rl_sar

中文文档

This repository provides a framework for simulation verification and physical deployment of robot reinforcement learning algorithms, suitable for quadruped robots, wheeled robots, and humanoid robots. "sar" stands for "simulation and real"

feature:

• Support legged_gym based on IaacGym and IsaacLab based on IsaacSim. Use framework to distinguish.
• The code supports both cpp and python, you can find python version in src/rl_sar/scripts

Click to discuss on Discord

Preparation

Clone the code

git clone https://github.com/fan-ziqi/rl_sar.git

Dependency

This project uses ros-noetic (Ubuntu 20.04) and requires the installation of the following ROS dependency packages:

sudo apt install ros-noetic-teleop-twist-keyboard ros-noetic-controller-interface ros-noetic-gazebo-ros-control ros-noetic-joint-state-controller ros-noetic-effort-controllers ros-noetic-joint-trajectory-controller

Download and deploy libtorch at any location

cd /path/to/your/torchlib
wget https://download.pytorch.org/libtorch/cpu/libtorch-cxx11-abi-shared-with-deps-2.0.1%2Bcpu.zip
unzip libtorch-cxx11-abi-shared-with-deps-2.0.1+cpu.zip -d ./
echo 'export Torch_DIR=/path/to/your/torchlib' >> ~/.bashrc

Install yaml-cpp and lcm. If you are using Ubuntu, you can directly use the package manager for installation:

sudo apt install liblcm-dev libyaml-cpp-dev

You can also use source code installation, click to expand

Install yaml-cpp

git clone https://github.com/jbeder/yaml-cpp.git
cd yaml-cpp && mkdir build && cd build
cmake -DYAML_BUILD_SHARED_LIBS=on .. && make
sudo make install
sudo ldconfig

Install lcm

git clone https://github.com/lcm-proj/lcm.git
cd lcm && mkdir build && cd build
cmake .. && make
sudo make install
sudo ldconfig

Compilation

Compile in the root directory of the project

cd ..
catkin build

If catkin build report errors: Unable to find either executable 'empy' or Python module 'em', run catkin config -DPYTHON_EXECUTABLE=/usr/bin/python3 before catkin build

Running

Before running, copy the trained pt model file to rl_sar/src/rl_sar/models/YOUR_ROBOT_NAME, and configure the parameters in config.yaml.

Simulation

Open a terminal, launch the gazebo simulation environment

source devel/setup.bash
roslaunch rl_sar gazebo_<ROBOT>.launch

Open a new terminal, launch the control program

source devel/setup.bash
(for cpp version)    rosrun rl_sar rl_sim
(for python version) rosrun rl_sar rl_sim.py

Where <ROBOT> can be a1 or a1_isaaclab or gr1t1 or gr1t2.

Control:

• Press <Enter> to toggle simulation start/stop.
• W and S controls x-axis, A and D controls yaw, and J and L controls y-axis.
• Press <Space> to sets all control commands to zero.
• If robot falls down, press R to reset Gazebo environment.

Real Robots

Example: Unitree A1

Unitree A1 can be connected using both wireless and wired methods:

• Wireless: Connect to the Unitree starting with WIFI broadcasted by the robot (Note: Wireless connection may lead to packet loss, disconnection, or even loss of control, please ensure safety)
• Wired: Use an Ethernet cable to connect any port on the computer and the robot, configure the computer IP as 192.168.123.162, and the gateway as 255.255.255.0

Open a new terminal and start the control program

source devel/setup.bash
rosrun rl_sar rl_real_a1

Press the R2 button on the controller to switch the robot to the default standing position, press R1 to switch to RL control mode, and press L2 in any state to switch to the initial lying position. The left stick controls x-axis up and down, controls yaw left and right, and the right stick controls y-axis left and right.

Or press 0 on the keyboard to switch the robot to the default standing position, press P to switch to RL control mode, and press 1 in any state to switch to the initial lying position. WS controls x-axis, AD controls yaw, and JL controls y-axis.

Train the actuator network

1. Uncomment #define CSV_LOGGER in the top of rl_real.cpp. You can also modify the corresponding part in the simulation program to collect simulation data for testing the training process.
2. Run the control program, and the program will log all data after execution.
3. Stop the control program and start training the actuator network. Note that rl_sar/src/rl_sar/models/ is omitted before the following paths.

   rosrun rl_sar actuator_net.py --mode train --data a1/motor.csv --output a1/motor.pt

4. Verify the trained actuator network.

   rosrun rl_sar actuator_net.py --mode play --data a1/motor.csv --output a1/motor.pt

Add Your Robot

In the following text, <ROBOT> represents the name of the robot

1. Create a model package named <ROBOT>_description in the rl_sar/src/robots directory. Place the robot's URDF file in the rl_sar/src/robots/<ROBOT>_description/urdf directory and name it <ROBOT>.urdf. Additionally, create a joint configuration file with the namespace <ROBOT>_gazebo in the rl_sar/src/robots/<ROBOT>_description/config directory.
2. Place the trained RL model files in the rl_sar/src/rl_sar/models/<ROBOT> directory.
3. In the rl_sar/src/rl_sar/models/<ROBOT> directory, create a config.yaml file, and modify its parameters based on the rl_sar/src/rl_sar/models/a1_isaacgym/config.yaml file.
4. Modify the forward() function in the code as needed to adapt to different models.
5. If you need to run simulations, modify the launch files as needed by referring to those in the rl_sar/src/rl_sar/launch directory.
6. If you need to run on the physical robot, modify the file rl_sar/src/rl_sar/src/rl_real_a1.cpp as needed.

Reference

unitree_ros

Citation

Please cite the following if you use this code or parts of it:

@software{fan-ziqi2024rl_sar,
  author = {fan-ziqi},
  title = {{rl_sar: Simulation Verification and Physical Deployment of Robot Reinforcement Learning Algorithm.}},
  url = {https://github.com/fan-ziqi/rl_sar},
  year = {2024}
}