direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Logo der TU Berlin

Inhalt des Dokuments

Noninvasive Blood Pressure Tracking

From Fachgebiet Regelungssysteme TU Berlin

Jump to: navigation, search


Abstract

Fig. 1: Non-invasive blood pressure measurement system. By controlling the blood flow to a constant small value, the balloon pressure follows the arterial blood pressure.

A noninvasive continuous blood pressure measurement technique that has been developed lately requires precise control of the blood flow through a superficial artery. The flow is measured using ultrasound and influenced via manipulating the pressure inside an inflatable air balloon which is placed over the artery.

This project is concerned with the design and evaluation of advanced controllers that are precise enough to enable the described indirect measurement principle. We combine classic feedback control methods with iterative learning control techniques that exploit the repetitive nature of the control task. Experimental validation is carried out in an artificial model of the cardiovascular system.

People involved

Cooperation

Funding

  • Technische Universität Berlin
  • Federal Ministry of Education and Research

Description

Blood pressure measurement is of vital importance in the diagnosis and treatment of many diseases. In particular, it is the cornerstone for the diagnosis, the treatment and the research on arterial hypertension. For more than one hundred years, the sphygmomanometry developed by Riva-Rocci and Korotkoff has been the most common method. Its limitations, however, are becoming increasingly evident and therefore alternative solutions are under investigation.

Our research partners from Charité Universitätsmedizin Berlin proposed a noninvasive method for continuous blood pressure measurement. However, this method requires tight control of the arterial blood flow that is measued by Doppler ultrasound. Due to measurement noise, traditional feedback control can only achieve limited bandwidth, and thus limited measurement accuracy. In order to improve the controller performance, we^ exploit the fact that the blood pressure curve is periodic and exhibits only small changes from one pulse to the next. More precisely, by looking at the pressure and flow curves of subsequent pulses batch-wise, we can adapt the balloon pressure curve of the next pulse based on the flow curve of the previous trials and thus introduce a feedback from pulse to pulse. This approach is known as Iterative Learning Control (ILC). The following figure shows one feasible controller structure with a cascade of classic feedback control and iterative learning control.

Fig. 2: Block diagram of a suitable controller structure. The plant outputs the pressure, which is controlled by the inner feedback loop, and the flow, which is controlled by the outer feedback loop. The outermost loop is an iterative learning controller (ILC), which learns in trials triggered by an edge detection of the flow signal.

While the blood pressure measurement should be as continuous as possible, a permanent reduction of the blood flow through the artery is not desired from a medical point of view. Therefore, we control the blood flow only in every fourth and fifth pulse, i.e. we alternate between letting three pulses pass and activating the controller for approximately two pulses, as depicted in the figure below. The double pulses in which the controllers are active are defined as the trials (or passes) of the ILC. In each of these trials, a feedforward control input will be applied which is updated between the trials based on measurement information from previous trials.

Fig. 3: Synchronization of the ILC trials with the pulses of the artificial cardiovascular system via edge detection. After the third rising edge of the flow signal, the two feedback control loops are closed. The ILC is activated after another time period $t_\text{wait}$ has passed, where $t_\text{wait}$ is chosen such that the feedback controllers are given enough time to converge, and such that each ILC trial starts at the end of a falling edge. All controllers are deactivated at the end of each trial, and the edge detection starts to count from zero.

The performance of the developed controllers is evaluated in a number of experiments using the laboratory model of the cardiovascular system, which includes an artificial artery and is capable of producing a physiologically pulsating blood pressure. Results show that, unlike classic feedback control, ILC is able to control the blood flow to a constantly low value.


Related Publications

T. Seel, T. Schauer, J. Raisch. Monotonic Convergence of Iterative Learning Control with Variable Pass Length. International Journal of Control, 90 (3):409–422, 2016.
T. Seel, S. Schneider, K. Affeld, T. Schauer. Design of a Learning Cascade Controller for a Continuous Noninvasive Blood Pressure Measurement System (in German). at - Automatisierungstechnik, 63 (1):5–13, 2015.
T. Seel, S. Weber, K. Affeld, T. Schauer. Iterativ lernende kaskadierte Regelung eines nichtinvasiven Blutdruck-Messsystems nach Penaz. In Workshop AUTOMED, Dresden, Germany, 2013.
T. Seel, S. Weber, K. Affeld, T. Schauer. Iterative Learning Cascade Control of Continuous Noninvasive Blood Pressure Measurement. In IEEE International Conference on Systems, Man, and Cybernetics, pages 2207-2212, Manchester, UK, 2013.

Recommend this page