The extended Kalman filter for continuous glucose monitoring

Abstract

Glucose monitoring is a problem with evolving solutions using sensors and data processing techniques that are continually improving. This paper presents the development and application of the extended Kalman filter to the problem of real-time estimation of blood glucose levels and is consistent with sensors based on optical, electrical, or chemical processes. The structure of the extended Kalman filter provides the capability to systematically accommodate new information as it develops. This flexibility and the potential for performance improvement are demonstrated by processing patient data files generated using the Medtronic (Northridge, CA) MiniMed continuous glucose monitoring system (CGMS).