Feedback control algorithms for automated glucose management in T1DM: The state of the art

Safety and performance of artificial pancreas (AP) systems rely on the design of advanced control algorithms. In this chapter, we provide an overview of various feedback control algorithms for glucose regulation proposed in the literature. We discuss how the safety and performance requirements are enforced in real-time closed-loop control algorithms through designing particular controller structures and operation constraints. Existing results on proportional-integral-derivative control, fuzzy logic control, model predictive and optimal control, and linear parameter-varying control are covered. With the need for long-term home use of the AP, controller adaptivity becomes an important factor to consider in AP algorithm design. In this regard, developments on long-term parameter adaptation and machine-learning-based control are also discussed.