Active implantable and wearable medical devices work as closed-loop control systems: sense a physiological variable, compare it to a target, and drive an actuator to close the gap, continuously.
error(t) = setpoint - measured(t)
actuator_output = Kp*error + noise_rejection
- Sensor/actuator nodes — glucose sensors, cardiac lead sensors, and their paired actuators (pump, pacing pulse) in the device.
- Control loop connectivity — how tightly sensor readings are fed back into actuator decisions each cycle.
- Controller gain — how aggressively the controller corrects a detected deviation from the physiological setpoint.
- Sensor noise — measurement noise the controller must filter out to avoid over-correcting.
Closed-loop insulin pumps ('artificial pancreas' systems) are exactly this control network in a wearable device, correcting blood glucose in near real time without requiring the patient to manually calculate every dose.