Page 308 - Glucose Monitoring Devices

Closed-loop glycemic control results 315

correspond with the physical definitions. In addition, maximum rates of change
constraints are defined for the parameters estimated simultaneously to the states
to avoid sudden changes in the parameter values due to measurement noise or
unknown disturbances that may result in inappropriate corrections.

Modulating insulin infusion
Estimates of the PIC can be used to moderate the aggressiveness of the MPC algo-
rithm and dynamically constrain the insulin infusion. The MPC thus explicitly
considers the insulin concentration in the bloodstream within the control law compu-
tation. A plasma insulin risk index can manipulate the weighting matrix for penal-
izing the amount of input actuation (aggressiveness of insulin dosing) depending
on the estimated PIC, thus suppressing the infusion rate if sufficient insulin is present
in the bloodstream (Fig. 15.3). As the plasma insulin concentration increases, the
penalty weight on the input action is also simultaneously increased as
R k ¼ RðPIRI k Þ, with PIRI k as the risk index at sampling instance k derived from
the estimated PIC. Incorporating PIC constraints in the optimal control problems
can prevent insulin stacking that may lead to hypoglycemia, which can yield a
safe and reliable AP system even in the presence of significant uncertainty in the
system.

Closed-loop glycemic control results

In this section, we illustrate the efficacy of an MPC formulation that employs adap-
tive models recursively identified through subspace-based techniques. The adap-
tive MPC incorporates variable weights in the objective function through the
glycemic and plasma insulin risk indexes. The adaptive MPC is also dynamically

FIGURE 15.3
The plasma insulin risk index for increasing the penalty weight on insulin infusion based
on plasma insulin concentration estimates.