Author:

Abstract:

This work compares various numerical methods to robustify periodic optimal control problems using the paradigm of Lyapunov differential equations. In this paradigm, estimates for state-covariance are obtained by solving the periodic Lyapunov equations for the a system linearized along a to-be-optimized trajectory, and are added to objective or constraints of the or iginal optimal control problem. For non-trivial dynamical systems, method details were found to be critical to obtain algorithms with reasonable time complexity. An application for time-optimal quadcopter flight is worked out numerically with the optimal-control tool CasADi, which was extended by the author to solve discrete periodic Lyapunov equations using the SLICOT library.