Page 293 - Glucose Monitoring Devices
P. 293

300    CHAPTER 15 Automated closed-loop insulin delivery




                         in a predefined zone, such as a desirable euglycemic zone. An inherent benefit of
                         control to the zone is that the pump actuation and infusion activity will be inherently
                         limited when the glucose levels are within the defined zone. The lack of corrective
                         suggestions when glycemic measurements are within the desired zone has the poten-
                         tial to reduce power consumption on handheld and mobile devices typically running
                         the control algorithms [14,29].
                            The novelty of the zone MPC algorithm lies in the cost function formulation.
                         Zone MPC, analogous to typical MPC algorithms, predicts the dynamic future tra-
                         jectory profile of the glycemic values using an explicit model of the glucose-insulin
                         dynamics and a candidate input trajectory to be optimized. In contrast to typical
                         MPC that drives the controlled variable to a specific fixed set-point, the optimization
                         problem in zone MPC attempts to maintain or drive the predicted outputs into a pre-
                         defined zone described by upper and lower bounds [29,58]. The optimization prob-
                         lem of zone MPC for a given current state x is

                                                V N ðxÞ¼  min Jðx; uÞ
                                                        x 0 ;u 0 ;.;x N
                            subject to:

                                                        x 0 ¼ x
                                               x kþ1 ¼ fðx k ; u k Þ; k˛I 0:N 1
                                                              k˛I 0:N
                                                 y k ¼ gðx k ; u k Þ
                                                 hðx k ; u k Þ  0; k˛I 0:N
                                                  x k ˛X; u k ˛U; k˛I 0:N
                         where the objective function JðxÞ is defined by
                                              N
                                             X   range T  range     s T        s
                                                y    Q k y
                                                 k      k
                                     Jðx; uÞ¼               þðu k   u Þ R k ðu k   u Þ
                                             k¼0
                         with y range  as the superposition of all the predicted glucose outputs exceeding the
                         permitted euglycemic range, and is given by
                                                      h              i
                                                range  lo   lb  hi  ub
                                               y    ¼ y   y ; y   y
                                                k      k       k
                            Further, y lo  and y hi  denote the aggregation of all predicted controlled outputs
                                    k     k
                                                                               lb
                                                                      ub
                                             lb
                         below the lower bound y and above the upper bound y , where y and y ub  char-
                         acterize the desired glycemic range. This aggregation is conducted by setting all pre-
                                                                        lb ub
                         dicted output values that are within the safe target range ½y y Š to zero as follows:
                                                                  lb
                                                  lo   y k  if y k < y
                                                  k
                                                 y ¼
                                                        0  otherwise
                         and
                                                                  ub
                                                 hi    y k  if y k > y
                                                 k
                                                 y ¼
                                                       0   otherwise
   288   289   290   291   292   293   294   295   296   297   298