0  On |                  Off |
    |  0%   42C    P8    25W / 450W |    709MiB / 24564MiB |      1%      Default |
    |                               |                      |                  N/A |
    +-------------------------------+----------------------+----------------------+
    ```

2. Install Anaconda:
    - Official Website: https://www.anaconda.com/products/distribution
    - Installation: https://www.anaconda.com/download/

3. Install Isaac Sim:
    - Version 2023.1.1 is recommended
    - Official Website: https://developer.nvidia.com/isaac/sim
    - Installation: https://docs.omniverse.nvidia.com/isaacsim/latest/installation/index.html

4. Install Orbit:
    - create symbolic link to Isaac Sim
    ```bash
    export ISAACSIM_PATH="${HOME}/.local/share/ov/pkg/isaac-sim-2023.1.1"
    ln -s ${ISAACSIM_PATH} _isaac_sim
    ```
    - create conda environment
    ```bash
    cd ./Wiki-GRx-Sim
    ./orbit.sh --conda  #default name is orbit

    ./orbit.sh --conda my_env  #your custom env name
    ```
    - install requirements
    ```bash
    conda activate orbit
    sudo apt install cmake build-essential
    ./orbit.sh --install
    ```

5. Install rsl-rl:
    ```bash
    cd ./Wiki-GRx-Sim/rsl-rl
    pip install -e .
    ```

6. Install fourier_tasks:
   ```bash
   cd ./Wiki-GRx-Sim/fourier_tasks
   pip install -e .
   ```

7. Install extra packages:
   ```bash
   pip install usd-core==23.2
   pip install botocore==1.19.52
   ```

8. Install VSCode
    - Official Website: https://code.visualstudio.com/
    - VSCode configuration
        - Open Wiki-GRx-Sim library in VSCode
        - Press Ctrl+Shift+P, choose Tasks: Run Task, select setup_python_env
        - after configuration, you can use the Go to Definition, Intellisense, etc. features of VSCode

### Training and Inference

0. Training example:
    ```bash
    ./orbit.sh -p source/standalone/workflows/rsl_rl/train.py --task Isaac-Walking-GR1-v1
    ```

1. Inference example:
    ```bash
    ./orbit.sh -p source/standalone/workflows/rsl_rl/play.py --task Isaac-Walking-GR1-Play-v1
    ```


### Notice

The training code here only shows how to control the robot's leg to walk, and the robot body is set fixed.
If you want to control the robot body to move, you need to modify the following files:

- urdf file: `./fourier_tasks/fourier_tasks/resources/robots/GR1/urdf/urdf/GR1T1.urdf`
- config file: `./fourier_tasks/fourier_tasks/resources/config/GR1/gr1_cfg.py`


Thank you for your interest in the Fourier Intelligence GRx Robot Model Repository.
We hope you find this resource helpful in your robotics projects!