README_install.md

Dependencies

Gradient-based MPC: It uses CasADI to define the model and acados to solve the optimal control problem. Sampling-based MPC: jax for both. The simulation environment is based on Mujoco.

You can install this repo with Pixi (preferred) or Conda.

---

Pixi installation

1. install pixi

2. create the environment:

pixi install
pixi shell

3. go to Common Steps section

Conda installation

1. install miniforge (x86_64 or arm64 depending on your platform)

2. create an environment using the file in the folder installation choosing between nvidia and integrated gpu:

conda env create -f mamba_environment.yml
conda activate quadruped_pympc_env

3. go to Common Steps section

Common steps between Pixi and Conda

4. clone the other submodules:

   git submodule update --init --recursive

5. go inside the folder acados and compile it pressing:

   cd quadruped_pympc/acados/
   mkdir build
   cd build
   cmake -DACADOS_WITH_SYSTEM_BLASFEO:BOOL=ON -DCMAKE_POLICY_VERSION_MINIMUM=3.5 ..
   make install -j4
   pip install -e ./../interfaces/acados_template
   

6. inside the file .bashrc, given your path_to_acados, put:

   export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:"/path_to_acados/lib"
   export ACADOS_SOURCE_DIR="/path_to_acados"
   

   Notice that if you are using Mac, you should modify the file .zshrc adding

   export DYLD_LIBRARY_PATH=$LD_LIBRARY_PATH:"/path_to_acados/lib"
   export ACADOS_SOURCE_DIR="/path_to_acados"
   

The first time you run the simulation with acados, in the terminal you will be asked to install tera_render. You should accept to proceed.

7. go to Quadruped-PyMPC initial folder and install it:

   pip install -e .
   

How to run - Simulation

1. activate your pixi or conda environment

2. go in the main Quadruped-PyMPC folder and press

   python3 simulation/simulation.py
   

In the file config.py, you can set up the robot, the mpc type (gradient, sampling..), its proprierties (real time iteration, sampling type, foothold optimization..), and other simulation params (reference, gait type..).

3. you can interact with the simulation with your mouse to add disturbances, or with the keyboard by pressing

arrow up, arrow down -> add positive or negative forward velocity
arrow left, arrow right -> add positive or negative yaw velocity
ctrl -> set zero all velocities

How to run - ROS2

1. activate your pixi or conda environment

2. you can run now the script

   python3 ros2/run_controller.py
   

3. if you want to test the above node with a simulator, for example to test ros2 delay, you can run

   python3 ros2/run_simulator.py
   

4. joystick

   ros2 launch teleop_twist_joy teleop-launch.py joy_config:='xbox'
   

5. general commands from terminal available here

6. for a real-robot deployment, use a nice state estimator