longjohncoder / acados

Fast and embedded solvers for nonlinear optimal control

acados

Fast and embedded solvers for nonlinear optimal control.

General

• acados offers interfaces to the programming languages C, Python, MATLAB and Octave
• Documentation can be found on docs.acados.org
• Forum: If you have any acados-related question, feel free to post on our forum discourse.acados.org.
• Citing acados: references can be found here and here.

Installation

1. Initialize all submodules

   git submodule update --recursive --init
   

2. Build and install acados. Both a CMake and a Makefile based build system is supported at the moment. Please choose one and proceed with the corresponding paragraph.

   CMake

   Set the BLASFEO_TARGET in <acados_root>/CMakeLists.txt. For a list of supported targets, we refer to https://github.com/giaf/blasfeo/blob/master/README.md . Install acados as follows

   mkdir -p build
   cd build
   cmake .. # with optional arguments e.g. -DACADOS_WITH_OSQP=OFF/ON -DACADOS_INSTALL_DIR=<path_to_acados_installation_folder>
   make install
   

   Make

   Set the BLASFEO_TARGET in <acados_root>/Makefile.rule. For a list of supported targets, we refer to https://github.com/giaf/blasfeo/blob/master/README.md . Install acados as follows

   make shared_library
   export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:<path_to_acados_folder>/lib
   make examples_c
   make run_examples_c
   

acados interfaces

acados written in C and offers interfaces to the programming languages C, Python, MATLAB and Octave.

C interface

• This includes all functionality in <acados_root>/interfaces/acados_c.
• Documentation can be found here: C API

MATLAB and Octave interface

• The interface is written in MEX
• It deals with the problem formulation described in this PDF
• To get started
  • install acados as described above
  • Octave users: provide a CasADi version in <acados_root>/external/casadi_octave/. We recommend version 3.4.5. For Matlab, a CasADi version should be provided in <acados_root>/external/casadi_matlab/, however, this is done automatically when using the examples in the getting_started folder. On Linux machines with Octave 4.4.1. or later this can be done as follows:

      cd external
      wget -q -nc --show-progress https://github.com/casadi/casadi/releases/download/3.4.5/casadi-linux-octave-4.4.1-v3.4.5.tar.gz
      mkdir -p casadi-octave
      tar -xf casadi-linux-octave-4.4.1-v3.4.5.tar.gz -C casadi-octave
    

  • Linux and MacOS users:
    • in a terminal navigate to <acados_root>/examples/acados_matlab_octave/getting_started
    • run source env.sh and start Matlab/Octave
  • Windows users:
    • start Matlab
    • run acados_examples_env.m in <acados_root>/examples/acados_matlab_octave
  • enjoy the examples in <acados_root>/examples/acados_matlab_octave/getting_started
• More documentation can be found on docs.acados.org/interfaces/

Python interface

• The interface is based on templated C code, header files and a Makefile, which are rendered with the templating engine Tera.
• The generated C code can be easily deployed on embedded platforms.
• The same problem formulation as for the MATLAB interface is used (see here).
• The ctypes package is used to interact with the rendered Code.
• CasADi is required and is downloaded automatically, when installing the Python interface using pip
• More documentation can be found on docs.acados.org/interfaces/