US 11,850,399 B2
Feedback predictive control approach for processes with time delay in the manipulated variable
Derrick K. Rollins, Ames, IA (US); and Yong Mei, Ames, IA (US)
Assigned to Iowa State University Research Foundation, Inc., Ames, IA (US)
Filed by Iowa State University Research Foundation, Inc., Ames, IA (US)
Filed on Sep. 9, 2021, as Appl. No. 17/447,242.
Application 17/447,242 is a continuation of application No. 16/508,934, filed on Jul. 11, 2019, granted, now 11,147,922.
Claims priority of provisional application 62/701,153, filed on Jul. 20, 2018.
Claims priority of provisional application 62/697,798, filed on Jul. 13, 2018.
Prior Publication US 2022/0016345 A1, Jan. 20, 2022
This patent is subject to a terminal disclaimer.
Int. Cl. A61M 5/172 (2006.01); G16H 20/17 (2018.01); A61M 5/142 (2006.01)
CPC A61M 5/1723 (2013.01) [A61M 5/14248 (2013.01); G16H 20/17 (2018.01); A61M 2205/3327 (2013.01); A61M 2205/3334 (2013.01); A61M 2205/3368 (2013.01); A61M 2205/50 (2013.01); A61M 2205/505 (2013.01); A61M 2205/52 (2013.01); A61M 2230/201 (2013.01); A61M 2230/50 (2013.01); A61M 2230/62 (2013.01); A61M 2230/63 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method of using an insulin delivery system comprising steps of:
determining an amount of insulin to administer using a feedback predictive control, wherein the feedback predictive control is stored on a machine readable non-transitory media;
wherein the feedback predictive control comprises a first model and a second model;
wherein the first model is a predictive model; and wherein the second model is a noise model, which models unmeasured disturbances and bias;
providing a predictive input and a measured input to the first model;
using the predictive input and measured input to provide a parametrized first model to provide an output, wherein the output is provided to the second model and to a PID controller, wherein the PID controller controls an insulin flow rate;
providing feedback to the PID controller from the second model;
administering insulin using the insulin delivery system based on the output and the feedback, wherein the administration of insulin is adjusted to nullify the effect of any disturbances on blood glucose concentration based on a predicted blood glucose value.