WO2013056319A1

WO2013056319A1 - A system and method for determining blood pressure

Info

Publication number: WO2013056319A1
Application number: PCT/AU2012/001287
Authority: WO
Inventors: Adel Ali S. AL-JUMAILY; Mohamed Odeh AL-JAAFREH
Original assignee: University Of Technology, Sydney
Priority date: 2011-10-21
Filing date: 2012-10-22
Publication date: 2013-04-25

Abstract

The present invention relates to a system and method for determining blood pressure. The system and method utilises a heart rate/blood pressure relationship which is established by analysis of data samples comprising heart rate values and blood pressure values from a plurality of subjects. The relationship is established using fuzzy logic processing. The relationship is used to determine blood pressure from a heart rate of patients.

Description

A SYSTEM AND METHOD FOR DETERMINING BLOOD PRESSURE

Field of the Invention The present invention relates to a system and method for determining blood pressure.

Background of the Invention

The measurement and detection of blood pressure (BP) is an important step in the examination and treatment of patients. Blood pressure is relevant to various

conditions such as hypertension, hypotension, artery stiffness, cardiovascular disease and other medical conditions. A number of methods are available to measure blood pressure. These include invasive methods such as cannulation, which requires a probe to be inserted into the artery of a patient. This method is accurate, but is invasive and not suitable for all situations.

Non-invasive methods are also available, such as the use of a pressure cuff. A problem with non-invasive methods is that accuracy is limited and depends very much on the skill of the medical practitioner taking the reading .

Summary of Invention

In accordance with a first aspect of the present invention, there is provided a method of determining blood pressure comprising the steps of:

obtaining a heart rate; and

determining a blood pressure result based on the heart rate and a heart rate/blood pressure relationship; wherein the heart rate/blood pressure relationship is established by analysis of data samples comprising heart rate values and blood pressure values from a plurality of subjects.

An advantage of at least an embodiment of the invention is that the blood pressure of a patient may be obtained accurately without requiring an invasive approach or a cuff.

In an embodiment, the step of analysis of the data samples comprises the step of fuzzy logic processing.

In an embodiment, the fuzzy logic processing comprises the step of obtaining fuzzifier variables from the obtained heart rate.

In an embodiment, the fuzzy logic processing further comprises the step of evaluation of a set of rules, wherein the rules are arranged to select a solution set of fuzzifier variables from the fuzzifier variables.

In an embodiment, the fuzzy logic processing further comprises the step of processing the solution set of fuzzifier variables to determine the blood pressure result.

In an embodiment, the step of processing the solution set of fuzzifier variables to determine the blood pressure results comprises a defuzzifying the solution set fuzzifier variables into the blood pressure result.

In an embodiment, the fuzzy logic processing uses an input fuzzy set which comprises a plurality of membership functions associated with a range of heart rate readings.

In an embodiment, the input fuzzy set further comprises a plurality of heart rate parameters.

In an embodiment, the plurality of heart rate parameters may include heart rates and their

classification within different bands, such as very low, low, medium, high, very high, the distribution of heart rates within these different bands and the overlapping areas of where a heart rate may be classified in more than one band.

In an embodiment, the fuzzy logic processing uses an output fuzzy set which comprises a plurality of membership functions associated with a range of blood pressure readings .

In an embodiment, the output fuzzy set further comprises a plurality of blood pressure parameters.

In an embodiment, the plurality of blood pressure parameters may include blood pressure readings and their classification within different bands, such as very low, low, medium, high, very high, the distribution of blood pressure within these different bands and the overlapping areas of where a blood pressure reading may be classified in more than one band.

In an embodiment, the range of heart rate readings and the plurality of heart rate parameters are optimized.

In an embodiment, the range of heart rate readings and the plurality of heart rate parameters are optimized by a first optimization process arranged to comprise the steps of: comparing the blood pressure result with a real blood pressure value determined from the data samples and, adjusting the range of heart rate readings and the plurality of heart rate parameters to reduce the

difference between the blood pressure result and the real blood pressure value.

In an embodiment, the range of blood pressure readings and the plurality of blood pressure parameters are optimized.

In an embodiment, the range of blood pressure values and the plurality of blood pressure parameters are optimized by a second optimization process arranged to comprise the steps of: comparing the blood pressure result with a real blood pressure value determined from the data samples and adjusting the range of blood pressure readings and the plurality of blood pressure parameters.

In an embodiment, the first and second optimization process comprises a multi particle swarm optimization process.

In an embodiment, the multi particle swarm

optimization process comprises a best vector.

In an embodiment, the best vector is associated with the heart rate/blood pressure relationship.

In an embodiment, the best vector comprises a function to determine the blood pressure result.

In an embodiment, the best vector is determined by comparing the blood pressure result with the real blood pressure value.

In an embodiment, the fuzzy logic processing is an interval type 2 fuzzy logic process.

In an embodiment, the plurality of membership functions is associated with a gaussian function.

In accordance with a second aspect of the present invention, there is provided a method of establishing a heart rate/blood pressure relationship comprising the step of analysis of data samples comprising heart rate values and blood pressure values from a plurality of subjects.

In an embodiment, the step of analysis of data samples include fuzzy logic processing.

In an embodiment, the fuzzy logic processing comprises a swarm analysis process.

In accordance with a third aspect of the present invention, there is provided an apparatus for determining a heart rate/blood pressure relationship comprising a processor arranged to analyze data samples from a

plurality of subjects, the analysis comprising:

a function arranged to correlate recorded heart rate values; and recorded blood pressure values for each of the plurality of subjects in the data samples to establish the heart rate/blood pressure relationship.

In an embodiment, the processor arranged to analyze data comprises a function for fuzzy logic processing.

In accordance with a fourth aspect of the present invention, there is provided an apparatus for measuring blood pressure comprising:

a processor arranged to determine a blood pressure result based on a patient' s heart rate and a heart rate/blood pressure relationship;

wherein the heart rate/blood pressure relationship is established by analysis of data samples from a

plurality of subjects, from a plurality of subjects, the analysis comprising:

a function arranged to correlate recorded heart rate values and recorded blood pressure values for each of the plurality of subjects in the data samples to establish the heart rate/blood pressure relationship.

In an embodiment, the processor arranged to analyze the data samples comprises a function for fuzzy logic processing .

In accordance with a fifth aspect of the present invention, there is provided an apparatus for determining cardiovascular parameters comprising at least one sensor arranged to detect at least one signal from a patient, wherein the at least one signal is used to determine each one of the heart rate, pulse wave velocity and oxygen saturation levels of the patient. In an embodiment, the at least one signal is a Photoplethysmograph signal.

In accordance with a sixth aspect of the present invention, there is provided a method for determining blood pressure comprising the steps of:

obtaining a heart rate from a Photoplethysmograph signal; and

In accordance with a seventh aspect of the present invention, an apparatus for determining blood pressure comprising :

a device arranged to obtain a heart rate; and a processor arranged to determine a blood pressure result based on the heart rate and a heart rate/blood pressure relationship;

wherein the heart rate/blood pressure relationship is established by analysis of data samples comprising heart rate values and blood pressure values from a plurality of subjects.

In an embodiment, the device arranged to obtain a heart rate is a Photoplethysmograph device.

In accordance with an eighth aspect of the present invention, there is provided an apparatus for determining blood pressure comprising:

a Photoplethysmograph device arranged to obtain at least one Photoplethysmograph signal, wherein the signal is processed to determine a heart rate; and a processor arranged to determine a blood pressure result based on the heart rate and a heart rate/blood pressure relationship;

wherein the processor and the Photoplethysmograph device is not calibrated in use.

Brief Description of the Drawings

Embodiments of the present invention will now be described, by way of example, with reference to the accompanying drawings in which:

Figure 1 is a block diagram of a system for

established a heart rate/blood pressure relationship in accordance with one embodiment of the invention;

Figure 2 is a block diagram of an embodiment of a system for establishing blood pressure by use of the heart rate/blood pressure relationship derived by the system of Figure 1;

Figure 3 is a flowchart illustrating an example process for determining the heart rate/blood pressure relationship established by the system of Figure 1;

Figure 4A is a chart illustrating the extension of a type 1 membership function to an interval type 2 primary membership function of a fuzzy logic system in accordance with one embodiment of the present invention;

Figure 4B is a chart illustrating the Gaussian interval type 2 primary membership function in accordance with the interval type 2 fuzzy logic system of Figure 4A;

Figure 4C is a block diagram of the type 2 fuzzy system of Figure 4A;

Figure 4D are charts illustrating the results of the fuzzy logic system in accordance with different stages of the operation of the interval type 2 fuzzy logic system of Figure 4A; Figure 5A is a chart illustrating the movement of particles during the operation of a particle swarm optimization process in accordance with one embodiment of the present invention;

Figure 5B is a flow chart illustrating the logical process of the particles swarm optimization process of Figure 5A;

Figure 6A is a chart illustrating an example primary membership function of an example input fuzzy set for the interval type 2 fuzzy system in accordance with one embodiment of the present invention; and

Figure 6B is a chart illustrating an example primary membership function of an example output fuzzy set for the interval type 2 fuzzy system of Figure 6A.

Detailed Description of an Embodiment of the Invention

A first aspect of an embodiment of the present invention relates to the analysis of a set of data samples to establish a heart rate/blood pressure relationship.

In an embodiment, the analysis of a set of data samples to establish a heart rate/blood pressure

relationship is carried out by a system for establishing a heart rate/blood pressure relationship. With reference to Figure 1, there is illustrated a system for establishing a heart rate/blood pressure relationship comprising a processor arranged to analyze data samples from a

plurality of subjects, the analysis comprising a function arranged to correlate recorded heart rate values and recorded blood pressure values for each of the plurality of subjects in the data samples to establish the heart rate/blood pressure relationship.

In this embodiment, the system is arranged to establish a HR/BP relationship such that the relationship is suitable for application to determine a blood pressure result of a patient by detection of the heart rate of the patient. Measurements and reviews undertaken by the inventor of recorded heart rate (HR) and blood pressure (BP) of individual patients over a large data sample of results have indicated that there is a relationship between the heart rate and blood pressure. However, the inventor has found that this relationship between the heart rate and blood pressure is a non-linear

relationship, in that the relationship is complex, multimode and is a multi-uncertainty relationship.

This relationship may be referred to as the heart rate/blood pressure relationship (or HR/BP relationship) . In the inventor's research and analysis, the inventor has devised embodiments of models, processing structures, instructions to establish this HR/BR relationship suitable for use in methods and apparatuses to provide a blood pressure result from a heart rate measured from a patient.

In this embodiment, the system comprises a computing apparatus 100 having suitable components necessary to receive, store and execute appropriate computer

instructions. The components may include a processing unit 102, read-only memory (ROM) 104, random access memory (RAM) 106, and input/output devices such as disk drives 108, input devices 110 such as an Ethernet port, a USB port, etc. Display 112 such as a liquid crystal display, a light emitting display or any other suitable display and communications links 114. The apparatus 100 includes instructions that may be included in ROM 104, RAM 106 or disk drives 108 and may be executed by the processing unit 102. There may be provided a plurality of communication links 114 which may variously connect to one or more computing or electronic devices such as servers, personal computers, terminals, wireless or handheld computing devices. At least one of a plurality of communications link may be connected to an external computing network through a telephone line or other type of communications link.

The apparatus 100 may include storage devices such as a disk drive 108 which may encompass solid state drives, hard disk drives, optical drives or magnetic tape drives. The apparatus 100 may use a single disk drive or multiple disk drives. The apparatus 100 may also have a suitable operating system 116 which resides on the disk drive or in the ROM of the apparatus 100.

In this embodiment, the apparatus 100 is arranged to receive a data sample 130 which includes a plurality of recorded heart rates and the corresponding recorded blood pressure readings of a plurality of subjects. Preferably, the data sample 130 includes a large number of samples of recorded heart rates and blood pressure readings for a large range of patients with different attributes and conditions. Examples of recorded hospital data of long termed monitored subjects in intensive care units (ICU) may be used as a suitable data sample 130. Other examples of data samples 130 may include recordings of heart rate and blood pressure values of patients being examined by a medical practitioner through suitable medical instruments.

Once the apparatus 100 receives the data sample 130, analysis of the data sample 130 is initiated. The apparatus may be programmed by software instructions to firstly randomly select a selection of data from the data sample depending on the size of the data sample 330. Once a selection is made, analysis of the correlation between the heart rate and the corresponding blood pressure is made, and established as a heart rate/blood pressure (HR/BP) relationship 133. This relationship may be established by statistical analysis, the application of mathematical models, computer logic problem solvers, fuzzy logic systems, swarm optimization processes or a

combination of anyone of these methods to devise a suitable HR/BP relationship 133 for use to determine a blood pressure result of a patient based on an obtained heart rate of the patient.

In one embodiment, the HR/BP relationship 133 may be in the form of mathematical statements, computer

instructions, codes or other form of indicia with a plurality of known variables, constants, estimated variables or any combination thereof. The HR/BP

relationship 133 may then be produced into a suitable format, such as computing instructions and transmitted for use by a device or apparatus to determine a blood pressure based on a measured heart rate. An embodiment of the methods and systems used to establish an example

embodiment of a heart rate/blood pressure relationship 133 is described herein this document with reference to

Figures 3 to 6B where it is illustrated the determination of a heart rate/blood pressure relationship 133 may use fuzzy logic system and/or particle swarm optimisation methods to process and evaluate the recorded heart rate and recorded blood pressure values within the data samples 130.

In a second aspect of an embodiment of the

invention, the established HR/BP relationship is utilised to determine a blood pressure result from an input patient heart rate. With reference to Figure 2, there is

illustrated an embodiment of an apparatus for measuring blood pressure comprising a processor arranged to

determine a blood pressure result based on a patient's heart rate and a heart rate/blood pressure relationship.

In this embodiment, the apparatus 200 comprises a processor 202 arranged to determine a blood pressure result based on a patient's heart rate 204 and a heart rate/blood pressure relationship 133, wherein the heart rate/blood pressure relationship is established by

analysing data samples comprising heart rate values and blood pressure values from a plurality of subjects.

In this embodiment, the processor 202 may be a computing processor or other form of logical unit capable of executing computing or machine instructions . The processor 202 may include a computing apparatus, computer, arithmetic logical devices or other processing means arranged to accept an input of a measured heart rate value 204. In one example, the heart rate value may be measured in real time 204A or entered manually from pre-existing database or readings 204B collected by a user.

Preferably, the heart rate is measured by a

Photoplethysmograph (PPG) device 206 arranged to directly measure various cardiovascular parameters of a patient. These parameters may include the heart rate (HR) , the pulse wave velocity (PWV) or the oxygen saturation level of a patient.

As shown in Figure 2, in one embodiment, the PPG sensor consists of an optical source and an optical sensor and is arranged to be engaged or clipped on to a subject's finger, limb or earlobe to measure the cardiovascular parameters. In one example, the optical source 208 is an infrared light emitting diode (LED) that acts as an optical transmitter and the optical sensor is a photo-diode 210 that acts as an optical receiver. From the sensor, a PPG signal 212 may be detected based on the amount of light, timing of the light or relevant attributes of the light received by the photo diode 210.

The PPG signal 212 comprises a number of physical signal components. These physical signal components include a DC component, AC component, the values of the DC and AC component and the frequencies of the AC component. Once a PPG device 206 is attached to a person, either through a limb, earlobe or other part of their body, the components of the PPG signals 212 are affected by the absorption of these components by the anatomy of the patient. Examples of such absorption include:

1. Absorption due to pulsated arterial blood of the patient;

2. Absorption due to non-pulsated arterial blood of the patient;

3. Absorption due to venous blood of the patient;

4. Absorption due to skin, bone, and tissues of the patient.

By detecting and measuring the changes in the components of the signal, these rates of absorption can be measured. As a result, the PPG signal 212 may be used to measure and/or compute the values of cardiovascular parameters including heart rate, pulse wave velocity, oxygen saturation level, and as explained with reference to Figure 3, blood pressure parameters.

In one example, the PPG signal 212 is used to measure the Heart rate (HR) by evaluating the AC component frequencies of the PPG signal 212. Once the frequency is measured, the heart rate can be derived by this equation:

HR = 60 x frequency of PPG signal.

In another example, the Pulse Wave Velocity (PWV) may also be measured by the use of PPG signals 212 received from the PPG device. The Pulse wave velocity (PWV) is a parameter which measures the velocity of the heart beat based on the standard expression for velocity calculation where the average velocity v of an object moving through a displacement (AD) during a time interval (AT) is described by the formula: AD/AT

In one embodiment, two PPG sensors are attached to the patient's finger, although other parts of the body are possible. The two PPG sensors are spaced with a known distance (AD) . Once the PPG sensors are in position, two output signals may be detected by the two PPG sensors .

This may be referred to as outputi and output2. By measuring the time difference between the two signals, the PWV can be measured. This relationship can be expressed by the following equation:

AD Distance between two PPG sensors

PWV =

|Time of beginning of output ₂ - Time of beginning of output _j |

In another embodiment, the PPG signals 212 may be utilized to determine the cardiovascular parameter of Oxygen Saturation level (S_p0₂ ) . From a clinical point of view, the Oxygen Saturation level (OSL) (also known as S_p0₂ or Sa O2 ₎ represents the amount of oxygen saturation in the vessel ' s blood .

In one example, to measure the OSL, the PPG signals

212 may be used to the Oxy-hemoglobin ( Hb0₂) optical absorption ratio and De-Oxy-hemoglobin (Hb) optical absorption ratio, where S_?0₂ is calculated by equation (Al) :

HbO^ where, S_p0₂ = Oxygen Saturation,

Hb0₂ = Oxy-hemoglobin,

Hb = De-Oxy-hemoglobin

The oxygen saturation, can also be defined in terms concentrations ( c) as: ^HbO

Sp02^: ^xl00% (A2)

c HbO +c Hb

2

from equation (Al) and after ignoring the percentage part :

= ^p°₂ l^cHbo, ^{+ c}m) (A3)

^CM> - (1 (S_p0₂)) (cHb0₂ ^+CHb (A4)

The optical absorption ratio Hb0₂ and Hb depends on medium concentration of Oxy-hemoglobin and

De-Oxy-hemoglobin and may be acquired by the PPG sensor (s) based on the Beer-Lambert Law. The Beer-Lambert Law states that if the light travels through a uniform medium, then the intensity of the light will attenuated by a predictable rate based on the extinction coefficient of the medium absorption (ε (λ) ) , concentration of the medium (c), and the length of optical path (d) . Therefore, the intensity of the light after traveling (I_out) is related to the original intensity of the light (Ii_n) as shown in equation (A5) :

j ₌ j (8( )xcx</)

^Aout ^Ain^c (A5)

Hence, when the optical light from the PPG sensors travels through the subject's finger, then the light intensity after travelling, can be found based on Beer-Lambert law by using formula (A6) :

*out ^~ in^e (A6)

After substituting the values of C^_Q^ and<¾_b in terms of S form equations (A5) and (A6) , respectively, then simplifying the whole expression, the following equation (A7) is derived: _=> ln^ = </x(c_Hb02 +^)^χ[_{ε Βθ2} (λ)^χ^_ρ0₂ + _είΚ(λ)(1- ^_ρ0₂))] _{(A7 )} in

In one example, to find the S_PC>2 of equation (A7); the ratio between light intensities of two PPG sensors with different wavelengths (Infrared and red) are required. The ratio (R) between light intensities of Red and IR diodes are expressed by:

After simplifying equation (A5) by substituting for both PPG sensors and rearranging the whole expression, the S_p02 can be valued by

The Ratio (R) in equation (A9) is calculated by DC and AC components of two different PG signals:

AC_R/DC_R

R= — (A10)

AC_IR/DC_IR

In this example, values of the extinction coefficients (s_Hbo2 and ε_Ηβ) for both wavelengths can be obtained from a spectrum graph based on the wavelength of the light used in the optical sensors. An example of an absorption spectrum graph can be found in J.G. Kim, X. Mengna and L. Hanli, "Extinction coefficients of hemoglobin for near-infrared spectroscopy of tissue", Engineering in Medicine and Biology Magazine, IEEE, Vol.: 24, No.: (2), pp.: 118-121, 2005. Once R and the extinction coefficients are found, the Oxygen Saturation Level can be calculated by equation A9.

In an example embodiment, the processor 202 is also arranged to receive a heart rate/blood pressure relationship 133 which may be in the form of a computer data set, computer instructions, machine code or mathematical instructions. The heart rate/blood pressure relationship 133 may be stored on storage media such as memory, hard disk, tape, non-volatile storage for execution by the processor 202 such that the measured heart rate from the PPG 206 or other heart rate input channels 204C can be analysed to determine a blood pressure result. The blood pressure result may form, in one embodiment, a suitable estimated blood pressure reading or the actual blood pressure of a patient with the inputted heart rate at the time of measurement of the heart rate .

In some embodiments, the apparatus 200 may be implemented into an individual standalone device incorporating a PPG sensor 206, a processor 202 and any data storage arranged the heart rate/blood pressure relationship 133 and the necessary instructions to execute the analysis of the heart rate with the HR/BP relationship 133. Alternative embodiments of the device may include communication connectivity by USB, Ethernet, WiFi or other forms of communication which allow the heart rate/blood pressure relationship 133 to be updated regularly. The communication connectivity may also be used to transmit cardiovascular parameters and blood pressure detected and determined by the device to be stored, recorded or analysed at a later time.

In an example use of the device, the PPG sensor 206 may be attached to a finger, limb or earlobe of a patient to detect a heart rate for the patient. Once the heart rate is detected, the heart rate is inputted into the processor 202 for processing with the heart rate/blood pressure relationship 133 to determine a blood pressure result. Once the blood pressure is detected, the blood pressure may be displayed, stored, further processed or utilised to diagnose the patient, monitor the patient or alert a medical practitioner through an alert system.

In an embodiment of the present invention, the analysis of data samples to establish a HR/BP relationship comprises two unique steps. The first of which includes the design and implementation of a fuzzy logic system arranged to conduct fuzzy logic processing on a plurality of parameters to model the HR/BP relationship based on data samples comprising of heart rate and blood pressure values to establish a blood pressure result. The second step comprises training and optimizing the parameters used in the fuzzy logic system by swarm analysis such that the fuzzy logic system can model accurately the true relationship between the heart rate and the blood pressure, and thereby provide a trained and optimized fuzzy logic system which can be used as an accurate HR/BP relationship for deployment in systems and apparatus which utilises the HR/BP relationship to measure the blood pressure of a patient.

In an alternative embodiment of the invention, the apparatus 200 may also be able to display, store and process information relating to the cardiovascular parameters which can be detected by the PPG device 206. The heart rate, PWV and OSL which can be measured and calculated based on the PPG signals 212 can be displayed, used, transmitted or stored to aid in the diagnosis of a patient .

An advantage of at least one embodiment of the invention is that the apparatus 200 does not require any form of calibration to determine the blood pressure or any of the cardiovascular parameters of a patient. Once the heart rate/blood pressure relationship is established and provided to the apparatus, the heart rate is the only parameter required to determine a blood pressure measurement for the patient. As such, the apparatus 200 may be used in various medical situations, such as emergency wards, where there is an insufficient time to calibrate a blood pressure measuring device. In addition, as the apparatus 200, in one embodiment, uses a PPG to record the necessary parameters, periodic equipment adjustments or servicing is generally not necessary to maintain the accuracy of the apparatus 200 in providing blood pressure measurements or other cardiovascular parameters .

With reference to Figure 3, there is illustrated a flow chart showing an embodiment of a method of determining a heart/rate blood pressure relationship comprising the step of analysing data samples from a plurality of subjects, the analyst including the step of correlating recorded heart rate values and recorded blood pressure values for each of the plurality of subjects in the data samples to provide the heart rate/blood pressure relationship, wherein the step of analysis of the data samples comprise the step of fuzzy logic processing.

By analysing these data samples 240 using fuzzy logic systems and further optimizing the fuzzy logic system with swarm analysis conducted by a particle swarm optimization process, the non-linear relationship between the heart rate and the blood pressure may be at least estimated or accurately modelled for application to a detected heart rate to determine a blood pressure result of a patient.

By way of introducing the embodiments of the invention, examples of the general operation of interval type 2 fuzzy logic systems and particle swarm optimization processes for swarm analysis are described with reference to Figures 4A to 5B . Embodiments of the operational parameters and settings of interval type 2 fuzzy logic systems and particle swarm optimization process for used in one embodiment of the present invention to determine one example of the HR/BP relationship are described with reference to Figures 3, 6A and 6B.

Fuzzy logic systems, or fuzzy systems are systems which are designed to accurate and logically estimate an output value based on an input value and an logical application of problem set which models a multi-variable problem set. Fuzzy logic systems may include a number of steps to devise an accurate output value. These include the steps of fuzzification, rule evaluation and rule firing and defuzzification .

Each of these steps include various requirements to prepare an input fuzzy set, output fuzzy set and a plurality of rules which model the behavior and relationship of the input and the output. In each of these process, the designer of a suitable fuzzy system is required to determine a suitable input and output fuzzy set which comprises, amongst other items, the distribution of suitable primary and secondary membership functions (PMF, SMF) which outline the necessary conditions and parameters to model the problem between an input value and an output value in which the fuzzy logic system attempts to solve in order to accurate estimate the input value and the output value. In general, there are two kinds of fuzzy systems. These include, type 1 fuzzy system (Tl Fuzzy System or Tl FS) which are the first generation of fuzzy systems and Interval Type-2 fuzzy logic system, or Type-2 Fuzzy logic system (T2 fuzzy systems) which includes more sophisticated fuzzy sets for input and output to the fuzzy logic system.

In theory T2 Fuzzy sets (T2 FS) are represented by the same distributions mentioned for Tl fuzzy sets. The T2 Fuzzy set is characterized by a (x) as Primary

Membership Function (PMF) for set A and the membership value for each element of A is fuzzy set; which is called a secondary membership function or membership grade, an example of the Type-2 system, Gaussian type-2 set is formed by moving the Gaussian type-1 set to left and right in a non-uniform manner as shown in Figure 4A.

As a practical case of the general T2 FS, the

Interval T2 FS (IT2 FS) is proposed to avoid the poor performance of the non-uniform sides of T2 FS PMF, by bounding the sides of T2 MF by Upper MF and Lower MF to form the Footprint of Uncertainty (FOU) of the IT2 PMF.

Moreover, the Gaussian IT2 PMF depends on these main parameters :

1. Lower mean rti! .

2. Upper mean m_u .

3. Lower standard deviation o± .

4. Upper standard deviation o_u .

These parameters control the shape of primary membership functions and FOU of IT2 FS and are shown in Figure 4B.

With reference to Figure 4C, an example of a T2 FS model is shown. In this example, the T2 FS model comprises 4 parts. These include:

1. Fuzzifier, which represents the input into T2 fuzzy sets by Primary Membership Function

(PMFs) and initializes input PMFs' parameters.

2. Rules, which include rules in the format of "IF

THEN ..." . These are rules which are used to connect fuzzified sets (A_x) of input with represented sets of output (0_X) .

3. An inference engine which combines the rules and forms a mapping between the input sets and the output sets .

4. A Defuzzification process which is used to produce crisp (discrete and precise) values of output. The defuzzification process may include two sub stages. The first stage is descending the type-2 fuzzy set to type-1 fuzzy set by type reduction methods; such as centroid, centre-of-sums and height. The second stage is using a Defuzzifier to find the output crisp value.

An IT2 FS may be designed to predict one output (U based on two inputs (X and Y) . The range of X variable is fuzzified to a few fuzzy sets according to linguistic expressions based on specialist opinions. Then these sets are represented as Primary Membership Functions (PMFs) μ(χ) with triangle distribution dependent on different means and same σ; such as A and sets, while the range of Y variable is fuzzified to a few fuzzy sets according to linguistic expressions. These sets are then represented as PMFs of μ (y) with triangle distribution dependent on different means and same o; such as 1 and 0 sets. "IF THEN ..." rules of this IT2 FS are set between inputs and output .

To increase the accuracy of the IT2 FS, the IT2 FS may be tested by two instant numbers x and y. The instant number x belongs to X variable range while instant numbers y belongs to Y variable range are supplied to IT2 FS. The testing of the IT2 FS may include the following steps:

Firstly, in accordance with the Fuzzification process, the instant number x has two membership grades μΑ( ^χ) and μκ ( ^χ) while the instant number y has two membership grades μι (y) and μο (y) ;

Secondly, the rules of the IT2 FS are fired; the μΑ( ^χ) combines with μ_∑ ^~(y) to form ϋι set and the μ¾ ( ^χ) combines with μο (y) to form O₂ ;

Thirdly, the Inference engine combines the nonzero membership grades of x and y; where the minimum combination of μ¾ ( χ) and μ_∑ iy) are formed 0_: set and the minimum combination of μϋ ( ^χ) and μο (y) are formed O₂ set, then the maximum combination of Ui set and IJ2 set are formed 0_C set, and;

Finally, the formed 0_C set is defuzzified by a centre of gravity method to find two crisp values C_- and C2 and then by a centre of sums method to find one crisp value C. These four steps are illustrated graphically as shown in Figure 4D.

Once an IT2FS model is designed, the model may be trained by an available database of input variables and output variables to minimize the error between the estimated output variables by Fuzzy system and the real output variables by adjusting the parameters of the PMF.

Once the IT2 FS model is trained, further known results between the input and the output may be used to test the accuracy of the IT2 FS model. If the results indicated a sufficient level of accuracy, then the IT2 FS model is deemed ready for duty and can be deployed into a problem situation to devise a suitable solution. Where the IT2 FS model fails to achieve a suitable level of accuracy, the model may be retrained until a suitable level of accuracy is devised, or redesigned by adjustment and variation of the parameters within the PMF, or the entire PMF itself.

With reference to Figures 5A and 5B, an introduction to the particle swarm optimization process, based on particle swarm optimization theory is herein described. Adjustments and modification of suitable parameters and solution particles may then be used in a modified particle swarm optimization process to determine an example embodiment of the HR/BR relationship.

The Particle Swarm Optimization (PSO) process is an intelligent technique that was inspired by the social actions of bees flocking, food seeking by birds, and fish crowding together. These actions are applied to the many insights of persons' actions and cognition.

Generally, the PSO process is linked to bird flocking, fish schooling, and swarming theory. Also PSO process is related to other computation and programming evolutionary techniques and has relations with evolution strategies.

In one aspect, PSO contributes many attributes with Genetic Algorithms (GA) because it is initialized by a population with random solutions and searches the optimal solution (s) . However, PSO is unlike GA because it does not have crossover operators that GA has.

Moreover, in some examples, the potential solutions of PSO are called particles. Each particle flies, moves, and searches through the problem space by tracking its best solution and global best solution that is achieved by a whole swarm population.

The PSO tracking process depends on the updated velocity and updated position of each particle, every iteration, toward the best solution through problem domain. The velocity is weighted, with separate numbers being generated toward two best locations based on evaluation functions.

The two best values are tracked by particle swarm are the particle best and the global best, the particle best (pbest) is the best location that the particle has achieved so far in the problem space. Global best (gbest) value is the best value obtained so far by any particle of the whole population.

Particle Swarm Optimization parts are population of particles, interconnection topologies, search algorithms, and evaluation functions. These fundamentals cooperate together to find the optimum solution of problem.

The normal population of PSO is twenty to fifty particles. This number is determined according to the problem size. PSO particles have interconnection topologies describing the communication among them to move within the problem domain to find an optimum solution. There are two topologies :

1. Sociometry topology or global best topology (gbest) where every particle is connected with all particles of population and influenced by the particle which has found best problem solution;

2. Local best topology (pbest) where every particle is affected by its previous position value .

Local best topology can converge separately on various optima solutions in problem space and it has less complexity than global best topology. Global best topology, however, is faster in finding an optimum solution to the problem.

The main fundamental of PSO is the search algorithms which adjust particles' velocities and positions by equations (i) and (ii) respectively to move toward pbests and gbest within domain boundaries as shown in Figure 5A. The equations are as follows:

V_n (t+1) = χ ^χ V_n (t) + cpi x randi ^χ (pbest_n - X_n) + φ₂ ^x rand2 ^x (gbest - X_r_) (i) ;

X_n (t+1) = X_n (t) + V_n (t+1) (ii) V_n is the displacement of particle's movement;

n is the current particle;

t is the iteration number;

χ is the constriction coefficient;

φι,2 are the acceleration constants;

randi,2 are random numbers in range [0,1];

pbest_n is the particle's best position;

X_n is particle's position within problem domain;

gbest is the global best within all particles.

Firstly, parameters' values of n, V_n=1 (t=l) , X_n=i (t=l), ψι,2, X, pbest and gbest are initialized to start particles' movements within problem space. Then particles' displacements are updated by equation (i) then the particles' positions are updated by equation (ii) .

Acceleration constants' values c i,2 manage particles' movements and the probability of finding optimum solution slowly with high accuracy or quickly with less accuracy; increasing ( i supports exploration and leads particles' movements towards pbest, while increasing cp₂ supports exploitation and particles' movements towards gbest. On the other hand, the construction coefficient χ balances between the effect of previous displacement, which is referred to cognitive rates, and the effect of interconnection topologies local and global best on current displacement, which is referred to social rates. Also, randi,2 are used to stochastically change the relative pull of pbest and gbest and to imitate the sight unpredictable element of nature swarm behavior. Besides these algorithms and their parameters, the particles' movements are surrounded by problem space - Maximum displacement (V_max) and boundary wall that control the particles' movement within problem space. There are three types of boundary wall: Absorbing Wall: when a particle steps out the boundary of the solution space, the velocity value is changed to zero and the particle will eventually be pulled back toward the allowed solution space. In this sense the boundary "walls" absorb the energy of particles trying to escape the solution space;

Reflecting Wall: when a particle steps out the boundary, the sign of the velocity is changed and the particle is reflected back toward the solution space;

Invisible Wall: It is a wall that allows particles to move without any physical restriction. However, particles that wander outside solution space are not evaluated for fitness.

Finally, the evaluation functions appraise the obtained solution success by adjusting global and local best values according to new best particles' positions. This operation is reiterated till the particles or some of them reached the optimum solution. The optimum solution goal is achieved by three ways:

1. Maximum Iteration: The PSO will stop when number of iterations reaches the number defined by the user;

2. Target fitness termination: The PSO is executed by a number of iterations that are defined by the user, PSO will stop when the target fitness reaches the target defined by the user, or when the number of iterations reaches the number defined by the user;

3. Minimum standard deviation (STD) : The PSO is executed by number of iterations that defined by the user, PSO will stop when the computed STD of all particles' fitness are less or equal to the user-defined min-STD or when the number of iterations reaches the number defined by the user .

In summary, PSO search algorithms and interconnection topologies control the particles' movements within the problem domain to find the optimum problem solution that is evaluated by evaluation rules . Stages of PSO algorithm are simply expressed on the flowchart as shown in Figure 2B.

In an embodiment of the invention as illustrated in Figure 3, to process this non-linear and complex relationship between the blood pressure values and the heart rate, an Interval Type-2 Fuzzy system (IT2FS) is designed to determine blood pressure values by processing the recorded heart rate values. Once the IT2FS is designed, the IT2FS may then be used as a heart rate/blood pressure relationship 133 for the apparatus 200 to determine blood pressure values based on the measured heart rate .

In one example, two different (IT2FSs) are each designed to determine different components of the blood pressure values. The components of blood pressure values include Systolic blood pressure (SBP) , Diastolic blood pressure (DBP) and Mean Blood Pressures (MAP) .

In this embodiment, a first IT2FS is designed to determine Systolic Blood Pressure (SBP) and a second IT2FS is designed to determine the Diastolic Blood Pressure (DBP) based on heart rate (HR) values. Accordingly, the output for first IT2FS is SBP and the output for second IT2FS is DBP.

Preferably, both IT2FSs have the same structure and the same input but have a different output (SBP and DBP) , both systems may be designed by the following steps:

1. Devising an input fuzzy set by dividing the possible range of HR values. The HR values are fuzzified to five singletons Gaussian Interval Type-2 Primary Membership Functions (PMFs) with the same standard deviation for all PMFs. The PMFS may take linguistic expressions being; very Low, Low, Healthy, High and very High as shown in Figure 6A. In one embodiment the parameters for the PMFs' are initialized in accordance with data and statistics from the American Health Association (AHA) standards . (These are available at AHA, "American heart association", http://www.americanheart.org) . Devising an output fuzzy set by dividing the possible range of SBP and DBP (BPP) values into five Gaussian Interval type-2 Primary Membership Functions (PMFs) . Each of the PMFs have two same standard deviation (ol, o2) for each PMF. These PMFs take the linguistic expressions: very Low, Low, Healthy, High and very High as shown in Figure 6B. In one embodiment the parameters for the PMFs' are initialized in accordance with data and statistics from the American Health Association (AHA) standards. (These are available at AHA, "American heart association", htt : //ww . americanheart . org) .

Fuzzy rules are located as one rule for each input fuzzy set, where each rule has the expression: "IF HR is A, THEN BPP is G". In one example, these rules have the effect of being associated with the fuzzified variables produced from the input PMF.

The Inference engine combines the rules and forms a mapping between HR fuzzy sets and each blood pressure values (SBP or DBP) fuzzy sets. The inference engine then finds the non-zero membership function values of HR to generate the T2 fuzzy set of the blood pressure values by depending at least in one example, on the non-zero membership function values of HR and a specified inference method which is the product sum inference method. In one example, the processing of the inference engine "fires" the fuzzy rules in which the conditions are met. From the "firing" of the fuzzy rules, the associated fuzzified variables are selected as a solution set of fuzzified variables for defuzzification.

The Defuzzification process of the IT2FS is implemented by following two steps:

(a) Descend the T2 fuzzy set of blood pressure values (SBP or DBP) by computing upper MF (y_r) by using the equation:

¹ ∑/

where y_r upper membership function (MF) of output

i is variable value changed from 1 to I

N is number of non-zero membership function values for input value created by firing the fuzzy rules

f ¹ is fuzzy membership value (s) of input value

y^∑r is mean value of matching output upper

MF based on fuzzy rules

and lower MF yi by next equation:

i is lower MF of output y¹ ₁ is mean value of matching output lower

MF based on fuzzy rules .

(b) Compute the output of IT2FS by

Centre of Sums as clefuzzification method as expressed in next equation:

y is output of Fuzzy system.

Then the output of IT2FS is y = SBP or DBP depending on which whether this was the first IT2FS or second IT2FS.

Subsequently, the MAP parameter is computed based on the following equation:

MAP = DBP + - (SBP - DBP)

3

In this embodiment, the two designed IT2FS are implemented on a data sample of twenty unhealthy subjects. In this regard, the results of the two IT2FS would estimate the blood pressure values (SBP or DBP) for generally unhealthy patients.

In another embodiment, although the IT2FS are able to estimate blood pressure values, preferably, the IT2FS are trained or optimized to improve on the accuracy of the estimated blood pressure values. To improve train and optimize the IT2FS, the estimates blood pressure values are provided by each of the IT2FS are compared with the actual blood pressure values recorded of the patient correlating with the recorded heart rate. This may be conducted in one example by retrieve a data sample a database of recorded heart rate and recorded blood pressure values. Once the data samples are randomly chosen, the estimated blood pressure values from the two IT2FS and the actual blood pressure values of a database for the same heart rate (HR) are compared to establish an error rate. Once this rate is determined, the parameters relating to the input and output fuzzy sets of each of the IT2FS may be optimized to minimize the difference between the error rate. The optimization process will thereby provide a trained and optimized IT2FS which may produce significantly more accurate result.

In one example, suitable data samples may be retrieved from the MIMIC database. This database is available in the Physio-bank or PhysioNet website at (http : // ww.physionet . orcj/physiobank/ ) . The website includes continuous records of cardiovascular signals and interrupted measurements; such as HR and MAP values, of subjects in an intensive care unit (ICU) . These records are extracted from bedside monitors that were attached to subjects and from subjects' medical records.

A sample of the results are as follows. The mean of HR, SBP and DBP values of each subject are given in the following table:

Table 1

(Table 1 - Sample SBP, DBP and HR values) These actual blood pressure values and heart rates may then be used to optimize each of the IT2FS. In one embodiment, the optimization process is a particle swarm optimization process arranged to optimize the parameters relating to the input and output fuzzy sets of each of the IT2FS .

In one example, the SBP values are estimated by utilizing IT2FS and 20 unhealthy subjects' readings which are divided into two groups: 10 subjects' measurement of HR and SBP are used for training the designed IT2FS by using a particle swarm optimization process while the remaining ten subjects' readings are used for testing the trained IT2FS to show error rate of the results.

With reference to Figures 6A and 6B, an example embodiment of the Primary Membership Function for the input and output fuzzy set of the first ITFS2 (SBP) includes parameters of: ml, m2, m3, m4 , m5, mil, ml2, ml3, ml4, ml5, mul, mu2, mu3, mu4, mu5, a, Oi and o₂. Initially, these parameters are assumed as the means of the parameter ranges obtained from the AHA database. An optimization process, which in this example is based on a particle swarm optimization process is utilized to find the optimum parameters' values within the AHA parameters' ranges by adjusting parameters' values within the parameters' ranges to achieve the smallest average of absolute differences between real and estimated values of the SBP. The particle swarm optimization process may be utilized to optimize the PMF parameters' values of IT2FS by the execution of a number of optimization and comparison steps. These steps include, without limitations :

1. Set the problem domain for the particle swarm optimization process as the database of 10 cases and boundaries of each range for each parameter. This may be initially set as the parameters of the AHA database;

Assign different swarm population for each PMF parameter of input; ml, m2, m3, m4, m5 and o; Assign different swarm population for each PMF parameter of output; mil, ml2, ml3, ml4, ml5, mul, mu2, mu3, mu4, mu5, Oi and o₂;

Set the population for each swarm of steps 2 and 3 within a suitable range; In this system, fifty particles are proposed to consider whole problem domain; n = 1 ... 50, although alternative number of particulars are possible; Initialize fifty random values for each parameter within the boundaries of each range; for example, initialize fifty random values located within the boundaries of High range. Complete the same step for all other parameters ;

Initialize same fifty random velocities for all parameters within range of [0, 1] with 0.02 step change and set constant value for Vma_x = 0 or 1;

Set constant values for parameters; χ = 1 and cpi = φ2 = 2, for use in the following equation in each Swarm;

V_n (t+1) = x x V_n (t) + cpi x randi ^χ (pbest_n-X_n) + φ2 ^x rand2 ^x (gbest - X_n) ;

Utilize two particles' topologies gbest and pbest to achieve the optimal solution or near optimal solution for all parameters, where the initial pbest vectors are equal to the combination of initial fifty particles' random values. While the initial gbest vector is equal to the initial pbest vector, that achieves the smallest mean of absolute differences between estimated SBP (ESBP) values by designed IT2FS and Real SBP values for each of the 10 cases;

Start the searching part of particle swarm optimization process. Set the number of iterations to be 100. (e.g. t = 1 ... 100);

Evaluate the vector of gbest by the designed IT2FS. If the mean of absolute differences between ESBP and Real SBP values for training ten cases' records is zero, where real HR and Real SBP values equal the mean of real HR and Real SBP records for each case's of the ten cases' records and/or if the initial gbest vector passes the evaluation check, then the optimal value is achieved and the optimization process is stopped. In all other cases, the optimization process continues;

Update the particle's velocity for each swarm separately by updating the V_n value as the particle's velocity in the equation of Step 7: V_n (t+1) = x x V_n (t) + cpi randi ^χ (pbest_n-X_n) + cp2 rand2 ^x (gbest - X_r_) -(i) so that the new V_n value can be used in the next iteration;

Absorbing wall may be used as a boundary wall to keep the particle within the problem domain; Update the particle's position for each swarm separately by updating the X_n value as the particle's position in this equate:

X_n (t+1) = X_n (t) + V_n (t+1) - (ii) so that the new X_n can be used in the next iteration. Absorbing wall may be used as a boundary wall to keep the particle within the problem domain; Repeat steps 11 and 12 with other particles to update the V_n(t+1), X_n(t+1) values for other 49 particles to use them as the new values for V_n(t), X_n(t) in equations (4-2) and (4-3) during next iteration;

Update the pbest_n vector; which will be used as the new pbest_n vector in equation (ii) during next iteration, based on using the updated particles' values for each parameter; which are achieved from equation (iii) , to estimate SBP values by designed IT2FS for ten cases' records. If the average of absolute differences between ESBP and Real SBP values for ten cases with the combination of new particles' positions is less than the average of absolute differences between ESBP and Real SBP values for ten cases with the combination of previous particles' positions, then the combination of new particles' positions is assigned as new value for pbest_n vector. Otherwise new pbest_n vector equals the previous pbestn vector;

Update the gbest vector which equals the pbest_n vector that achieves the smallest mean of absolute differences between ESBP by IT2FS and Real SBP values; where each value of that vector will be used as a new gbest value for corresponding swarm respectively in equation (ii) during next iteration;

Test the new gbest vector by the evaluation check as in step 10;

Repeat steps 10, 11, 12, 13, 14 and 15 to update the V_n(t+1), X_n(t+1) for each particle and for all swarms also to update the pbest_n vectors and the gbest vector for whole system during every iteration till pass the evaluation check and the achieved optimal values or reach the maximum iteration (t = 100) ;

21. Finally assigns each value of gbest vector to corresponding PMF parameter, respectively. These values of PMF parameters form the trained IT2FS .

Once the training process is completed, the trained IT2FS which has been optimized by the optimization process is ready to be used to estimate SBP values. The trained IT2FS may then be codified into a machine code, computer software or procedure steps as an example of the heart rate/blood pressure relationship 133, which may be utilized by the apparatus 200 to determine a blood pressure result from a measured heart rate.

In one embodiment, the accuracy of this example of the heart rate/blood pressure relationship 133 derived from the trained IT2FS was tested by comparing the determined SBP and DBP values obtained from an application of the HR/BP relationship 133 for any given heart rate. The remaining 10 subject's records are used to test the HR/BP relationship 133. The difference between the results from the HR/BP relationship 133 is then compared with the results of the actual recorded data from the database .

These results are shown in the following table

Table 2

(Table 2 - Sample of SBP - Results of trained IT2FS)

The accuracy of this embodiment of the HR/BP relationship 133 may be calculated by equation:

Accuracy — 100%^x (1 ^— average of absolute difference between Estimated and Real values ^ average of Real values

In this embodiment, the HR/BP relationship 333

5.98

accuracy 100 (1- 109 ) 94.51

In an embodiment, the DBP values are estimated by the second IT2FS (DBP) by using the same 20 unhealthy subjects' measurements which are divided to two groups: 10 subjects' measurements of HR and DBP are used for training the IT2FS by using the optimization process whilst the remaining ten subjects' measurements are used for testing the trained IT2FS. In this embodiment, the PMF parameters of the second IT2FS (DBP) are assumed as the means of AHA parameters' ranges for each parameter of the recorded DBP. The optimization process, which in this example uses the particle swarm optimization process is utilized to find the optimum parameters' values within the AHA parameters' ranges .

The PMF parameters of IT2FS were adjusted within AHA ranges of DBP by using same optimization process outlined above for the first IT2FS (SBP) to achieve the smallest average of absolute differences between real and estimated values of DBP.

Likewise, the second IT2FS, which is transformed into an embodiment of the HR/BP relationship 133, is used to determine DBP values for the test group. These determined DBP values, HR values, Real DBP values and the absolute differences between determined DBP and Real DBP are shown in the following table

Table 3

(Table 3 - Sample DPB results of trained IT2FS) The accuracy of the HR/BP relationship 133 for DBP is calculated by equation:

Accurac — 100%^x (1 ^— average of absolute difference between Estimated and Real values ^ average of Real values

In this embodiment, the HR/BP relationship 133 for

DBP accuracy = 100 % ^χ ( ) = 91.48

After determining the SBP and DBP results, the MAP values can be computed based on determined SBP and DBP with this equation:

MAP = DBP + - x (SBP - DBP)

3

In this embodiment, the determined MAP values, HR values, Real MAP values and the absolute differences between determined MAP and Real MAP for test group are shown in the following table. Table 4

(Table 4 - Sample MAP results from trained IT2FS) The accuracy of estimated MAP values is calculated by equation:

In this embodiment, the accuracy of estimated MAP values = 100 % ^χ ) = 94.79

In these embodiments, the results have indicated that the accuracy of the determined value by this

embodiment of HR/BP relationship 133 is sufficiently accurate to provide useful results. These embodiments of the HR/BP relationship 133 may then be used in apparatus 200 to determine a blood pressure based on the heart rate measured from a PPG or otherwise obtained through other means. These embodiments are particularly advantageous in that they are able to obtain accurate results of blood pressure without the risk of invasive procedures. In addition, the use of the HR/BP relationship 133 with the PPG does not require calibration, and thereby increasing the efficiencies of measuring the blood pressure results of a patient.

In an alternative embodiment, the determination of the blood pressure values for healthy patients may be carried out by use of a similar heart rate/blood pressure relationship wherein the data samples comprise mainly healthy patients. By changing the database to use healthy subjects, a similar heart rate/blood pressure relationship can be used accurately to determine the blood pressure for healthy patients.

Although not required, the embodiments described with reference to the figures can be implemented as an application programming interface (API) or as a series of libraries for use by a developer or can be included within another software application, such as a terminal or personal computer operating system or a portable computing device operating system. Generally, as program modules include routines, programs, objects, components and data files assisting in the performance of particular

functions, the skilled person will understand that the functionality of the software application may be

distributed across a number of routines, objects or components to achieve the same functionality desired herein .

It will also be appreciated that where the methods and systems of the present invention are either wholly implemented by computing system or partly implemented by computing systems then any appropriate computing system architecture may be utilised. This will include stand alone computers, network computers and dedicated hardware devices. Where the terms "computing system" and

"computing device" are used, these terms are intended to cover any appropriate arrangement of computer hardware capable of implementing the function described.

It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the

invention as broadly described. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.

Any reference to prior art contained herein is not to be taken as an admission that the information is common general knowledge, unless otherwise indicated.

Claims

CLAIMS :

1. A method of determining blood pressure comprising the steps of:

obtaining a heart rate; and

2. A method in accordance with Claim 1, wherein the step of analysis of the data samples comprises the step of fuzzy logic processing.

3. A method in accordance with Claim 2, wherein the fuzzy logic processing comprises the step of obtaining fuzzifier variables from the obtained heart rate.

4. A method in accordance with Claim 3, wherein the fuzzy logic processing further comprises the step of evaluation of a set of rules, wherein the rules are associated with fuzzifier variables to select a solution set of fuzzifier variables.

5. A method in accordance with any one of Claims 4, wherein the fuzzy logic process further comprises the step of processing the solution set of fuzzifier variables to determine the blood pressure result.

6. A method in accordance with any one of Claims 2 to 5, wherein the fuzzy logic processing uses an input fuzzy set which comprises a plurality of membership functions associated with a range of heart rate readings.

7. A method in accordance with Claim 6, wherein the input fuzzy set further comprises a plurality of heart rate parameters.

8. A method in accordance with any one of Claims 2 to

7, wherein the fuzzy logic processing uses an output fuzzy set which comprises a plurality of membership functions associated with a range of blood pressure readings.

9. A method in accordance with Claim 8, wherein the output fuzzy set further comprises a plurality of blood pressure parameters.

10. A method in accordance with any one of Claims 6 to 9, wherein the range of heart rate readings and the plurality of heart rate parameters are optimized.

11. A method in accordance with Claim 10, wherein the range of heart rate readings and the plurality of heart rate parameters are optimized by a first optimization process arranged to comprise the steps of:

comparing the blood pressure result with a real blood pressure value determined from the data samples; and adjusting the range of heart rate readings and the plurality of heart rate parameters to reduce the

difference between the blood pressure result and the real blood pressure value.

12. A method in accordance with any one of Claims 8 to 11, wherein the range of blood pressure readings and the plurality of blood pressure parameters are optimized.

13. A method in accordance with Claim 12, wherein the range of blood pressure values and the plurality of blood pressure parameters are optimized by a second optimization process arranged to comprise the steps of:

comparing the blood pressure result with a real blood pressure value determined from the data samples; and adjusting the range of blood pressure readings and the plurality of blood pressure parameters.

14. A method in accordance with any one of Claims 10 to 13, wherein the first and second optimization process comprises a multi particle swarm optimization process.

15. A method in accordance with Claim 14, wherein the multi particle swarm optimization process comprises a best vector .

16. A method in accordance with Claim 15, wherein the best vector is associated with the heart rate/blood pressure relationship.

17. A method in accordance with Claims 15 or 16, wherein the best vector comprises a function to determine the blood pressure result.

18. A method in accordance with any one of Claims 14 to

17, wherein the best vector is determined by comparing the blood pressure result with the real blood pressure value.

19. A method in accordance with any one of Claims 2 to

18, wherein the fuzzy logic processing is an interval type 2 fuzzy logic process.

20. A method in accordance with any one of Claims 6 to 19, wherein the plurality of membership functions is associated with a gaussian function.

21. A method of establishing a heart rate/blood pressure relationship comprising the step of analysis of data samples comprising heart rate values and blood pressure values from a plurality of subjects.

22. A method in accordance with Claim 21, wherein the step of analysis of the data samples comprises the step of fuzzy logic processing.

23. A method in accordance with Claims 22, wherein the step of fuzzy logic processing comprises the step of obtaining fuzzifier variables from the obtained heart rate .

24. A method in accordance with Claim 23, wherein the step of fuzzy logic processing further comprises the step of evaluation of a set of rules, wherein the rules are associated with the fuzzifier variables to select a solution set of fuzzifier variables.

25. A method in accordance with any one of Claim 24, wherein the step of fuzzy logic processing further comprises the step of processing the solution set of fuzzifier variables to determine the blood pressure result .

26. A method in accordance with any one of Claims 22 to

25. wherein the step of fuzzy logic processing uses an input fuzzy set which comprises a plurality of membership functions associated with a range of heart rate readings.

27. A method in accordance with Claim 26, wherein the input fuzzy set further comprises a plurality of heart rate parameters.

28. A method in accordance with any one of Claims 22 to 27, wherein the step of fuzzy logic processing uses an output fuzzy set which comprises a plurality of membership functions associated with a range of blood pressure readings .

29. A method in accordance with Claim 28, wherein the output fuzzy set further comprises a plurality of blood pressure parameters.

30. A method in accordance with any one of Claims 26 to 29, wherein the range of heart rate readings and the plurality of heart rate parameters are optimized.

31. A method in accordance with Claim 30, wherein the range of heart rate readings and the plurality of heart rate parameters are optimized by a first optimization process arranged to comprise the steps of:

difference between the blood pressure result and the real blood pressure value.

32. A method in accordance with any one of Claims 28 to

31. wherein the range of blood pressure readings and the plurality of blood pressure parameters are optimized.

33. A method in accordance with Claim 32, wherein the range of blood pressure values and the plurality of blood pressure parameters are optimized by a second optimization process arranged to comprise the steps of:

34. A method in accordance with any one of Claims 30 to 33, wherein the first and second optimization process comprises a multi particle swarm optimization process.

35. A method in accordance with Claim 34, wherein the multi particle swarm optimization process comprises a best vector .

36. A method in accordance with Claim 35, wherein the best vector is associated with the heart rate/blood pressure relationship.

37. A method in accordance with Claims 35 or 36, wherein the best vector comprises a function to determine the blood pressure result.

38. A method in accordance with any one of Claims 34 to

37, wherein the best vector is determined by comparing the blood pressure result with the real blood pressure value.

39. A method in accordance with any one of Claims 22 to

38, wherein the fuzzy logic process is an interval type 2 fuzzy logic process .

40. A method in accordance with any one of Claims 26 to 39, wherein the plurality of membership functions is associated with a gaussian function.

41. An apparatus for determining a heart rate/blood pressure relationship comprising a processor arranged to analyse data samples from a plurality of subjects, the analysis comprising a function arranged to correlate recorded heart rate values and recorded blood pressure values for each of the plurality of subjects in the data samples to establish the heart rate/blood pressure relationship .

42. An apparatus in accordance with Claim 41, wherein the processor arranged to analyse data comprises a function for fuzzy logic processing.

43. An apparatus in accordance with Claims 42, wherein the fuzzy logic processing comprises fuzzifier variables obtained from the obtained heart rate.

44. An apparatus in accordance with Claim 43, wherein the fuzzy logic process further comprises a procedure for evaluating a set of rules, wherein the rules are

associated with the fuzzifier variables to select a solution set of fuzzifier variables.

45. An apparatus in accordance with any one of Claims 43 or 44, wherein the fuzzy logic process further comprises a function arranged to process the solution set of fuzzifier variables to determine the blood pressure result.

46. An apparatus for measuring blood pressure

comprising : a processor arranged to determine a blood pressure result based on a patient' s heart rate and a heart rate/blood pressure relationship;

47. An apparatus in accordance with Claim 46, wherein the processor arranged to analyse the data samples comprises a function for fuzzy logic processing.

48. An apparatus in accordance with Claims 47, wherein the fuzzy logic processing comprises fuzzifier variables obtained from the obtained heart rate.

49. An apparatus in accordance with Claim 48, wherein the fuzzy logic processing further comprises a procedure for evaluating a set of rules, wherein the rules are associated with the fuzzifier variables to select a solution set of fuzzifier variables.

50. An apparatus in accordance with any one of Claims 48 or 49, wherein the fuzzy logic processing further

comprises a function arranged to process the fuzzifier variables to determine the blood pressure result.

51. An apparatus in accordance with any one of Claims 47 to 50, wherein the fuzzy logic processing uses an input fuzzy set which comprises a plurality of membership functions associated with a range of heart rate readings.

52. An apparatus in accordance with Claim 51, wherein the input fuzzy set further comprises a plurality of heart rate parameters.

53. An apparatus in accordance with any one of Claims 47 to 52, wherein the fuzzy logic processing uses an output fuzzy set which comprises a plurality of membership functions associated with a range of blood pressure readings .

54. An apparatus in accordance with Claim 53, wherein the output fuzzy set further comprises a plurality of blood pressure parameters.

55. An apparatus in accordance with any one of Claims 52 to 54, wherein the range of heart rate readings and the plurality of heart rate parameters are optimized.

56. An apparatus in accordance with Claim 55, wherein the range of heart rate readings and the plurality of heart rate parameters are optimized by a first

optimization process arranged to comprise:

a comparator arranged to compare the blood pressure result with a real blood pressure value determined from the data samples; and

an adjusting function arranged to adjust the range of heart rate readings and the plurality of heart rate parameters to reduce the difference between the blood pressure result and the real blood pressure value.

57. An apparatus in accordance with any one of Claims 54 to 56, wherein the range of blood pressure readings and the plurality of blood pressure parameters are optimized.

58. An apparatus in accordance with Claim 57, wherein the range of blood pressure values and the plurality of blood pressure parameters are optimized by a second optimization process arranged to comprise:

a comparator arranged to comparing the blood pressure result with a real blood pressure value

determined from the data samples; and

an adjusting function arranged to adjust the range of blood pressure readings and the plurality of blood pressure parameters.

59. An apparatus in accordance with any one of Claims 56 to 59, wherein the first and second optimization process comprises a multi particle swarm optimization process.

60. An apparatus for determining cardiovascular

parameters comprising:

at least one sensor arranged to detect at least one signal from a patient;

wherein the at least one signal is used to determine each one of the heart rate, pulse wave velocity and oxygen saturation levels of the patient.

61. An apparatus in accordance with Claim 60, wherein the at least one signal is a Photoplethysmograph signal.

62. A method for determining blood pressure comprising the steps of:

obtaining a heart rate from a Photoplethysmograph signal; and

63. An apparatus for determining blood pressure comprising:

64. An apparatus in accordance with Claim 63, wherein the device arranged to obtain a heart rate is a

Photoplethysmograph device.

65. An apparatus for determining blood pressure comprising:

a Photoplethysmograph device arranged to obtain at least one Photoplethysmograph signal, wherein the signal is processed to determine a heart rate; and

a processor arranged to determine a blood pressure result based on the heart rate and a heart rate/blood pressure relationship;

66. A computer program, comprising instructions for controlling a computer to implement a method of

determining blood pressure in accordance with any one of Claims 1 to 20.

67. A computer readable medium, providing a computer program in accordance with Claim 66.

68. A data signal, comprising a computer program in accordance with Claim 66.

69. A computer program, comprising instructions for controlling a computer to implement a method for establishing a heart rate/blood pressure relationship in accordance with any one of Claims 21 to 40.

70. A computer readable medium, providing a computer program in accordance with Claim 69.

71. A data signal, comprising a computer program in accordance with Claim 69.