Simulation studies on neural predictive control of glucose using the subcutaneous route.

A novel strategy for closed-loop control of glucose using subcutaneous (s.c.) tissue glucose measurement and s.c. infusion of monomeric insulin analogues was developed and evaluated in a simulation study. The proposed control strategy is an amalgamation of a neural network and nonlinear model predictive control (NPC) technique. A radial basis function neural network was used for off-line system identification of Nonlinear Auto Regressive model with eXogenous inputs (NARX) model of the glucoregulatory system. The explicit NARX model obtained from the off-line identification procedure was then used to predict the effects of future control actions. Numerical studies were carried out using a comprehensive model of glucose regulation. The system identification procedure enabled construction of a parsimonious network from the stimulated data, and consequently, design of a controller using multiple-step-ahead predictions of the previously identified model. According to the simulation results, stable control is achievable in the presence of large noise levels and for unknown or variable physiological or technical time delays. In conclusion, the simulation results suggest that closed-loop control of glucose will be achievable using s.c. glucose measurement and s.c. insulin administration. However, the control limitations due to the s.c. insulin administration makes additional action of the patient at meal time necessary.