Non-invasive glucose monitoring from measured inputs

In recent years, there has been tremendous progress in the technological advancement of continuous-time glucose devices. However, even the most successful devices are at least semi-invasive, requiring subcutaneous insertion of a sensor, that affects patient comfort in addition to reoccurring costs for sensor placement. The objective of this work is the development of a continuous-time glucose modeling system that infers glucose levels from non-invasive measured inputs of activity, stress, and food components. The method appears to have excellent long term stability and the correction algorithm for using measured glucose is quite promising.