Closed-loop control based on predicted trajectories

Abstract

Disclosed herein are techniques for closed-loop control based on predicted trajectories. The techniques may involve predicting a trajectory of a glucose level of a patient based on past observations of the glucose level. The techniques may further involve determining a cost expression based on the predicted trajectory. The cost expression may be based on a duration of time that the trajectory is predicted to extend outside of a target glucose concentration range. Additionally, the techniques may involve affecting, based on the cost expression, at least one future command to at least one infusion pump so as to shorten the duration of time that the trajectory is predicted to extend outside of the target glucose concentration range.

Dependencies are inferred from the claim text.