Complexity reduction of receding horizon control

Abstract

The off-line solution of constrained optimal control problems has garnered much attention because implementation can be realized with a simple look-up table. In this paper, a detailed complexity analysis of algorithms used to compute the explicit optimal control solution is given. Based on this analysis, two novel algorithms to compute controllers of significantly lower complexity are presented. Stability and feasibility of the closed-loop system are guaranteed by both algorithms. Extensive simulation results suggest that, on average, the controller complexity can be reduced by orders of magnitude at the cost of a performance decrease below 0.5%.