Simple adaptive PI controller development and evaluation for mean arterial pressure regulation

Abstract

The automatic control of mean arterial pressure (MAP) in post-cardiac surgery has many advantages for patients and medical staff. However, it is a challenging problem for the control engineering and biomedical engineering community. Successfully developed controllers in the research environment have not been successful in the market because of their complex implementation, complicated operation, and inadequate capabilities performance across the range of patient variabilities. This paper describes the development and testing of a simple embedded adaptive PI controller for automatic MAP regulation of a known range of patients. The developed controller is a synergistic combination of simple and proved schemes that have complementary characteristics; it is a parameter estimator for identification of patient MAP parameters, calculation of patient sensitivity, and adjustment of gains of a specific proportional-integral controller for the identified patient. The controller was implemented in Matlab/Simulink for desktop simulation along with a MAP model that allows the variation of the type of patient tested. In addition to the desktop simulation, an embedded version of the controller implemented in ARM Cortex M3 low-cost microcontroller was tested on a hardware-in-the-loop (Hil) simulation along with a real-time MAP model embedded in ARM Cortex A7 single board computer. Both simulations (desktop and hardware-in-the-loop) presented similar results, meeting the established MAP regulation performance criteria: settling time less than 600 s, overshoot less than 10 mmHg, and steady-state error about setpoint less than 5 mmHg. The results achieved are as follows: settling time between 170 and 330 s, overshoot between 1 and 4 mmHg, and steady-state errors about setpoint less than 3 mmHg, across the range of MAP patients considered. The controller developed in this work is a viable alternative to the complicated identification and control techniques normally employed for automatic MAP regulation. An adaptive, simple PI controller for adjusting the MAP and having low computational cost was developed and embedded in a low-cost microcontroller. The controller was evaluated via desktop and Hil simulation across the known range of MAP parameters of patients, with successful regulation of MAP in all cases considered.