Physiologic-state-adaptive recovery of aortic blood pressure and flow using blind 2-channel IIR cardiovascular system identification

Abstract

This paper presents the development and analysis of a method to identify a two-channel cardiovascular system using two distinct peripheral blood pressure signals. The method is able to characterize the upper- and lower-limb arterial path dynamics as well as the aortic root impedance, and recover the aortic blood pressure and flow signals fed to it. The blind system identification and input de-convolution algorithms for a class of two-channel infinite impulse response systems are developed and applied to a gray-box model of a two-channel cardiovascular system. Persistent excitation condition, model identifiability and asymptotic variance are analyzed to quantify the method’s validity and reliability. Experimental results based on 83 data segments obtained from a swine subject show that the cardiovascular dynamics can be identified very accurately and reliably, and the aortic blood pressure and flow signals are stably recovered from two distinct peripheral blood pressure signals under diverse physiologic conditions. The benefit of the proposed method is demonstrated by comparing it to a predetermined transfer function describing the cardiovascular dynamics at nominal physiologic conditions.