WO2017129495A1

WO2017129495A1 - Pulse rate measurement module and method

Info

Publication number: WO2017129495A1
Application number: PCT/EP2017/051252
Authority: WO
Inventors: Paul Aelen; Teun Van Den Heuvel; Erik BRESCH; Maarten Petrus Joseph KUENEN
Original assignee: Koninklijke Philips N.V.
Priority date: 2016-01-28
Filing date: 2017-01-20
Publication date: 2017-08-03

Abstract

The present disclosure relates to a pulse rate measurement module (40) for use in a vital signs monitoring system (10) and to a corresponding method, the module (40) comprising an interface (42) operatively coupled with a pressure measurement unit (26) that detects a cuff pressure signal (30) of a wearable cuff (14) which is arranged to pressurize a measurement site of a subject, a signal separator (46) arranged to extract a potentially pulse rate indicative signal portion from the cuff pressure signal (30), an autocorrelator (48) arranged to autocorrelate the separated potentially pulse rate indicative signal portion, and a pulse rate processor (50) arranged to process the pulse rate based on temporal information linked with processed correlation coefficients. The disclosure further relates to a method for controlling the inflation rate of a wearable cuff for an inflation-based non- invasive blood pressure measurement system and to a vital signs monitoring system (10) arranged for blood pressure measurement.

Description

Pulse rate measurement module and method

FIELD OF THE INVENTION

The present disclosure relates to the field of vital signs monitoring, particularly to the field of pulse rate measurement. An exemplary field of application for pulse rate measurement approaches can be found in the context of blood pressure monitoring. In this respect, embodiments of the present disclosure relate to refinements of non- invasive blood pressure measurement methods and devices. Generally, non-invasive blood pressure measurement is also referred to as NIBP. More particularly, in more specific embodiments, the present disclosure relates to refinements in blood pressure measurement systems that are suitable for measuring a blood pressure during the inflation of a cuff, and to operation control methods for such blood pressure measurement devices. Generally, non- invasive blood pressure measurement may be referred to as methods of and approaches to detect arterial blood pressure in a mediate fashion without the need of obtrusive measures applied to the body of a subject. BACKGROUND OF THE INVENTION

The pulse is the regular throbbing of arteries, caused by the successive contractions of the heart. The pulse rate is the rate of the pulse, usually stated in beats per minute (BPM). The pulse rate can vary according to physical needs of a subject, including the need to absorb oxygen and excrete carbon dioxide. The pulse rate is a fundamental vital sign and also referred to as one of the primary vital signs. Hence, the pulse rate as such is indicative of the well-being of the monitored subject. Further, the pulse rate may be utilized as an auxiliary measure for operating medical equipment, including monitoring or measurement devices that address further vital signs (e.g. blood pressure, respiratory rate and temperature), respective derivatives and/or more specific health indicators.

In the following, main aspects of the disclosure will be described and detailed in the context of blood pressure measurement and monitoring, while further fields of application can be seen in the medical and health technology field, in the field of leisure and sports equipment, and, more generally, in the field of vital signs monitoring. As opposed to invasive blood pressure measurement, non-invasive blood pressure measurement is a method of measuring the blood pressure of a human indirectly. The most established non- invasive blood pressure measurement methods require an inflatable cuff to be placed around a limb, where the pressure in this cuff is changed to infer blood pressure. Generally, respective devices may be referred to as sphygmomanometers. Within the sphygmomanometers, there are two categories of determining blood pressure.

A first method for non-invasive blood pressure measurement is the so-called auscultatory-based method. Auscultation measurements are based on listening to the sounds of the artery during the period that the cuff pressure is changed. These sounds are referred to as so-called Korotkoff sounds, as Korotkoff was the first person that linked these sounds to blood pressure. Sounds are typically measured with a stethoscope that is placed at the brachial artery. First, the cuff is inflated until the flow of the blood in the artery is blocked, which can be observed by the absence of Korotkoff sounds. Then the cuff is deflated. Users (usually medical staff) may determine the blood pressure by listening to the change of the Korotkoff sounds. The pressure at the occurrence of the first sound is usually referred to as the systolic blood pressure. During further deflation, sounds first become louder and decrease in amplitude later on. The pressure at which the last sound is present is usually referred to as the diastolic blood pressure.

A second method for non-invasive blood pressure measurement is the so- called oscillometric method. In this method, the systolic and diastolic blood pressure values are dependent on the amplitudes of the pressure oscillations in the blood pressure cuff. The amplitudes of these pressure oscillations are measured at different cuff pressures, where the systolic and diastolic blood pressure values are usually calculated as the cuff pressure where the pressure oscillations are a certain fraction of the maximum of the pressure oscillations. Conventionally, the cuff pressure is changed by first blocking the flow of blood by inflating the cuff. Then, the cuff is deflated slowly during which the oscillation amplitudes are measured. Currently, most of the electronic automatic blood pressure measuring devices measure the blood pressure based on the oscillometric method which has been widely used in clinical practice and in outpatient or ambulant environments.

US 2014/0309541 Al discloses an oscillometric blood pressure measurement device comprising a cuff that, when worn on a blood pressure measurement area, pressurizes an artery in the measurement area at a pressure of a fluid in the cuff; a piezoelectric pump that increases the pressure within the cuff; a deflating unit that reduces the pressure within the cuff; a pressure detection unit that detects a cuff pressure that is the pressure within the cuff; and a control unit, wherein the control unit is arranged to determine an amplitude and a frequency of a voltage applied to the piezoelectric pump; carry out control so that a voltage at the amplitude and frequency determined by the determination unit is applied to the piezoelectric pump; and calculate a blood pressure value based on the cuff pressure detected during inflation when the cuff pressure is increased by the piezoelectric pump. Consequently, US 2014/0309541 Al describes a blood pressure measurement device that is arranged to analyze the inflation stage so as to derive the blood pressure.

US 2013/0289421 Al discloses a method for evaluating whether an

oscillometric signal representative of pressure oscillations in the vasculature of a patient is associated with special conditions, said method comprising obtaining an oscillometric signal at a location on an extremity of the patient; determining a ratio using a value associated with a first frequency component of the oscillometric signal and a value associated with a second frequency component of the oscillometric signal; comparing the ratio to a threshold value; associating a first diagnostic class from a plurality of diagnostic classes with the oscillometric signal when a first outcome of the comparison to the threshold value results from said comparing the ratio to a threshold value; and associating a second diagnostic class from a plurality of diagnostic classes with the oscillometric signal when a second outcome of the comparison to the threshold value results from said comparing the ratio to a threshold value.

US 2004/0186386 Al relates to a patient monitoring method and system that determines blood pressure and pulse rate. The document discloses a method of determining a pulse rate of a patient, the method comprising acquiring and storing measured information for at least one pulse at each of a plurality of pressure steps; determining and storing quality values for the at least one pulse at each of the plurality of pressure steps; analyzing pulse matching criteria for the plurality of pressure steps; selecting measured information from a set of pulses that have quality values greater than a predetermined amount and meet a pulse matching criteria; and determining the pulse rate based on the selected measured information.

KURT BARBE ET AL: "Analyzing the Windkessel Model as a Potential Candidate for Correcting Oscillometric Blood-Pressure Measurements", IEEE

TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, IEEE SERVICE CENTER, PISCATAWAY, NJ, US, vol. 61, no. 2, 1 February 2012, pages 411-418, relates to improvements in noninvasive automatic oscillometric blood-pressure meters (NIBP), involving a correction or calibration of the oscillometric blood-pressure meters in case hypertension or hypotension patients. A correction method is proposed, involving a databased Fourier series approach to model the oscillometric waveform, and using the Windkessel model for the blood flow to correct the oscillometric blood-pressure meters.

Conventionally, oscillometric blood pressure measurement methods utilize the deflation stage of a measurement wherein the pressure is gradually reduced so as to determine blood pressure indicative signals. Oscillometric blood pressure measurements are basically arranged to derive characteristic blood pressure values such as systolic blood pressure and diastolic blood pressure, in a mediate fashion from measured values and empirical parameters which are typically device-dependent.

Deflation stage based measurement is well established but is at the same time to some extent uncomfortable since the subject (or patient) is exposed to a relatively high cuff pressure for a certain amount of time. Further, since the deflation stage is used, first a defined maximum pressure level needs to be achieved starting from which the deflation procedure may be initiated. Therefore, the patient is exposed to a higher maximum cuff pressure than basically required for the blood pressure measurement as such. Further, the inflation of the cuff takes some time and also the deflation regime takes a considerable amount of time which further increases the discomfort for the patient (or subject).

Inflation stage based blood pressure measurements may reduce the discomfort as the blood pressure measurement may be accomplished in less time as the inflating part of cuff pressure is used rather than the deflating part. Once the pressure range of interest is passed at the inflation stage, pressure relief may be initiated so as to reduce the overall measurement time and the pressure-based discomfort.

However, also in the domain of inflation-based non- invasive blood pressure measurement methods and devices, there is room for improvement. Generally, an appropriate inflation rate and/or deflation rate needs to be defined based on a trade-off between measurement time, user discomfort, reliability and accuracy. As an example, operating the cuff, particularly pressurizing the cuff at a certain (inflation/deflation) rate may be performed dependent on a current pulse rate of the to-be-monitored subject. To this end, additional sensors may be necessary which make respective monitoring systems more complicated and more expensive. Consequently, there is also room for improvement in pulse rate

measurement methods and devices.

Particularly inflation-based non- invasive blood pressure measurement poses several challenges. The benefit is that the discomfort of the monitored subject and the measurement time may be considerably reduced. However, in practice, when it comes to implementation, several issues may emerge. For instance, the desired inflation (or deflation) rate is subject to a trade-off between a sufficient number of oscillations that are covered in the inflation stage such that a reliable measurement of the blood pressure (for instance systolic blood pressure (SBP) and diastolic blood pressure (DBP) and mean arterial pressure (MAP)) is enabled. On the other hand, it is a goal to reduce the measurement duration and inflation time. Hence, appropriate control of the inflation stage (and in deflation based systems of the deflation rate) should be performed.

Since the frequency of the oscillations in the cuff pressure signals corresponds to the pulse (or pulse rate) of the to-be-monitored subject, pulse rate measurement is a beneficial measure to further improve non- invasive blood pressure measurement. Once the subject's pulse rate is measured, the inflation rate (or the deflation rate) may be associated thereto. Hence, the cuff may be pressurized or de-pressurized as a function of the measured pulse rate.

SUMMARY OF THE INVENTION

It is therefore an object of the present disclosure to provide a pulse rate measurement module which is arranged to be used in a vital signs monitoring system, particularly in a blood pressure measurement system. More preferably, at least in some embodiments, the pulse rate measurement module is operable based on a limited variety of input signals, preferably based on a cuff pressure signal which is anyway present in a blood pressure measurement arrangement. In a more general context, it is an object of the present disclosure to present a pulse rate measurement module which does not require a distinct, separate measurement signal which is, as such, not required for the main signal or vital sign a respective vital signs monitoring system typically addresses. Further, a corresponding pulse rate measurement method, a vital signs monitoring system that implements a respective pulse rate measurement module, and a corresponding vital signs monitoring method shall be provided.

More particularly, it is an object of the present disclosure to seek for improved approaches to non-invasive blood pressure measurement which may be implemented in respective non- invasive methods and devices. Particularly, improved methods and devices for inflation stage-based non- invasive blood pressure measurement shall be presented.

More particularly, it is an object of the present disclosure to provide an improved inflation strategy for an inflation-based blood pressure measurement procedure which addresses both accuracy and robustness issues. Further, the methods and devices in accordance with the present disclosure preferably enable the measurement of a subject's blood pressure within a short amount of time. Preferably, discomfort for the monitored subject may be even further reduced without adversely affecting the monitoring accuracy.

More generally, in accordance with some aspects of the present disclosure, at least some drawbacks that are inherent to prior art inflation based blood pressure

measurement approaches shall be addressed and mitigated. Further, at least in some aspects of the present disclosure, alternative approaches to signal processing in inflation based oscillatory blood pressure measurement devices and methods shall be presented.

In accordance with a first aspect of the present disclosure a pulse rate measurement module for use in a vital signs monitoring system is presented, the module comprising an interface operative ly coupled with a pressure measurement unit, wherein the interface detects a cuff pressure signal of a wearable cuff which is arranged to pressurize a measurement site of a subject, a signal separator arranged to extract a potentially pulse rate indicative signal portion from the cuff pressure signal, an auto correlator arranged to autocorrelate the separated potentially pulse rate indicative signal portion, and a pulse rate processor arranged to process the pulse rate based on temporal information linked with processed correlation coefficients.

Main aspects of the present disclosure address a fast and efficient way to at least approximately determine the pulse rate. While it is acknowledged that pulse rate measurement devices (for instance heart rate monitor watches) are generally known as such, which may be referred-to in the context of the present disclosure as single-purpose standalone pulse rate measurement devices, a field of application of pulse rate measurement as presented herein is the provision of supplementary or auxiliary signals which may be utilized to facilitate and improve the detection and measurement of further vital signs and related signals. Therefore, an exemplary application can be seen in the context of non- invasive blood pressure measurement. Pulse rate measurement as such is particularly suitable for inflation-based non- invasive blood pressure measurement. However, it is not necessarily required to derive and provide highly accurate measurements. By contrast, at least a certain level of accuracy may be sufficient to significantly improve the operation of the above vital signs monitoring system by, so-to-say, linking the control thereof with measured pulse rate values or, at least, estimates.

As indicated above, the pulse rate measurement module within the context of the present disclosure utilizes the cuff pressure signal which is anyway present in respective vital signs monitoring systems. However, the potential field of application is not limited to respective blood pressure measurement arrangements. Further fields of application can be seen where an auxiliary pulse rate signal may improve the operation.

In accordance with the present disclosure, vital signs monitoring systems may be provided which do not need to implement, or to be coupled with, additional external pulse rate sensors.

By way of example, the inflation speed or inflation rate of the blood pressure measurement device may be set and controlled as a function of the measured pulse rate. As a consequence, it may be ensured that the blood pressure measurement covers a sufficient number of oscillations while it is, at the same time, ensured that the measurement, particularly the inflation stage, requires only a short amount of time. This may significantly reduce the subject's discomfort while maintaining or even enhancing the measurement accuracy.

By using the cuff pressure signal as such as an input signal for the pulse rate measurement, a fast measurement can be established.

The pulse rate may be computed based on temporal information that is obtainable from processed correlation coefficients that form a result of the autocorrelation. The temporal information may be indicative of a time lag between characteristic signal portions (extreme values, peaks, curve shapes, etc.). Hence, based on the autocorrelation applied, time lag information or reciprocal frequency information may be obtained.

Eventually, the pulse rate may be obtained therefrom.

In accordance with the above embodiment, the cuff pressure signal as such is processed so as to extract temporal information and/or frequency information therefrom. An autocorrelation may be applied to the signal of interest, preferably subsequent to preprocessing measures. Autocorrelation may be also referred to as serial correlation.

Autocorrelation may be regarded as a correlation of a signal with a delayed (lagged) version of itself. In this way, periodically occurring signal patterns may be detected, such as the pulse rate which is reflected in the cuff pressure signal. Consequently, correlation coefficients may be processed which are indicative of similarities between an original signal and its lagged version. For instance, the correlation coefficients may represent a result of the autocorrelation for a certain delay (or, equivalent ly, time lag). Assuming that the pulse rate is basically constant or quasi-constant for a defined period of time, a high correlation coefficient may be obtained at the time lag corresponding to the time between consecutive oscillations. As a result, the pulse rate may be obtained as the reciprocal of this lag. The pulse rate measurement module may comprise a signal processing section which may implement the signal separator, the autocorrelator and the pulse rate processor. Consequently, the above components may be implemented by hardware, by software, or by a combination thereof. Further, the signal processing section may form part of a main processing or control section of a vital signs monitoring system. Further, the pressure measurement unit that supplies the cuff pressure signal may be the unit that is anyway provided in the vital signs monitoring system. However, alternative embodiments may be envisaged wherein the pressure measurement unit is a distinct unit which, however, processes the cuff pressure signal (or a derivative thereof) which is anyway present in a blood pressure measurement arrangement. Further, the pressure measurement unit may be arranged to detect the cuff pressure signal in a direct or mediate fashion.

Further, the autocorrelator may be arranged to autocorrelate the separated potentially pulse rate indicative signal portion in a basically moving fashion. Consequently, time periods or moving windows may be defined for which the autocorrelation is performed, thereby having a repeated indicator of the pulse rate.

In an exemplary embodiment of the module, the signal separator is arranged to separate the potentially pulse rate indicative signal portion from a non- indicative signal portion obtained from an overall pressurizing regime. By way of example, the cuff pressure signal may involve pulse rate indicative cuff oscillations and/or derivative signals. In a refinement of the above embodiment, the signal separator is arranged to filter the potentially pulse rate indicative signal, wherein for instance a bandpass filter may be applied thereto. The bandpass filter may attenuate or remove signal portions outside a defined frequency range which represents expected pulse rate values, e.g. 30 to 180 beats per minute (BPM).

In addition, or in the alternative, the signal separator may be arranged to remove a defined signal portion which may for instance represent a ramp of the cuff pressure signal at the inflation or deflation stage. Hence, a curve fitting procedure may be applied to the cuff pressure signal, for instance a polynomial fit. For instance, depending on the signal characteristics, a first degree polynomial fit may be applied, wherein the respective signal portion may be removed or attenuated. Further, in some embodiments, a second degree polynomial fit may be applied. Needless to say, further curve fitting measures may be applied so as to define a non-indicative signal portion and to attenuate or remove the same.

In another embodiment of the module, the autocorrelator is arranged to detect pulse rate indicative periodic patterns in the potentially pulse rate indicative signal portion. In yet another embodiment of the module, the autocorrelator is further arranged to repetitively apply a running autocorrelation based on a defined moving time interval. In a refinement of the above embodiment, particularly for adult subjects, the defined moving time interval is in the range of 2.0 s to 10 s (seconds) wherein a repetition frequency of the autocorrelation is in the range of about 0.5 Hz (Hertz) to about 200 Hz. It should be ensured that the moving time interval is long enough to cover at least two complete pulse cycles.

In another exemplary embodiment of the module, the signal separator is arranged to separately process the moving periods, prior to the autocorrelation.

In still a further embodiment of the module, the autocorrelator is further arranged to derive a correlation coefficient for time lags corresponding to the pulse rate range to be measured.

In yet another embodiment of the module, the pulse rate processor is further arranged to identify pulse rate indicative correlation coefficients. This may involve, at least in some embodiments, that the pulse rate processor is further arranged to detect and remove repetitive patterns in the correlation coefficients that are related to multiples of the pulse rate, and to assess corresponding correlation coefficients so as to detect the pulse rate of interest.

In yet another embodiment of the module, the pulse rate processor is arranged to remove repetitive signals in accordance with the following equation:

ACco_r lag) = AC(lag) - max(AC (lag/2); AC(lag/3); AC(lag/5)) wherein ACcorr is a processed corrected autocorrelated curve, wherein an original AC curve is time-scaled (expanded in time) by a factor of 2, 3 and 5, and wherein the respective time-scaled AC curves are subtracted from the original AC curve.

In yet another embodiment of the module, there is further provided a pulse rate validity processor arranged to assess the validity of a potential pulse rate indicative correlation coefficient. In an exemplary refinement, the pulse rate validity processor is further arranged to process a signal stability indicative signal. In this way, validity assessment may be performed which further improves the pulse rate measurement accuracy and reliability.

In a refinement of the above embodiment, validity assessment is based on at least one of the following:

determination of an absolute value of a candidate correlation coefficient, determination of a relative value of a candidate correlation coefficient, and determination of an amount of time passed since the start of the pulse rate estimation relative to the processed period. Needless to say, validity assessment may be based on a combination of these approaches.

Preferably, the validity assessment further comprises the verification of whether the determined values comply with defined thresholds.

The passed amount of time may be indicative of the validity of the correlation coefficient since the passed time period is indicative of a covered number of heart beats. For instance, it may be defined that at least three heart beats are required so as to reliably process the pulse rate. Therefore, the processed pulse rate value (frequency or beats per minute) may be linked to the time period underlying the autocorrelation. For instance, assuming that a pulse rate of 40 BPM (beats per minute) is calculated, three single beats would basically span a period of 4.5 s (seconds). Consequently, in accordance with the above criteria, at least 4.5 s should have been passed before the pulse rate of 40 BPM is considered acceptable. In a further example, if pulse rate estimation is started at t = To seconds and the method requires to have detected at least three pulses since T₀ before the measured pulse rate can be accepted as valid, a pulse rate of 60 beats/minute within 3 seconds after starting time (i.e., t = T₀ + 7 s) cannot be accepted as valid. On the other hand, a measured instantaneous pulse rate of 120 beats/minute may be accepted as valid after 1.5 seconds from the starting time (i.e. t = To + 1.5 s).

Further, in accordance with another embodiment, validity assessment is based on further processing the potentially pulse rate indicative signal portion, preferably of a defined period thereof, and to assess whether processed values comply with defined thresholds, wherein said further processing involves at least one of the following:

determination of a signal amplitude or energy of the processed period, - determination of a signal variance of the processed period,

determination of a signal to noise ratio of the processed period, and processing morphologic features of the processed period.

Needless to say, the above-indicated validity assessment measures may be regarded as alternatives, but may be also applied in any appropriate combination.

In yet another refinement of the module, the pulse rate validity processor is further arranged to process a signal stability indicative signal, wherein an observation period for processing the signal stability indicative signal is preferably longer than the defined period for the autocorrelation. Preferably, at least in some embodiments, the observation period covers a multiple of the defined period for the autocorrelation. In the alternative, a fixed predetermined period of time may be used which is sufficiently long to cover a number of autocorrelation periods. In other words, in order to classify an estimated pulse rate value as a valid one, it is assessed whether the pulse rate value can be estimated for a prolonged period of time, at least within a defined tolerance range for the value. In addition, it is assessed whether the involved autocorrelation-based features are sufficiently stable and consistent over time.

Generally, the signal stability indicative signal may represent temporal stability of one of the above-indicated validity assessment criteria, preferably in relation to a defined observation period. Hence, validity assessment, on the one hand, involves the derivation and/or determination of validity values and, at least in some embodiments, further tracks and observes the temporal stability of the validity values.

In a further aspect of the present disclosure, a vital signs monitoring system arranged for blood pressure measurement is presented, the system comprising:

a wearable cuff adapted to pressurize a measurement site of a subject, - a pressurizing unit arranged to supply and operate the wearable cuff,

a pulse rate measurement module in accordance with at least one embodiment as disclosed herein, the pulse rate measurement module being operatively coupled with the wearable cuff, and

a blood pressure determination module operatively coupled with the pulse rate measurement module, the blood pressure determination module being arranged to calculate a blood pressure value based on the cuff pressure signal detected by the pressure measurement unit when the cuff pressure is increased or decreased by the pressurizing unit in accordance with a defined pressurizing regime,

wherein the pressurizing unit is arranged to apply a pressurizing regime dependent on the measured pulse rate.

Consequently, a blood pressure measurement arrangement may be provided which is arranged for adjusting an inflation or deflation rate dependent on another vital sign so as to improve the measurement accuracy and to reduce the subject's discomfort during measurement, particularly by speeding up the measurement.

Preferably, the vital signs monitoring system is exclusively provided with cuff pressure sensors, preferably with a single cuff pressure sensor. In other words, no additional separate heart rate or pulse sensor is provided. Hence, blood pressure measurement may be improved in a cost-efficient way. In a refinement of the above system, a control section is provided which is operable to inflate the cuff at an inflation rate as a function of a measured pulse rate, wherein the pulse rate is calculated prior to the blood pressure measurement based on the cuff pressure signal.

Hence, in accordance with the above embodiment, at an initial stage of the measurement, an initial pulse rate value is measured based on which an initial inflation rate is defined. However, also within the inflation stage as such, when the cuff is pressurized, pulse rate measurement may continue. Consequently, a current inflation rate may be adapted accordingly.

In another embodiment, the pulse rate calculation or detection is initiated in a lagged fashion upon starting the inflation of the cuff. Consequently, a low-pressure state of the cuff in which basically no indicative pressure signal may be detected can be omitted.

Generally, the pulse rate measurement may be performed while inflating or deflating the cuff, wherein it is preferred that a corresponding pressure ramp signal is removed for further processing. However, at a generally steady state of the cuff pressure, pulse rate measurement may be performed as well, provided that cuff oscillations are present which may be detected.

In another embodiment of the above system, the blood pressure determination module is operable to determine the pulse rate indicative signal at a level of the cuff pressure which is below a defined minimum pressure required for the calculation of the blood pressure value. Hence, it may be ensured that at least an initial measurement of the pulse rate can be performed before the blood pressure measurement actually begins.

In another embodiment of the above system, the pressurizing regime includes an inflation rate proportional to the measured pulse rate. In one exemplary embodiment, the inflation rate is preferably determined in accordance with the following equation:

Ri [mmHg/sJ = k [mmHg/beat] x HR [BPMJ /60 wherein Ri is the inflation rate,

wherein HR is the measured pulse rate,

wherein k is the proportionality factor, and wherein the factor k is in a range of 2 to 20. Preferably, the factor k is in a range of 5 to 15. More preferably, the factor k is in a range of about 7 to 12. In yet another aspect of the present disclosure, a method for pulse rate measurement for use in a vital signs monitoring system is presented, the method comprising:

detecting a cuff pressure signal of a wearable cuff which is arranged to pressurize a measurement site of a subject,

- processing the pulse rate, including:

extracting a potentially pulse rate indicative signal portion from the cuff pressure signal,

auto correlating the separated potentially pulse rate indicative signal portion, and

- processing the pulse rate based on temporal information linked with processed correlation coefficients.

In yet another aspect of the present disclosure, a method for controlling the inflation rate of a wearable cuff for an inflation-based non- invasive blood pressure measurement system is provided, the method comprising the steps of:

- operating a pressurizing unit, particularly a pump or valve, to supply pressure to a wearable cuff which is arranged to pressurize a measurement site of a subject, wherein the pressurizing unit is arranged to apply a pressurizing regime dependent on a control signal, detecting the pulse rate according to a method in accordance with at least one embodiment of the pulse rate measurement method as disclosed herein, and

- processing and controlling the control signal of the pressurizing unit to set an inflation rate of the wearable cuff as a function of a measured pulse rate.

In yet another aspect of the present invention there is provided a computer program which comprises program code means for causing a computer to perform the steps of the methods as discussed herein when said computer program is carried out on that computer.

The program code (or: logic) can be encoded in one or more non-transitory, tangible media for execution by a computing machine, such as a computer. In some exemplary embodiments, the program code may be downloaded over a network to a persistent memory unit or storage from another device or data processing system through computer readable signal media for use within the system. For instance, program code stored in a computer readable memory unit or storage medium in a server data processing system may be downloaded over a network from the server to the system. The data processing device providing program code may be a server computer, a client computer, or some other device capable of storing and transmitting program code. As used herein, the term "computer" may stand for a large variety of processing devices. In other words, also mobile devices having a considerable computing capacity can be referred to as computing devices, even though they provide less processing power resources than standard "computers". Needless to say, such a "computer" can be part of a medical device and/or system. Furthermore, the term "computer" may also refer to a distributed computing device which may involve or make use of computing capacity provided in a cloud environment. The term "computer" may also relate to medical technology devices, fitness equipment devices, and monitoring devices in general, that are capable of processing data.

Preferred embodiments of the disclosure are defined in the dependent claims. It should be understood that the claimed method and the claimed computer program can have similar preferred embodiments as the claimed system and as defined in the dependent system claims. BRIEF DESCRIPTION OF THE DRAWINGS

These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter. In the following drawings

Fig. 1 shows a general layout of an exemplary vital signs monitoring system; Fig. 2 shows a simplified block diagram illustrating an embodiment of a vital signs monitoring system implementing a pulse rate measurement module;

Fig. 3 shows a block diagram illustrating several steps of an embodiment of a pulse rate measurement method;

Fig. 4 shows a block diagram illustrating several steps of an exemplary embodiment of a method for operating a blood pressure measurement system;

Fig. 5 shows two charts exemplifying a signal autocorrelation in accordance with at least some embodiments as discussed herein,

Fig. 6 shows an exemplary chart illustrating correlation coefficients which are provided for further processing so as to obtain pulse rate information,

Fig. 7 shows a block diagram illustrating several steps of an embodiment of an exemplary algorithm for a pressure based detection of the pulse rate of a subject, and

Fig. 8 shows several charts exemplifying autocorrelation signals and signal processing/enhancing measures applied thereto. DETAILED DESCRIPTION OF THE INVENTION

In accordance with the current disclosure, embodiments of pulse rate measurement modules and methods are provided which may be implemented in an integral fashion in vital signs detection and monitoring systems. As used herein, an integral (or embedded) arrangement involves that no additional sensors are required for the pulse rate measurement. More particularly, already present or anyway required sensors may be operated and utilized so as to process the desired pulse rate signal which even may be regarded as a by-product. However, the pulse rate measurement significantly enhances and facilitates the operation of respective vital signs monitoring systems.

In the following, main aspects and insights of the invention will be illustrated and further elucidated with reference to exemplary embodiments of inflation-based NIBP methods and devices. It should be noted that the following exemplary embodiments and descriptions shall be not interpreted in a limiting sense. Rather, the skilled person may readily transfer and broaden the respective specific embodiments and the components as well as process steps disclosed therein so as to arrive at the general concept of the present disclosure.

More particularly, further fields of application can be seen for the pulse rate measurement approaches discussed herein. Needless to say, also deflation-based NIBP methods and devices may profit from an integrated pulse rate measurement.

Reference is made to Fig. 1 illustrating an exemplary measurement environment for a vital signs monitoring system 10. Further reference is made to Fig. 2 showing a block diagram illustrating an exemplary layout of another embodiment of a vital signs monitoring system 10. As indicated above, the vital signs monitoring systems 10 are configured for illustrative purposes as blood pressure measurement systems. However, the systems 10 may serve further, different purposes in other applications. Needless to say, combined vital signs monitoring systems may be envisaged, particularly in a clinical environment.

With respect to the general structure and to general features of exemplary conventional inflation-based NIBP devices, reference is made to US 2014/0309541 Al disclosing an oscillometric blood pressure measurement device which may be operated to measure the blood pressure at the inflation stage of a cuff.

In accordance with the simplified illustration of Fig. 1, the system 10 includes a control device 12 and a clamp unit or cuff 14 which may be attached to a measurement site of a subject 18, particularly at a body limb 20 thereof. Typically, the cuff 14 may be attached to an upper arm or forearm of the subject 18. The cuff 14 may be coupled with the control device 12 via at least one supply line 16 which is arranged for pressurizing the cuff 14 and, consequently, the measurement site of the subject 18. Supply lines 16 may be provided for inflating and for deflating the cuff 14. For inflating the cuff 14, a pump may be provided at or may be coupled with the control device 12. For deflating the cuff 14, a deflation valve may be provided at or coupled with the control device 12.

Further, the control device 12 may implement or may be coupled with at least one pressure sensor so as to detect the current pressure of the cuff 14. Typically, the cuff 14 is supplied with air or another appropriate fluid.

An exemplary layout of a control device 12 for a vital signs monitoring system 10 is illustrated in Fig. 2 by means of a simplified block representation.

As already indicated above, the control device 12 and the cuff 14 that may be arranged to be wrapped around a limb 20 of the subject may be connected via at least one supply line 16 which may be also referred to as pressure supply line.

In some exemplary systems, an interface between the cuff 14 and the control device 12 may be provided such that both entities may be separated.

In accordance with at least some aspects of the present disclosure, the control device 12 comprises a pressurizing unit 24 which may comprise a pump or at least one inflation and/or deflation valve. Further, a pressure measurement unit 26 is provided which is configured for detecting a current pressure at the measurement site which is attributable to the subject's blood pulsation. Since the pressure measurement unit 26 is arranged for measuring/monitoring the pressure of the pressurizing fluid, blood pressure measurement is performed in a mediate way.

The pressurizing unit 24 may be arranged to pressurize the cuff 14 in accordance with a defined pressurizing regime. This may involve a defined inflation regime and/or a defined deflation regime. Further, the pressurizing unit 24 may be arranged for keeping or maintaining a basically constant pressure level at the cuff 14. Due to blood pulsation at the subject, slight oscillations may be present in the pressure signal detected by the pressure measurement unit 26. Accordingly, the pressure measurement unit 26 is arranged for deriving a blood pressure indicating value, as for instance explained in US 2014/0309541 Al.

Further, as discussed within the context of the present disclosure, a cuff pressure signal 30 provided by the pressure measurement unit may be delivered to a blood pressure determination module 28 which is arranged to calculate and process the desired blood pressure signal. For instance, systolic and diastolic blood pressure values may be signals of interest. Further, a mean arterial pressure (MAP) may be calculated based on measured systolic and diastolic blood pressure values. To this end, the blood pressure determination module 28 is arranged for detecting and processing the so-called cuff oscillations that are attributable to blood pulsations at the measurement site.

At the pressurizing unit 24, a pressure control unit 34 is provided which is arranged for controlling a pump or an input valve of the pressurizing unit 24 so as to apply a defined pressurizing regime to the cuff 14.

As an input value for the pressure control unit 34, a pressurizing signal 32 may be provided which may be particularly influenced by the pulse rate of the observed subject 18. Based on the measured pulse rate, the to-be-applied inflation rate and/or deflation rate or another operating parameter of the pressurizing unit 24 may be adapted and adjusted so as to improve the measurement procedure.

A pulse rate measurement module 40 is provided which is arranged for receiving the cuff pressure signal 30 and to process/calculate the desired pressurizing signal 32.

Via at least one interface 42, the cuff pressure signal 30 may be supplied to the pulse rate measurement module 40. The pulse rate measurement module 40, as shown in the exemplary embodiment of Fig. 2, comprises a signal processing section 44 which implements a signal separator 46, an autocorrelator 48, a pulse rate processor 50, and a pulse rate validity processor 52. The respective components of the signal processing section 44 may be implemented by hardware, by software, or by an appropriate combination thereof.

Generally, the control device 12 may be provided with processing capacity so as to implement the processing and control function of the pressurizing unit 24, the pressure measurement unit 26, the blood pressure determination module 28 and/or the pulse rate measurement module 40. However, at least some components may be implemented in discrete form. A common processing arrangement or capacity may be provided so as to process the signals required for operating the pressurizing unit 24, the pressure measurement unit 26, the blood pressure determination module 28 and the pulse rate measurement module 40.

The signal separator 46 is arranged for extracting a potentially pulse rate indicative signal portion of the cuff pressure signal 30. To this end, the signal separator 46 may be arranged for splitting the cuff pressure signal 30 into a pulse rate indicative portion and a non- indicative portion. Generally, the signal separator 46 may be arranged for removing or attenuating non-indicative signal portions. The autocorrelator 48 is arranged to apply an autocorrelation processing to the signal provided by the signal separator 46. Consequently, periodic (harmonic) patterns in the signal may be detected which are potentially attributable to the pulse rate.

To this end, a pulse rate processor 50 may be provided which is arranged to examine the signal provided by the autocorrelator 48 so as to detect prominent frequencies which may be also referred to as pulse rate values or pulse rate candidate values.

At least in some embodiments, a pulse rate validity processor 52 is provided so as to assess the validity of the pulse rate values or pulse rate candidate values provided by the pulse rate processor 50.

Exemplary embodiments and refinements of the pulse rate measurement module 40 and associated methods will be introduced and discussed further below.

Fig. 3 shows, by means of a simplified block diagram, a method of pulse rate measurement which is suitable for being implemented in a vital signs monitoring system.

Initially, at a step S10, a cuff pressure signal of a wearable cuff (blood pressure cuff) is detected which serves as an input signal for the pulse rate measurement. For instance, the cuff pressure signal may be supplied via an interface.

Further, a pulse rate processing step S12 may follow which may comprise several substeps S14, S16, S18 and S20.

In a first substep S14, a potentially pulse rate indicative signal portion may be extracted from the blood pressure signal. To this end, filtering, signal splitting, signal portion removal and further signal processing approaches may be utilized. Since a general level of the cuff pressure signal is at least approximately known, respective signal portions (ramps and/or steady portions) may be removed such that a remaining portion is at least substantially composed of cuff oscillations which are attributable to the blood pulsation of the subject.

A further substep S16 addresses signal processing by means of autocorrelation. More particularly, the separated potentially pulse rate indicative signal portion is autocorrelated so as to detect prominent periodic signal patterns which indicate the desired pulse rate.

Consequently, in a further substep SI 8, autocorrelated signals may be further processed so as to derive temporal and/or frequency information therefrom which is linked with dominant correlation coefficients. Consequently, the pulse rate may be calculated. At least, pulse rate candidates may be obtained. Therefore, at least in some embodiments, a validity assessment substep S20 may follow which includes applying validity assessment measures to pulse rate candidates. Preferably, validity assessment also involves assessing an overall signal stability which may involve a temporal signal stability.

Fig. 4 illustrates, by means of a simplified block diagram, a method of blood pressure measurement, which involves a detection of a pulse rate of the monitored subject. The method involves a step S50 which includes operating a pressurizing unit so as to supply pressure to a cuff which may be wrapped around a limb of a subject. The cuff is arranged to pressurize a measurement site which enables the detection of so-called cuff oscillations which are attributable to the subject's blood pulsation. Operating the pressurizing unit may involve applying a defined pressurizing rate or pressurizing regime.

In a further step S52, based on a cuff pressure signal which is indicative of a current pressure at the cuff, a - so to say - auxiliary or supplemental signal may be detected, namely the pulse rate of the subject. Hence, the step S52 may involve the steps S10 to S20 as discussed in connection with Fig. 3. As a result, a pulse rate may be provided.

In a further step S54, the wearable cuff, particularly the pressurizing unit associated thereto, may be processed and controlled dependent on the measured pulse rate. For instance, an inflation and/or deflation rate of the wearable cuff may be set as a function of the measured pulse rate. This may have the advantage that a beneficial trade-off between signal accuracy, measurement accuracy reliability, and measurement time may be achieved.

A further step S56 may follow which involves the calculation and derivation of the subject's blood pressure. Again, reference is made to US 2014/0309541 Al .

Further reference is made to the illustrative exemplary signal charts shown in Fig. 5 and in Fig. 6.

In Fig. 5, a lower chart represents a cuff pressure signal 80 illustrating exemplary cuff oscillations. An upper chart of Fig. 5 illustrates a dot plot 82 of correlated signals based on the cuff pressure signal 80. The chart illustrating the cuff pressure signal 80 includes a time axis 84 and a pressure axis 86. Further, windows 88 are indicated illustrating respective time periods based on which the autocorrelation of the cuff pressure signal 80 is performed. Hence, the charts 80, 82 in Fig. 5 are shown in a synchronized state.

Consequently, also the dot plot 82 includes a time axis 90 which basically corresponds to the time axis 84 of the cuff pressure signal 80. A correlation strength signal is indicated by an axis 92. Based on the correlation strength signal, a correlation coefficient plot may be formed, refer also to the plot 100 of Fig. 6. Signal dots 94 form the dot plot 82, wherein a respective signal dot 94 is associated with a corresponding time period or window 88. The dots 94 represent selected correlation coefficients from the chart illustrated in Fig. 6. Further reference is made to Fig. 6 illustrating an exemplary correlation coefficient plot 100 which may be obtained based on the autocorrelation indicated in connection with Fig. 5.

The plot 100 includes a lag axis 102 which may be referred to as an inverse of a frequency axis. Consequently, in a direction towards the origin of the time axis 102, frequency values become higher while corresponding (inverse) time values which are indicative of a (time) lag become shorter. Therefore, for illustrative purposes, also a frequency axis 112 and a pulse rate axis 114 is provided in Fig. 6 to illustrate the relationship between lag time (in seconds), frequency (in Hertz) and pulse rate (in beats per minute).

Further, the correlation coefficient plot 100 is characterized by a correlation coefficient strength axis 104. A frequency range of interest is indicated in Fig. 6 by 106. The frequency range of interest covers for instance a range between 30 BPM (corresponding to 0.5 Hz and to 2.0 s lag) and 180 BPM (corresponding to 3.0 Hz and to 0.33 s lag) which may be regarded as a general example for potential pulse rate values of a (human) subject.

In Fig. 6, exemplary correlation coefficient values are indicated by PI, P2, P3,

P4 and P5.

Several signal processing measures may be applied to the signal 100 within the range 106 so as to attenuate or remove basically non- indicative signal portions. By way of example, (decreasing or increasing) signal slopes close to the borders of the window 106 may be removed. For instance, the decreasing slope between the values PI (border crossing) and P2 (local minimum) may be removed since no peak value is present in that signal portion.. Rather, the peak value is outside the potentially indicative frequency window 106. Hence, even if for instance the point PI exhibits the highest signal level within the window 106, it is readily apparent that the point PI is not indicative of the desired pulse rate signal. Basically the same applies to the increasing slope between the point P3 and a low-frequency border of the window 106.

The point P3 indicates a zero-crossing of the signal 100 which may be used as a further indicator for a signal portion which may be removed.

Fig. 7 shows, by means of a further simplified block diagram, an exemplary algorithm for the detection of the pulse rate of a subject based on a detected cuff pressure signal.

A step SI 00 addresses signal conditioning. More particularly, a cuff pressure indicative input signal may be processed accordingly. Signal conditioning may involve a signal filtering procedure, for instance a band-pass filtering or a comb filtering. In addition or in the alternative, mean signal portions may be removed so as to enhance high-frequency signal fluctuations within a defined pulse rate frequency band.

Further, an already applied pressure regime (e.g. inflation ramp, deflation ramp, steady portion) may be removed from the received signal. The step SI 00 may be also referred to as signal separation step.

In a further step SI 02, an autocorrelation of the signals computed in the step SI 00 is performed. The step SI 02 may involve a normalization which can be achieved by dividing obtained autocorrelated signals by a base-autocorrelation at a lag of t = 0 s

(seconds).

A further step SI 04 may follow which includes further processing the autocorrelated signals. By way of example, a defined frequency range (lag range) may be applied so as to define a generally indicative period. Further, defined signal portions (e.g. slopes at the edges or boundary portions of the defined frequency range may be removed.

A further autocorrelation processing may involve the removal of so-called double-peaks. For instance, the removal of the double peaks may be performed in accordance with the following equation:

ACco_r lag) = AC(lag) - max(AC (lag/2); AC(lag/3); AC(lag/5)) wherein AC_corr is a processed corrected autocorrelated curve, wherein an original AC curve is time-scaled (expanded in time) by a factor of 2, 3 and 5, and wherein the respective time-scaled AC curves are subtracted from the original AC curve.

In this context, additional reference is made to Fig. 8 showing several charts I, II and III illustrating the above approach. In the charts I, II and III, axes 150, 160, 170 are time or lag axes. Axes 152, 162, 172 are signal strength axes indicating the strength of the processed correlation coefficient signals dependent on the lag. In the chart I, a signal plot 154 is provided. In the chart II, a signal plot 164 is provided which is based on the signal plot 154 of chart I. In the chart III, a signal plot 174 is provided which is based on both the plots 154 and 164 of the charts I and II. The plot 154 is a correlation coefficient plot and may basically correspond to the signal plot 100 of Fig. 6. In the plot 154, multiple-peak (or double-peak) signals may be present, exemplified by signal peaks 156, 158 in the chart I. The presence of double peak signals 156, 158 may complicate the derivation of the desired pulse rate as integer multiples of the period that corresponds to the real rate are also amplified in the autocorrelation procedure. In accordance with the above aspect, a time-scaled (expanded) signal (plot 164) may be used which is based on the original signal (plot 154). In other words, assuming that the lag axes 150, 160 show the same scale, an actual signal strength value at a certain lag of the plot 164 may for instance correspond to a signal strength value at half the lag (1/2 lag) of the plot 154, see the position of the peak 156 in the charts I and II. As shown above, also further fractions of the lag (e.g., 1/2, 1/3 and 1/5 in the above equation) corresponding to integer multiples of the period may be used.

The chart III illustrates a corrected autocorrelation signal plot 174 resulting from a subtraction of the plot 164 of the chart II from the plot 154 of the chart I. Lag values 178, 180 indicated at the lag axis 170 correspond to the position of the peaks 156, 158 of the plot 154 at the lag axis 150 in the chart I. It can be seen that the peak 158 basically has been canceled out in the resulting plot 174. A remaining peak 176 basically corresponds to the peak 156 of the original plot 154. However, the position indication 178 of the original peak 156 does not necessarily have to perfectly match the position of the peak 176 at the lag axis 170, due to the processing.

Further, clipping may be applied to the AC curves so as to obtain positive values. In this context, further reference is made to T. Tolonen and M. Karjalainen: A computationally efficient multi pitch analysis model. IEEE Trans, on Speech and Audio Processing, 8(6):708-716, 2000. In this reference, appropriate signal processing measures are presented and explained in the context of audio signal processing. It has been observed that the above approach is a suitable measure for the enhancement and dressing of the

autocorrelation signals described herein.

Needless to say, alternative processing measures may be applied to the autocorrelation signals.

Again referring to Fig. 7, a further step SI 06 may follow which involves a computation of a time period which corresponds to the lag that is associated with the detected autocorrelation peak. The computed period is a candidate period for the desired pulse rate signal.

In a further step SI 08, validity assessment may be applied which involves a determination of a validity of the period processed in the step S 106.

The assessment of the validity of the computed period may be based on one or more of the following:

variance of recently estimated periods,

signal duration passed since the start of the period estimation, calculated energy and/or amplitude of the signal,

computation of a normalized peak of the autocorrelation,

computation of a normalized peak difference of the autocorrelation (for instance, computation of a peak value minus neighboring values within a lag of about 0.2 s), and

time-averaged peak and/or peak difference of the autocorrelation. An estimation step SI 10 may follow wherein it is estimated whether or not a period is valid, in case a period is found to be sufficiently valid, the pulse rate may be calculated as an inverse of the computed period. The pulse rate may be specified as a BPM value (beats per minute) or as a Hz value (frequency in Hertz).

In the following, further aspects of the present disclosure are explained and discussed with reference to embodiments in the field of medical technology, particularly in the field of blood pressure measurement. Respective arrangements are known as NIBP devices (non-invasive blood pressure measurement). Further, so-called iNIBP arrangements are known which process the blood pressure during the inflation stage of the cuff.

In the context of NIBP, the inflation and/or deflation control (generally referred to as pressurizing control) has a main influence on the patient's (referred to as subject herein) discomfort and/or the measurement accuracy. It has been observed that pressurizing control dependent on the heart rate (pulse rate) of the subject is a promising measure.

Therefore, main aspects and embodiments of the present disclosure relate to novel approaches to pulse rate measurement based on a running autocorrelation.

In the following, a pulse rate measurement method is described which is related to an inflation (more generally: pressurization) of a NIBP cuff at a defined rate related to the pulse rate. This ensures a certain target number of beats within an envelope (to determine DBP and SBP, as discussed above). Therefore, in accordance with this method, inflation based NIBP is improved and no separate envelope determination is required with an additional method (e.g. step down deflation), while subject discomfort is minimized.

However, the pulse rate measurement as discussed herein may be used as well in the context of deflation based NIBP applications.

In accordance with exemplary embodiments, the detection of the pulse rate comprises the following main steps:

separation of the inflation ramp and cuff oscillations,

calculation of a running autocorrelation of the cuff oscillations to determine the pulse rate. checking or assessing the validity of the pulse rate.

Further, the pulse rate measurement may be performed in several contexts, i.e. at different pressure and pressure rate levels. Further, the pulse rate-dependent inflation rate may be chosen differently.

The three steps as indicated above as well as further aspects and insights will be elucidated and further detailed in the following sections.

Separation of inflation ramp and cuff oscillations

In one embodiment, the set cuff pressure signal (base signal in accordance with the pressurizing regime) and cuff oscillations (attributable to the subjects blood pulsation) may be separated by use of a filter. To this end, several options may be envisaged.

In one embodiment, a band pass-filter is chosen that retains the signal frequencies that are related to the pulse rate.

In another embodiment of the signal separation, a signal ramp (e.g. inflation ramp) is removed from the oscillations by processing a polynomial fit of the cuff pressure signal and, consequently, removing the polynomial fit from the actual cuff pressure waveform. This removal may be done repetitively on a certain period of the data (as described further below in the running autocorrelation section, e.g. on periods of 4.0 s with 0.5 s overlap).

Preferably, the polynomial order is chosen such that the resulting polynomial describes the inflating cuff pressure, but does not include the cuff oscillations. For instance, a polynomial order of 1 (linear) or 2 (quadratic) could be taken. The advantage of this method is that the cuff oscillation morphology will not be disturbed, i.e. there is no different amplification at different frequencies.

Further, also a combination of the above may be applied at the signal separating stage.

Running autocorrelation of cuff oscillations

In accordance with this aspect, a running autocorrelation (as discussed above in connection with Fig. 6) of the pre-processed cuff oscillations (or their derivative) is processed. This may be achieved by selecting a certain period and repetitively calculating the autocorrelation of the signal in that period. The length of the period may be chosen such that a correlation can be calculated in a reliable fashion. In a preferred embodiment, this length corresponds to at least two beats (e.g. 4.0 seconds if the minimum pulse rate to be measured is 30 BPM). As the correlation, in practice, changes relatively slowly (e.g. every new heart beat, not every millisecond), the repetition frequency does not need to be very high. In one embodiment, a new value is calculated when a complete new beat is available (i.e. every 0.33 s if the maximum to be measured pulse rate is 180 BPM).

In one embodiment, at an initial stage of the measurement, the period length may be gradually increased, i.e. not starting at the preferred period length (e.g. 4.0 s), but starting for instance at about 1.0 second and increasing the period every new measurement cycle until the preferred period length is reached.

In this context, the above introduced Fig. 5 illustrates an example of the running autocorrelation: the correlation is calculated for a certain period and gives a single value (i.e. the dots in the upper chart of Fig. 5). After a certain period of time an update is done on newly available data, thereby constructing a running autocorrelation signal as illustrated in the upper chart of Fig. 5.

In one embodiment, the pressure ramp removal as discussed above (e.g. by fitting a polynomial function to the pressure ramp and subtracting this function from the actual pressure), also may be done repetitively on these blocks of data, i.e. the ramp may be removed repetitively for each autocorrelation period.

Further, the calculated autocorrelation values are processed to obtain a potentially valid measure of the pulse rate (refer to Fig. 6 as already discussed above). This may involve one or more of the following approaches.

In one embodiment, the autocorrelation coefficient of the cuff oscillation signal is calculated for lags (corresponding to an inverse of respective frequencies) corresponding to a pulse rate between a defined minimum and maximum, resulting in an autocorrelation function. The actual calculation period may be a longer period but preferably values between the pulse rate limits are considered. Other values may be removed.

Values at the start and end of the autocorrelation function may be removed if they are associated with peaks outside the interval/range of interest. This may involve one ore more of removing all declining values at the left side of the correlation, up to the first non- declining value and/or removing all increasing values at the right side of the correlation, starting from the last decline. Further, instead of removing up to the first non-declining / from the last declining value, the declining / increasing values up to the first / from the last zero crossing may be removed which gives similar results. The signal conditioning measures discussed herein relate to a temporal signal chart. A frequency-based signal chart would involve an inverse representation and therefore require similar (possibly inverse) signal measures. Hence, at a detected period (corresponding to a certain pulse rate) at the maximum correlation is a potential pulse rate to be found. As indicated above, due to the nature of the autocorrelation, additional peaks may be found at integer multiples of the detected period (i.e. the lag corresponding to the true period). Hence, these peaks correspond to integer fractions (1/2, 1/3, 1/4, etc) of the detected pulse rate. In order to prevent detection of the pulse rate at the incorrect peak, e.g., that at half the true pulse rate, the pulse rate may be updated, e.g. to double the potential pulse rate, if the correlation at said double frequency (double pulse rate) is large enough (e.g. in relation to the correlation at the potential pulse rate, as a fraction, or as an absolute difference). This may be applied as, in some cases, the highest correlation value may be found at double the pulse rate period. For example, if the true period is 1.0 s (i.e., the pulse rate is 60 bpm), an incorrect peak may be observed at the lag 2.0 s (corresponding to a pulse rate of 30 bpm). If a peak at a lag of 2 s is detected, it should then be checked whether there is also a peak at 1 s, 0.67 s, 0.5 s, 0.4 s... (i.e., at the lag values for which 2.0 seconds is an integer multiple, namely 2/2, 2/3, 2/4, 2/5...).

In this context, the above introduced Fig. 6 elucidates an example of a correlation coefficient chart illustrating correlation coefficients as a function of the lag (frequency) between the processed autocorrelated signal samples. Given the exemplary correlation coefficients of the cuff oscillation signal, only the values between a defined minimum and maximum pulse rate are kept. In the exemplary embodiment, the defined range is between 30 and 180 BPM, as indicated by reference numeral 106. For the sake of clarity, it is worth noting that in Fig. 6 the highest pulse rate value is closest to the vertical axis, as the horizontal time axis is measured in a lag period, not in a frequency.

As a further processing measure, the autocorrelation period is further narrowed, so as to remove high values that are related with peaks outside the defined range/interval of interest. For instance, in Fig. 6 the value PI is the highest value, but is related to a maximum at lag 0 and is therefore not regarded and treated as a potentially indicative peak.

Consequently, values to the right of the first value (at 180 BPM in Fig. 6) that have a declining slope (i.e. until the local minimum at value P2) may be removed. A similar removal is done at the right side of the autocorrelation (from 30 BPM to the left in this example). Instead of looking for the first increasing value, one may also search for a zero crossing (as done on the right side at value P3). The remaining signal (i.e. the dashed curve between values P2 and P3 in the chart) is the autocorrelation function which contains candidate peaks of interest. Accordingly, a search for the maximum peak is performed (value P4 in this example). This peak may in some cases however be related to double the length (half the frequency) of the correct period. As a repetitive signal is processed in the autocorrelation, repetition at double the correct period is also present.

Therefore, at least in some embodiments, if the peak at half the potential pulse rate period (value P5) in this exemplary chart of Fig. 6 is sufficiently high (e.g. a certain factor of value P4 or a fixed offset from value P4), then the value P5 is taken as the actual correct pulse rate instead of the value P4. If this criterion is not met, the value P4 may be taken as the potential pulse rate. Any potential pulse rate is associated with a correlation coefficient (i.e. the amplitude of the peak of value P4 or value P5) which involves information on the validity of the peak.

For instance, to detect that point P4 corresponds to the correct pulse, an algorithm is suggested which assesses whether another local maximum (i.e. value P5) is present in the correlation curve at about twice the frequency (or half the lag). If a value is found, then for instance a ratio between the amplitudes of value P4 and value P5 is calculated. Further, it may be assessed whether the calculated value is above a given threshold. If so, the measured pulse rate may be considered as being correct. Otherwise, the measured pulse rate candidate is discarded. Basically, this algorithm avoids to mistakenly regard point P5 as being the measured pulse rate, in case, for some reason, it would present an amplitude larger than that of P4.

Assessment of the validity of the measured pulse rate

The validity of the found pulse rate for each period is verified by one or more of the below described approaches.

Absolute magnitude of the autocorrelation coefficient

In accordance with this exemplary approach, an absolute value of the autocorrelation as such gives a measure of validity of the found pulse rate. A high value may be considered as an indicator of a high reliability of the found pulse rate; a low value respectively indicates a low reliability. A defined threshold value may be used to decide whether a found pulse rate is valid or not.

Relative magnitude of the autocorrelation coefficient

This exemplary approach is suitable in cases were the correlation value is always high, for instance when some DC signal component is still present. Hence, a relative magnitude assessment may provide some indication on how peaky or smooth the found correlation is. This may be for instance performed by comparing the correlation coefficient at the candidate pulse rate with the minimum or mean value of the observed correlation period. A resulting value, e.g. a ratio, a fraction or an absolute difference, may be defined as a measure of a level of unevenness, where a corresponding threshold may allow a conclusion as to whether a found pulse rate candidate is valid or not.

Signal Energy and related values

In accordance with this exemplary approach, a criterion on signal energy may be provided. For instance, signal energy may be processed in addition to absolute magnitude information or relative magnitude information. Hence, even in situations where the absolute /relative correlation value as such may indicate a certain validity level, potentially wrong conclusions may be avoided. For example, in the absence of heart beats, the main period of the noise within the signal might be found and mistakenly associated with a pulse of the monitored subject. A signal energy measure may be calculated on the cuff oscillation itself, and/or its derivative, for the same period to which the correlation was applied. In the context of signal energy assessment, different features of signal energy may be used.

For instance, a value corresponding to a maximum minus minimum value of a period of interest (or a part thereof) may be processed. A threshold value may be defined so as to allow a conclusion as to whether or not a minimum amplitude level is reached. The threshold value may be obtained from statistical data and/or empirical knowledge.

For instance, a value corresponding to the signal energy level may be processed, e.g. an integral value, a variance value, a standard deviation value, a root mean square value, etc. Again, a threshold value may be defined so as to allow a conclusion as to whether or not a minimum energy is reached.

For instance, a signal to noise ratio may be used which may be calculated by dividing the signal energy by a noise energy. Signal and noise energy may be separated from the cuff oscillation signal by applying a high/low-pass filtering to the cuff oscillation signal and using the low-pass filtered signal as the actual energy signal and the high-pass filtered signal as the noise signal. The cut-off frequency should be chosen such that the desired oscillation frequency is basically contained in the actual (low-pass filtered) signal.

For instance, morphologic features of the signal may be used for validity assessment. As used herein, morphologic features involve, but are not limited to, magnitude of derivatives, zero crossings, a comparison to a template pressure pulse, a combination thereof, and a distribution thereof. Signal morphology considerations may basically require certain knowledge of a number of heart beats. These and other features may be calculated over a complete correlation period. In the alternative, or in addition, these and the other features may be calculated for a part of the correlation period related to the minimum pulse rate to be found.

For example, it may be assessed whether the signal energy is present at the complete period or only at a part thereof. For instance, if an exemplary complete period lasts 5.0 s (seconds), the amplitude may be checked every 2.0 s (related to a pulse rate of 30 BPM to be found) to see if in every processed sub-period of the complete period the amplitude threshold is reached.

Steady state assessment

In accordance with this exemplary approach, temporal steadiness of validity measures and of the actual pulse rate itself may be assessed.

As already indicated hereinbefore, a general insight in this context is that the pulse rate is quasi-stable over time, given a typical monitoring timeframe and environment. Therefore, a first, general assumption is that a basically consistent pulse rate should be measured over time. In case short-term deviations, particularly fluctuations, are detected, a general indication for lacking signal consistency is present. Several consistency assessment measures may be envisaged, which may form alternatives or which may be applied in combination.

For instance, it may be assessed whether involved/detected/processed signal features meet certain thresholds for a prolonged period of time.

Further, it may be assessed whether time-averaged and/or low-pass filtered features meet certain thresholds. Preferably, the applied averaging time corresponds to or should be at least consistent with the above mentioned prolonged period of time.

For instance, a prolonged time period for steady state assessment may take 1.0 s, whereas the averaging time is at about 1.0 s, 1.5 s or 2.0 s.

Another option is to assess whether detected period values (corresponding to pulse rate values) are sufficiently stable over time. For instance, a variation of maximal 10 % over a timeframe of 1.0, 1.5 or 2.0 seconds may be an appropriate choice in an exemplary embodiment.

Steady state assessment may involve a determination whether the values are above/below a certain threshold and/or a determination whether the values deviate/drift less than defined by a certain threshold value. An assessment time period may be defined which may involve, for instance, 1.5 s which would correspond to 5 calculations of the above validity values if the repetition period of measurements is 0.3 s. Therefore, a steadiness- related validity criterion may indicate whether all calculated values are below/above a defined threshold and/or within a defined range. In case sufficient steadiness has been detected, a pulse rate candidate may be marked as a valid measure.

Generally, at least in some embodiments, determining validity of the detected pulse rate candidate may be regarded as a combination of one or more of the several criteria discussed further above, and the detected steady state condition as explained hereinbefore. The previous criteria provide information as to whether the measured pulse rate at a given instant meets a given magnitude or energy level with respect to a threshold, while the steady state condition requires valid data points during a given period of time. Consequently, an even higher reliability level may be achieved.

Pressure and inflation rate during pulse rate determination An exemplary application of the pulse rate measurement method in the context ofNIBP, more particularly inflation-based NIBP, will be discussed hereinafter. Generally, pulse rate determination may be performed at different cuff pressures and cuff pressure rates.

In one exemplary embodiment, the pulse rate measurement is performed at cuff oscillations at substantially high amplitudes. However, as the inflation rate during envelope detection should be related to the pulse rate, the pulse rate should be processed below a pressure where the envelope has to be detected, for instance below the diastolic pressure.

Further, it can be envisaged to perform pulse rate measurement at a constant pressure (i.e. no inflation) or at a defined inflation (or deflation) speed. The advantage of determining the pulse rate at a fixed pressure is that the subtraction of the inflation pressure ramp may be omitted. A potential issue related thereto may be that at the particular pressure, in some cases, oscillations might be of relatively low amplitude.

In one exemplary embodiment, pulse rate measurement is started when the flow into the cuff is basically stable. In periods of quick changes (e.g. due to sudden activation/enabling of the pump) some transient effects might introduce a big oscillation that may obscure or slow down the pulse rate measurement. Hence, pulse rate measurement should be started at least at a fixed amount of time after a sudden activation of the pump.

Given the above, regarding an appropriate time slot for the pulse rate measurement, a number of viable embodiments is presented hereinafter.

In an exemplary embodiment, the cuff pressure is quickly ramped up to a set point pressure that is high enough that pressure oscillations are present (detectable), but low enough to be below the diastolic pressure. For instance, this pressure can be a defined fraction (e.g. 0.9) of the diastolic pressure for 95% of the patient population (from statistics, e.g. 33 mmHg, leading to a threshold of 33 x 0.9 = 30 mmHg). Needless to say, for a different population, e.g. people suffering from cardiac diseases, elderly people, children or even neonates, this basically results in another value. Generally, the set point pressure may be derived from statistical data and/or empirical data. Pulse rate oscillations may be determined at this defined (basically fixed) pressure level. It is further noted in this context that in accordance with this embodiment ramp removal at the initial stage is not necessarily needed as the pressure during pulse rate determination is fixed. However, at least a DC signal portion may be removed. When pulse rate is known, pressure is increased at a pulse rate-dependent inflation speed.

In another exemplary embodiment, pressure is quickly ramped up to a set point pressure that exhibits basically high cuff oscillations, and may be above diastolic pressure. As discussed above, this pressure level might be related to some fraction of a representative systolic pressure for a defined group of patients. For example, this pressure value may be defined to be at 80 mmHg. Subsequently, pulse rate oscillations are determined at the fixed pressure level. When the pulse rate is found, pressure is deflated to a level below diastolic pressure (e.g. 30 mmHg). Subsequently, the cuff pressure is again increased at a defined pulse rate-dependent inflation speed. The envelope is constructed at the portion of pressurizing regime after (interim) deflation.

In yet another illustrative embodiment, pressure is quickly ramped up to a set point pressure that is high enough that pressure oscillations are present, but low enough to be below diastolic pressure. For instance, this pressure value may be defined to be at 30 mmHg. After reaching the pressure threshold, pressure may inflated slowly (e.g. 5 mrnHg/s) while pulse rate measurement is performed. Subsequently, one of the following pressurizing regimes may be applied:

In one embodiment, the cuff pressure is increased at a defined pulse rate- dependent inflation speed wherein both the envelope of the pulse rate measurement period and the envelope of the subsequent period are used.

In an alternative embodiment, the cuff is deflated to a level below diastolic pressure (e.g. 30 mmHg) after which pressure is again increased, but in accordance with at a pulse rate dependent inflation speed. Then the envelope may be constructed at the part after the deflation stage.

Pulse rate dependent inflation speed Once the pulse rate is determined, the inflation speed may be adapted or linked to the pulse rate. In an exemplary embodiment, the inflation speed is chosen proportional to the pulse rate (or: heart rate, HR): Inflation Speed [mmHg/s] = k [mmHg/beat] x HR [BPM] /60

As it is a general goal to have a certain number of beats (cuff oscillations) in a blood pressure measurement envelope, the proportional factor k may be adapted to reflect this. For instance, a number of about 10 beats in an envelope is a practical definition. Hence, the actual value of the factor k may be defined as follows:

In an exemplary embodiment which is applicable to a normal blood pressure level of about 120/80 (i.e. pulse pressure amplitude of 120-80 = 40), envelope between 40 and 130, and an inflation speed target rate of 9 mmHg/beat results in 10 beats in an envelope from which 4.5 beats are in the range between 80 and 120 mmHg. In this context, a relative envelope width (also referred to as REW) may be utilized. The relative envelope width may be regarded as a difference between the maximum and minimum value of the envelope that is required, divided by the pulse pressure. In the given example the relative envelope width is 2.25 (i.e. (130-40)/(120-80) = 2.25).

In another exemplary embodiment which is applicable to a hypotensive patient with a blood pressure of 75/45 (envelope between 30/90), an inflation speed target rate of 9 mmHg/beat results in 7 beats in an envelope from which 3 beats are in the range between 45 and 75 mmHg.

The above exemplary calculations illustrate that a factor of 9 mmHg/beat basically results in a sufficient number of beats in normal and hypotensive patients. Patients with a high pulse pressure (SBP-DBP, as is more likely in hypertensive patients) will have more beats. Therefore in a preferred embodiment, the factor k is 9 mmHg/beat. More generally, the factor k may be selected from a range between 2 and 20 [mmHg/beat], depending on an actual application.

In an alternative embodiment, the factor k may be calculated for different situations and/or applications based on the lowest expected pulse pressure (LEPP, in

[mmHg]), the desired relative envelope width (REW), and a minimum required number of beats per envelope (NBPE): k [mmHg/beat] = LEPP [mmHg] x REW /NBPE [beats] For illustrative purposes, exemplary values for LEPP = 30 mmHg, REW = 1.4, NBPE = 7 would result in a factor k = 6 mmHg/beat.

In exemplary practical applications, the pulse rate is measured and validated first (e.g. based on one or more criteria and a steady state condition assessment), and then the cuff may be inflated at constant speed. Further, also during inflation of the cuff, the pulse rate may be monitored and hence a more recent value of pulse rate may be determined and used to update the inflation speed (which would then be a variable speed). However, also a constant or quasi-constant inflation speed, based on the measured pulse rate, may result in a significantly increased NIBP performance.

Pulse rate stability check

For a successful and reliable blood pressure measurement, it is helpful to ensure that a relatively stable pulse rate during the course of the measurement is present. Generally, pulse rate instability may result from health physiological responses to e.g. stress or exercise, or from arrhythmias or other heart problems.

In order to ensure a successful and reliable measurement, the pulse rate determination may be continued during the inflation period. For instance, pulse rate determination may be performed in a repetitive fashion, based on a most recent period of the signal (e.g. 5.0 seconds, as described further above). In this way, a temporal progress of the subject's pulse rate over the course of the measurement may be obtained, and checked for stability. For instance, standard deviation values, maximum absolute difference values, and/or defined ranges may be calculated and assessed.

Also in this context, threshold values and/or ranged may be defined for an assessment of the measured values. If it is then determined that the pulse rate is not sufficiently stable, respective values may be flagged as potentially corrupted values, and presented together with the displayed measurement result. In case of device-internal use, the occurrence of corrupted features may be used to initiate a stop of and to restart the

measurement. For instance, a restarted measurement may be based on a lower inflation rate to account for the detected uncertainty.

Alternatively, the device may abort and postpone an actual measurement process for a defined period, e.g. for 5 minutes. Hence, the measurement can be re-initiated at a more suitable moment in time involving improved measurement conditions. Further strategies for responding to pulse rate instability may be envisaged within the general context of the present disclosure. While the invention has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive; the invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims.

In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. A single element or other unit may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

A computer program may be stored/distributed on a suitable medium, such as an optical storage medium or a solid-state medium supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems.

Any reference signs in the claims should not be construed as limiting the scope.

Claims

CLAIMS:

1. A pulse rate measurement module (40) for use in a vital signs monitoring system (10), the module (40) comprising:

an interface (42) operatively coupled with a pressure measurement unit (26) that detects a cuff pressure signal (30) of a wearable cuff (14) which is arranged to pressurize a measurement site of a subject,

a signal separator (46) arranged to extract a potentially pulse rate indicative signal portion from the cuff pressure signal (30),

an autocorrelator (48) arranged to autocorrelate the separated potentially pulse rate indicative signal portion, and

- a pulse rate processor (50) arranged to process the pulse rate based on temporal information linked with processed correlation coefficients.

2. The module (40) as claimed in claim 1, wherein the signal separator (46) is arranged to separate the potentially pulse rate indicative signal portion from a non- indicative signal portion describing an overall pressurizing regime.

3. The module (40) as claimed in claim 1 or 2, wherein the autocorrelator (48) is arranged to detect pulse rate indicative periodic patterns in the potentially pulse rate indicative signal portion.

4. The module (40) as claimed in any of the preceding claims, wherein the autocorrelator (48) is further arranged to repetitively apply a running autocorrelation based on a defined moving time interval, and wherein the defined moving time interval is preferably in the range 2.0 to 10 s, wherein a repetition frequency of the autocorrelation is in the range of about 0.5 Hz to about 200 Hz.

5. The module (40) as claimed in any of the preceding claims, wherein the pulse rate processor (50) is arranged to identify pulse rate indicative correlation coefficients, and wherein the pulse rate processor (50) is further arranged to detect and remove repetitive patterns in the correlation coefficients, and to assess corresponding correlation coefficients so as to detect the pulse rate.

6. The module (40) as claimed in any of the preceding claims, further comprising a pulse rate validity processor (52) arranged to assess the validity of a potential pulse rate indicative correlation coefficient, and wherein the pulse rate validity processor (52) is preferably arranged to calculate a signal stability indicative signal.

7. The module (40) as claimed in claim 6,

wherein validity assessment is based on at least one of the following:

determination of an absolute value of a candidate correlation coefficient, determination of a relative value of a candidate correlation coefficient, and determination of a an amount of time passed since the start of pulse rate estimation relative to the processed period,

and comprises the verification of whether the determined values comply with defined thresholds, and/or

wherein validity assessment is based on further processing the potentially pulse rate indicative signal portion, preferably of a defined period thereof, and to assess whether processed values comply with defined thresholds, wherein said further processing involves at least one of the following:

determination of a signal amplitude or energy of the processed period, determination of a signal variance of the processed period,

8. The module (40) as claimed in claim 6 or 7, the pulse rate validity processor is further arranged to process a signal stability indicative signal describing the stability of the measured pulse rate, wherein processing the signal stability indicative signal is based on at least one of the following:

- computation of autocorrelation values for a defined period,

computation of time-averaged autocorrelation values, based on absolute or relative autocorrelation values,

computation of low-pass filtered autocorrelation values, based on absolute or relative autocorrelation values, and a combination thereof,

wherein it is further assessed whether processed values, involving at least one of the above computed values, for a defined period of time, preferably a prolonged period of time.

9. A vital signs monitoring system (10) arranged for blood pressure measurement, the system (10) comprising:

a wearable cuff (14) adapted to pressurize a measurement site of a subject, a pressurizing unit (24) arranged to supply and operate the wearable cuff (14), - a pulse rate measurement module (40) as claimed in any of claims 1 to 8, the pulse rate measurement module (40) being operatively coupled with the wearable cuff (14), and

a blood pressure determination module (28) operatively coupled with the pulse rate measurement module (40), the blood pressure determination module (28) being arranged to calculate a blood pressure value based on the cuff pressure signal (30) detected by the pressure measurement unit (26) when the cuff pressure is increased or decreased by the pressurizing unit (24) in accordance with a defined pressurizing regime,

wherein the pressurizing unit (24) is arranged to apply a pressurizing regime dependent on the measured pulse rate.

10. The system (10) as claimed in claim 9, wherein a control unit (34) is provided which is operable to inflate the cuff (14) as a function of a measured pulse rate, and wherein the pulse rate is calculated prior to the blood pressure measurement based on the cuff pressure signal (30).

11. The system (10) as claimed in claim 9 or 10, wherein the blood pressure determination module (28) is operable to determine the pulse rate indicative signal at a level of the cuff pressure which is below a defined minimum pressure required for the calculation of the blood pressure value.

12. The system (10) as claimed in any of claims 9 to 11, wherein the pressurizing regime includes an inflation rate proportional to the measured pulse rate, and wherein the inflation rate is preferably determined in accordance with the following equation: Ri [mmHg/s] = k [mmHg/beat] x HR [BPM] /60 wherein Ri is the inflation rate,

wherein HR is the measured pulse rate,

wherein k is the proportionality factor, and

wherein the factor k is in a range of 2 to 20, preferably in the range of 5 to 15, more preferably in the range of 7 to 12.

13. A method for pulse rate measurement for use in a vital signs monitoring system (10), the method comprising:

detecting a cuff pressure signal (30) of a wearable cuff (14) which is arranged to pressurize a measurement site of a subject,

processing the pulse rate, including:

extracting a potentially pulse rate indicative signal portion from the cuff pressure signal (30),

processing the pulse rate based on temporal information linked with processed correlation coefficients.

14. A method for controlling the inflation rate of a wearable cuff (14) for an inflation-based non- invasive blood pressure measurement system (10), wherein the wearable cuff (14) is arranged to pressurize a measurement site of a subject, the method comprising the steps of:

- operating a pressurizing unit (24), particularly a pump or valve, to supply pressure to a wearable cuff (14) which is arranged to pressurize a measurement site of a subject, wherein the pressurizing unit is arranged to apply a pressurizing regime dependent on a control signal,

detecting the pulse rate according to the method of claim 13; and

- processing and controlling the control signal of the pressurizing unit (24) to set an inflation rate of the wearable cuff (14) as a function of a measured pulse rate.

15. Computer program comprising program code means for causing a computer to carry out the steps of the method as claimed in claim 13 or 14 when said computer program is carried out on a computer.