Sensor fusion and distributed platform development for artificial pancreas

Abstract

This paper describes a distributed sensor platform for a new breed of artificial pancreas devices. In recent work, a multi-variable adaptive algorithm has been proposed which incorporates physical activity of the patients for accurate prediction and control of glucose levels. In order to facilitate this algorithm, we integrate a smartphone and multiple sensors including activity trackers and a glucose monitor into a distributed system. The proposed sensor platform provides realtime data access for the artificial pancreas control algorithm hosted on a remote device.