PolyMPC: An Efficient and Extensible Tool for Real-Time Nonlinear Model Predictive Tracking and Path Following for Fast Mechatronic Systems. Developed at the Automatic Control laboratory EPFL as part of the AWESCO project.
Information on installation, algorithms and API can be found here: polympc.rtfd.io
PolyMPC an open-source C++ library for real-time embedded nonlinear optimal control that combines high computational performance, modularity and a simple and intuitive user interface. PolyMPC is a lightweight collection of optimised and loosely coupled tools for nonlinear optimisation and optimal control which includes several quadratic programming (QP) and nonlinear programming (NLP) solvers, quasi-Newton Hessian approximation and Hessian regularisation methods, a polynomial interpolation and approximation module and an implementation of the Chebyshev pseudospectral collocation method, a continuous algebraic Riccati equation (CARE) solver, linear quadratic regulator (LQR), and more.
Petr Listov - [email protected]