The Blood Pressure Control System with Abnormal Control Phase Recognition

Abstract

A study is presented of a blood pressure control system with safety feature in therapeutic practice. Using fuzzy logic, the physician's heuristic rules for control were transformed into automatic control rules. Fuzzy controller has two input variables (the difference between the actual and target pressures, the increment of blood pressure) and one output variable (the variation in the drug dosage) expressed as a fuzzy variable by establishing corresponding membership functions. 25 control rules were converted into fuzzy control rules. As safety feature, the control phases recognition function was developed by using TMSTR (Time-sequential Multiple State Transition Representation) technique which is devised to represent time-sequential data transition with IF-THEN rule structure. The function is observing time-trend of blood pressure and drug infusion rate during control. The function takes over control from the fuzzy controller when abnormal control phases are recognized. To evaluate the present system, the pressure elevation simulation and animal experiments were performed. The blood pressure was elevated from the ordinary level (about 100 mmHg) to 150 mmHg by infusing the vasopressor. Excellent blood pressure control was achieved in both of simulation and experiments, and the overall effectiveness of the present control system was verified.