Developing a dynamic bio-inspired intelligent system to support diabetic patients

This paper proposes a new biologically-inspired phenomenological model of insulin release that can be used to continuously monitor a person with diabetes and therefore recommend insulin dose and predict events. Given all the parameters and conditions that have to be met in order to achieve stability, the proposed insulin-release system will be considered as highly non-linear and dynamical. The resulting model will study the mechanisms that keep the system from reaching a chaotic state (diabetes), based on the property of self-organised criticality. This property was selected since it has been previously identified and studied in other dynamical biological systems. The proposed model will then be used as part of a model-based reasoning system that aims to support patients with diabetes