MPC for weakly nonlinear system
Nonlinear Model predictive control (NMPC) is a powerful control method for the control of constrained, nonlinear system. NMPC is based on the solution of an optimal control problem at each time instance. Often the solution of the optimal control problem requires to discretize the underlying continuous-time, nonlinear systems. This discretization can be computationally expensive. Therefore methods to reduce this effort are of great interest in order to enable or improve the applicability of these methods. One method to reduce this effort is to use integration schemes tailored to specific problem classes.
In this work we want to consider weakly nonlinear systems, i.e. systems in which the system response is dominated by the linear dynamics. The idea of this work is to use integration schemes tailored to weakly nonlinear systems in order to improve the speed. The main aim is to investigate the benefit of using such tailored integration approaches for the discretization of different nonlinear systems and to compare it with standard integration schemes. An extension to ODE sensitivity computation is possible.
If you are interested in this topic, please contact me.Helpful/Required Prerequisites:
Project Start:
Estimated time requirements:
| Literature search | 20% | |
| Systems theory | ca 30% | |
| Implementation/Test: | ca 50% |
Contact:
