Controlled Invariant Sets for Gaussian Process State Space Models
For control engineers, this provides a data-driven approach to guarantee safety with probabilistic guarantees for nonlinear systems with unknown dynamics.
We compute probabilistic controlled invariant sets for nonlinear systems using Gaussian process state space models, and propose a semidefinite programming scheme for designing state-feedback controllers that maximize the probability of trajectories staying within the set while satisfying input constraints. The method is validated on a quadrotor in simulation and on a physical platform.
We compute probabilistic controlled invariant sets for nonlinear systems using Gaussian process state space models, which are data-driven models that account for unmodeled and unknown nonlinear dynamics. We propose a semidefinite programming scheme for designing state-feedback controllers that maximize the probability of the trajectories staying within a probabilistic controlled invariant set while satisfying input constraints. The results are validated on a quadrotor, both in simulation and on a physical platform.