Adaptive PID Control for Chronotropic Efficiency in Cardiac Pacemakers

Jasmeet Bhatia, Rohil Deolalikar, Cole Wener, Andrew Zeller, Cameron Zeller, Gert Cauwenberghs

Abstract: Chronotropic Incompetence (CI) has recently been discovered as a sign of near-future heart failure. Having the ability to correct CI could avoid or delay chronic heart failure. Pacemakers account for the heart’s failed electrical conduction for specific cardiac arrhythmias such as Tachy-Brady Syndrome, Atrial Fibrillation, and many other conditions, most of which arise from a failing sinoatrial (SA) node. Because of the increasing relevance of cardiovascular disease and, as a result, pacemakers, modeling patients with similar conditions as a control system is of great importance. We introduce an adaptive pacemaker controller which embeds a proportional-integral-derivative (PID) controller inside an additional least-mean-squares (LMS) adaptive control loop that dynamically adjusts the PID control parameters for increased chronotropic efficiency, more rapidly adjusting the heart rate to help patients recover with CI after exertion. We derive the optimal learning rates for the LMS adaptive loop to arrive at optimal PID parameters for the control loop based on the signal statistics of the loop error and target control bandwidth to return the current heart rate to the desired heart rate.