Fast Predictive Control: Real-time Computation and Certification

Abstract

Pushing predictive controllers, which require the solution of an optimization problem at each sampling interval, into the millisecond range opens up both new possibilities, as well as new challenges for control. Computational limits placed on standard convex solvers invalidate basic assumptions made when proving the stability, or invariance of constrained control laws and as a result cannot be used in fast, real-time implementations with confidence.This extended abstract will present two optimization methods that can be certified to stabilize a system and enforce constraints, even for extremely short computational times or in the presence of time jitter. We'll report on two fast ‘code-generation' toolboxes: FiOrdOs, for first-order, and FORCES, for second order methods and demonstrate their capabilities through application examples.