The Artificial Pancreas Project developed by SUPELEC and the University Hospital of Rennes is focused on the evaluation of an original control algorithm that computes appropriate subcutaneous insulin infusion in response to continuous glucose measurements The results during a 24-hour closed-loop period will be compared with those of the open-loop. The implementation of this algorithm named Error Dynamics Shaping (EDS) requires as a preliminary a phase of identification of the parameters of the model (insulin pharmacokinetics and effect on glucose metabolism) followed by a phase of optimization i.e. a personalized adjustment of the damping parameters of EDS.

The present study aims to formalize this procedure of identification/optimization of the parameters of the model and controller EDS. They will be determined in type -1 diabetic patients using insulin therapy and under conditions free from disturbances such as those generated by the glucidic meals, i.e. during a period of glucidic fast. Three variables will determine this procedure: blood glucose, insulin concentration and insulin doses.

Inclusion Criteria:

• Type 1 diabetic patients (age : 18-70)
• Diabetes duration higher than three years
• Insulin pump with Humalog®
• Hospitalized for a 24-hour glucidic fast test
• Written informed consent

Exclusion Criteria:

• Pregnancy or nursing
• Acute infectious disease
• Corticoid treatment
• Creatinine clearance <40 ml/mn
• Patients using Apidra® or Novorapid®
• People major being the object of a legal protection (safeguard of justice, supervision, trusteeship)