Development and in silico validation of the PID-algorithm for the artificial pancreas with intraperitoneal insulin delivery

Abstract

Background: The efficacy of the treatment of type 1 diabetes can be markedly improved using artificial pancreas (AP), which is a technology to automatically control blood glucose levels. Aim: In this paper, we propose the construction of a controller for controlling the automated delivery of insulin in AP based on a proportional–integral–derivative (PID) algorithm using intraperitoneal (IP) insulin delivery. Methods: The project used rapid-acting insulin in the IP space when setting up a PID controller with feedback to ensure the safe and efficient delivery of insulin. The controller was configured to satisfy feedback insulin present in blood. Controller check was performed In Silico using the metabolic simulator UVA|Padova T1DMS on 10 virtual patients. Results: The proposed controller design has time to reach 83% within the glycaemic range of 70–140 mg/dl (3.9–7.8 mmol/l), without time spent in hypoglycaemia. Conclusions: In a future study we plan to test this controller in vivo to evaluate its performance in vivo.

Keywords

• аrtificial pancreas
• intraperitoneal insulin infusion
• in silico
• control algorithm
• insulin pump
• preclinical studies