WO2016207672A2

WO2016207672A2 - Method and device for determining fetal heart sounds by passive sensing and system for examining fetal heart function

Info

Publication number: WO2016207672A2
Application number: PCT/HU2016/000040
Authority: WO
Inventors: Imre Bartos; Imre JUHÁSZ; János KÁLDI; Sándor SERES; István Székely; Miklós TÖRÖK
Original assignee: Pentavox Mernoki Menedzsment Es Kereskedelmi Kft
Current assignee: Pentavox Mernoki Menedzsment Es Kereskedelmi Kft
Priority date: 2015-06-25
Filing date: 2016-06-23
Publication date: 2016-12-29
Anticipated expiration: 2017-12-25
Also published as: HU230994B1; WO2016207672A3; CN106793996A; CN106793996B; HUP1500298A2

Abstract

In the method according to the invention a phonocardiographic signal, obtained by a passive acoustic sensor (10) from a maternal abdominal wall, is pre-filtered, amplified, digitized, digitally filtered by means of a programmable amplifier and filter unit (20), is stored in an intermediate memory and autocorrelation is performed by means of an autocorrelation unit (30) in a time window of a predetermined size. The signals obtained as a result of autocorrelation are fed to an input of a fuzzy expert system (40), wherein local maximums of the signals, temporal location of the local maximums and variation of the temporal location of the local maximums are determined by the fuzzy expert system (40), and, applying those as input parameters or signals of the fuzzy expert system, those are classified into probability groups - utilizing a fuzzy rule set of a decision logic stored in a knowledge base and biologically expected data - and are evaluated.

Description

METHOD AND DEVICE FOR DETERMINING FETAL HEART SOUNDS BY PASSIVE SENSING AND SYSTEM FOR EXAMINING FETAL HEART

FUNCTION TECHNICAL FIELD

The invention relates to a method for determining fetal heart sounds by passive sensing and to a device for carrying out the method, which device can be preferably utilized in home conditions, and can also be preferably applied in telemedicine. The invention further relates to a system for examining fetal heart function.

BACKGROUND ART

It is known that clearly interpretable measurement results of fetal heart function can only be obtained through tests with a longer duration (approximately 20 minutes); with such duration there is a chance for detecting, or noticing in due time, phenomena related to possible symptoms of heart disorders.

The type of measurement most frequently used nowadays is the ultrasonic method based on the Doppler principle, which is an active method and thereby involves continuous energy irradiation for a non-negligible period of time and is not suited for tests carried out at home without supervision. There is a very limited choice of passive methods adapted for examining fetal heart function, primarily due to problems related to physical interfacing with and access to the maternal body, and secondly due to the fact that the tests require substantial expertise on the part of the persons performing them and last but not least, due to the excessive costs of such tests.

To circumvent the problems arising with ultrasonic methods and eliminate the need for an expert person being present during the test, phonocardiographic methods, such as those disclosed in US 6,551 ,251 and US 6,749,573 were developed. These testing methods apply multiple acoustic sensors, with the interpretable fetal heart sound signals being produced by processing the signals of the sensors separately and combining them. However, due to problems related to placing the sensors on the maternal abdomen and supporting them against the skin the application of multiple sensors is cumbersome and results in a poor signal to noise ratio, thereby allowing measurements with only limited accuracy and failing to fulfil the requirements of home and telemedicine application.

Known solutions also include the method, device and triple-chamber acoustic sensor according to EP 0850014 and US 6,245,025, which allow for sufficiently accurate fetal heart function examining and registration of heart function signals starting from a certain age of the fetus that is determined by a number of circumstances, and are also suitable for telemedicine applications. However, with more complex cases there arises a need for improving sensitivity and accuracy, and for the possibility to carry out phonocardiography measurements starting at an earlier stage of pregnancy. According to known solutions the signals produced by periodic heartbeats are intended to be filtered from the detected noise-laden signals by means of autocorrelation, which results in the reduction of lower-amplitude non-periodic signals (essentially, noise) but sensitivity and accuracy cannot be increased. In a number of solutions error-correction mathematical methods are applied for filtering the sounds produced by heartbeats from the noisy signals, but even these are incapable of improving effective sensitivity and accuracy.

DISCLOSURE OF THE INVENTION

The object of the invention is to provide a phonocardiography method and device that provides improved sensitivity, reliability and accuracy for determining the sounds of fetal heart function, and thereby renders fetal heart function tests simpler, providing the possibility of carrying out such tests at an earlier developmental stage.

Our further objective is to provide a method that allows for obtaining clearly interpretable results from tests carried out without supervision, at home or in telemedicine. Related to that, a still further objective of our invention is to provide an improved acoustic sensor that allows for optimal detection of fetal heart sounds for subsequent evaluation also in tests carried out without medical supervision, at home, or in telemedicine.

The invention is based on the recognition that autocorrelation - which reduces non- periodic noise - is not primarily applied for enhancing signals resulting from the periodic heart sounds, but rather for further noise filtering, such that the amplitude of low-amplitude signals produced by autocorrelation is increased by means of further processing, and then the temporal location of the signals is determined with high accuracy.

A further recognition of the invention is that such an improved sensor is suited for fulfilling our objectives that, in addition to being easily operable by anyone, can provide optimal acoustic coupling to the maternal lower abdominal skin. Therefore, in the solution according to the invention fetal heart sounds are determined in the following way: a phonocardiographic signal, obtained by a passive acoustic sensor from a maternal abdominal wall, is pre-filtered, amplified, digitized, digitally filtered,

stored in an intermediate memory, autocorrelation is performed in a time window of a predetermined size, and then local maximums of the signals obtained as a result of autocorrelation are determined, temporal location of the local maximums and variation of the temporal location of the local maximums are determined, and, applying the results as input parameters or signals of a fuzzy expert system, those are classified into probability groups - utilizing a fuzzy rule set of a decision logic stored in a knowledge base and biologically expected data - and are evaluated, the numeric results of the evaluation are stored for further processing in an input and output memory and in a control memory.

In a preferred way of carrying out the method according to the invention, the values stored in the control memory are applied for adjusting the gain of the amplifier and the parameters of the digital filter to achieve a maximum amplitude for the fetal heart rate signal and a minimum for disturbance signals. In a further preferred way of carrying out the method according to the invention the rules applicable for making inferences are selected applying the knowledge base and the inference system of the fuzzy expert system and based on the data stored in the knowledge base, and based on that, the temporal distance (time difference) between the signals and thereby the values of fetal heart rate (FHR) are determined by the fuzzy sets.

In a preferred way of carrying out the method according to the invention the digitized sound signal derived from fetal heart sounds is transposed to a higher frequency and the transposed sound is audibly displayed.

In a further preferred way of carrying out the method according to the invention the fetal heart rate signal previously stored in memory is displayed, data - primarily personal and medical data - are assigned to the stored heart rate signal, and the data are stored locally or remotely for evaluation and archiving.

The device constituting a further object of the invention is essentially characterised in that

it comprises an acoustic sensor adapted to be placed on a maternal abdominal wall and for converting acoustic signals of fetal heart function to electric signals, and further comprises

a programmable amplifier and filter unit comprising

a programmable analogue amplifier connected to the acoustic sensor, an analogue filter connected to the output of the analogue amplifier, an A/D converter connected to the output of the analogue filter, a digital filter connected to the output of the A/D converter, and an intermediate memory connected to the output of the digital filter, and

an autocorrelation unit having a delay circuit, a multiplier circuit and an integrator circuit, wherein the input of the delay circuit and a first input of the multiplier circuit are interconnected and constitute the input of the autocorrelation unit, and are connected to the output of the programmable amplifier and filter unit, the output of the delay circuit being connected to a second input of the multiplier circuit, and the output of the multiplier circuit being connected to the input of the integrator circuit, wherein the output of the autocorrelation unit is constituted by the output of the integrator, the device further comprising

a fuzzy expert system comprising a parameter normalization unit adapted for receiving data signals and converting them to input signals, a fuzzification unit that is connected to the output of the parameter normalization unit and is adapted for converting the input signals to an input signal complying with fuzzy rules, a fuzzy inference system connected to the output of the fuzzification unit and comprising a decision logic corresponding to the inference process, a knowledge base unit connected with the fuzzy inference system via a bidirectional link and comprising memory storing a data set defining boundary conditions and logical circuits defining decision rules, and a defuzzification unit that is connected to the output of the fuzzy inference system and is adapted for converting the logical decisions to a data signal, the input component of the fuzzy expert system being the parameter normalization unit that is connected to the output of the autocorrelation unit,

wherein the defuzzification unit of the fuzzy expert system comprises

a first output providing temporal distance data of the fetal heart rate (FHR) and a second output providing desired parameters of the digital filter, the first output being connected to the input of an input/output MEM3 memory, and the second output being connected via a control MEM2 memory to the gain control input of the analogue amplifier and to the filtering control input of the digital filter, wherein the output of the programmable amplifier and filter unit is connected to an acoustic transducer adapted for converting the fetal heart sounds to an audible signal, and a modem providing bidirectional data traffic is connected to the input/output MEM3 memory.

The essence of the invention can be summarized in that the objective is realized by signal processing carried out applying a series-connected programmable amplifier and filter unit, autocorrelation unit, and fuzzy expert system, and being optimized by controlling amplification and filtering. A preferred embodiment of the device comprises a microprocessor circuit comprising an input A/D converter, the circuit comprising the A/D converter, the digital filter and the intermediate MEM1 memory of the programmable amplifier and filter unit, and also the autocorrelation unit, the fuzzy expert system, as well as the control MEM2 memory and the input/output MEM3 memory. A preferred embodiment can also be contemplated wherein the microprocessor is implemented and programmed as the microprocessor of a mobile phone, iPOD, or other computing device. In a further preferred embodiment of the device the input/output MEM3 memory comprises an external data input.

In a further preferred embodiment of the device the acoustic sensor comprises a housing having an internal space divided into an open first chamber and closed second and third chambers, a sound conduction opening disposed in a separator wall between the first and second chambers, an electromechanical acoustic transducer disposed in the third chamber, the sensor membrane of the electromechanical acoustic transducer being disposed in a wall situated between the second chamber and the third chamber, the first chamber having an edge adapted to fit against a maternal abdominal wall, wherein the first chamber is closed in a position where it is fitted against the maternal abdominal wall, and wherein a respective low-diameter pressure equalization opening, opening to the outside air, is disposed in the side wall of each of the first chamber and of the third chamber. The fetal heart function examining system according to the invention is essentially characterised in that it comprises one or more fetal heart rate examining devices according to the invention, a respective interfacing modem connected to the output thereof, a central computer adapted for medical evaluation and archiving, one or more computers adapted to be used by treating physicians and/or mobile phones capable of data communication and display, the units being interconnected through an internet network via a data connection carrying bidirectional data traffic. BRIEF DESCRIPTION OF THE DRAWINGS

The essence of the invention is described in detail below with reference to the accompanying schematic drawings, where

Fig. 1 shows an essentially disturbance-free acoustic signal of fetal heart sounds,

Fig. 2 illustrates major steps of the method according to the invention,

Fig. 3 is a block diagram of the device according to the invention,

Fig. 4 is a block diagram of a preferred embodiment of the device according to Fig. 3, partly implemented with a microprocessor,

Fig. 5 shows a schematic cross-sectional view of a preferred embodiment of the acoustic sensor used in the device, and

Fig. 6 is a possible arrangement of the system according to the invention.

MODES FOR CARRYING OUT THE INVENTION

Fig. 1 shows a disturbance-free acoustic (phonocardiographic) signal of fetal heart sounds. In the figure there can be easily seen the two characteristic signals S1 and S2 corresponding to the opening (S1) and closure (S2) of the fetal heart valve; the instantaneous heart rate (the instantaneous FHR) being determined on the basis of the time difference between the subsequent signals S1 or S2.

In Fig. 2 the major steps of the method according to the invention are illustrated. In the course of the method, a phonocardiographic signal obtained from a maternal abdominal wall by means of a passive sensor is amplified, digitized, filtered, and is subjected to an autocorrelation process applying a time window of a predetermined size, and the signal sequence obtained as a result of autocorrelation is processed by a fuzzy expert system.

During processing, the signals obtained as a result of autocorrelation - being an input parameter of the fuzzy expert system - are classified in probability groups utilizing a rule set of a decision logic stored in a knowledge base (the rule set comprising biologically expected signal ranges and the corresponding probabilities of occurrence), and the signals are then evaluated. Local maximums, temporal location of the local maximums and changes of temporal location of the local maximums of the signals obtained as a result of processing are determined. By means of a first output signal of the fuzzy expert system the amplification gain is adjusted to achieve an optimum gain, a second output signal of the fuzzy expert system is applied for adjusting filter parameters such that minimal noise is achieved, and a third output signal, comprising the frequency of the fetal heart sound, is stored for further evaluation and processing.

Fig. 3 shows a block diagram of a device according to the invention.

The device has an acoustic sensor 10 and a programmable amplifier and filter unit 20, wherein the output of the acoustic sensor 10 is connected to the input of the programmable amplifier and filter unit 20. The programmable amplifier and filter unit 20 has a controllable-gain analogue amplifier 22, an analogue filter 23, an AID converter 24, a programmable digital filter 25, and intermediate MEM1 memory 26, all connected in series connection. The analogue amplifier 22 and the digital filter 25 have respective control inputs adapted for adjusting the gain of the analogue amplifier 22, as well as one or more parameters - such as the frequency range, slope, damping, etc. - of the programmable digital filter 25.

The device further comprises an autocorrelation unit 30 having a delay circuit 31 , a multiplier circuit 33, and an integrator 35. The output of the programmable amplifier and filter unit 20 is constituted by the output of the intermediate MEM1 memory 26 that is connected on the one hand to the input of the delay circuit 31 , and on the other hand, to a first input of the multiplier circuit 33, wherein the output of the delay circuit 31 is connected to a second input of the multiplier circuit 33. The output of the multiplier circuit 33 is connected to the input of an integrator 35, wherein the output of the integrator 35 is connected to the input of a fuzzy expert system 40. The input unit of the fuzzy expert system 40 is a parameter normalization unit 41 , the output of which being connected to the input of a fuzzification unit 43, the output of the fuzzification unit 43 being connected to a first input of a fuzzy inference system 45. The fuzzy expert system 40 further comprises a knowledge base module 47 connected to the fuzzy inference system 45 by a connection providing bidirectional data transfer. The output of the fuzzy inference system 45 is connected to a defuzzification interface 49, wherein one output of the defuzzification interface 49 is connected via a control memory 60 to the respective control inputs of the analogue amplifier 22 and of the digital filter 25 of the programmable amplifier and filter unit 20, and another output thereof is connected to the input of an input/output MEM3 memory 50. The memory 50 has an external data input 90, and an interfacing modem 80 is connected to the memory 50.

An acoustic display unit 70 adapted for converting fetal heart sounds to audible signals is also connected to the output of the programmable amplifier and filter unit 20.

The device operates by converting, by the acoustic sensor 10, the fetal heart sounds detected in an acoustic manner at the maternal abdominal wall into an electric signal, amplifying the converted signal in a controlled manner to the desired level by the analogue amplifier 22, and - since they fall outside the frequency range of fetal heart sound signals - applying an analogue filter 23 for filtering out higher-frequency components from the detected signal. The amplified and filtered signals are converted by the A/D converter 24 into a digital signal, and the digitized signal is further filtered applying a programmable digital filter 25, filtering out the components falling outside the frequency range of the useful signal, while - taking into account the age of the fetus, the build of the mother and the position of the fetus - the signals falling into the useful range are adjusted to an optimum amplitude. Filtering on the one hand reduces low-frequency components, maternal heart sounds and other disturbance signals, and on the other hand also reduces the higher-frequency components of external disturbances. The amplifier gain and the filtering parameters are programmed for a maximum fetal heart sound signal and a minimal disturbance signal. The digital signal sequence thereby obtained is stored in the intermediate MEM1 memory 26.

The correlation function is produced by the autocorrelation unit 30 according to the following formula:

where

t stands for time

T stands for delay time (time distance between clock signal pul

R(T) stands for the autocorrelation function

v(t) stands for the original function

v(t-T) stands for the delayed original function

The delay time of the delay circuit 31 is selected to be so small that the applied resolution of the fetal heart sound signal can result in an autocorrelated signal having a sufficient accuracy. As a result of autocorrelation periodic signals retain their periodicity, while non- periodic signals (i.e. mostly noise) are significantly reduced in amplitude.

The intermediate MEM1 memory 26 of the programmable amplifier and filter unit 20 is connected to the input of the autocorrelation unit 30 and also to the input of the input/output MEM3 memory 50 and to the acoustic display unit 70 adapted to display fetal heart sounds.

The length and size of the autocorrelation function corresponds to the minimum biologically possible heart rate and to the sampling values, i.e. in our case it contains the fetal heart sound signals of at least four fetal heartbeats. Applying a parameter normalization unit 41 of the fuzzy expert system 40, parameter normalization is performed on the output signal obtained from autocorrelation, which comprises the steps of

- detecting peaks in the signal (at least four amplitude maximums are detected),

- ordering the detected peaks by their amplitude,

- selecting at least four signals with maximum amplitude.

The selected signals most probably constitute a fetal heart signal sequence. Because heart sound signals may be masked by noise, the detected peaks (maximums) have to be tested for really being characteristic of heart function. It may happen, by way of example, that the fetal heart sound signal cannot be found due to noise, but it can still be detected or determined utilizing the fetal heart sound signal sequence. In these cases the measured signal sequence does not have to be discarded, but rather the location of the fetal heart sound signal should be determined utilizing the other accepted values.

The signals obtained from this parameter normalization operation are fed to the input of the fuzzification unit 43. In the course of fuzzification the information content of the signals is converted into fuzzy input signals according to fuzzy rules.

By means of the fuzzy input signals and on the basis of inference rules and data, such as physiological rules and previous measurement data, defined in the fuzzy language and stored in the knowledge base module 47, the fuzzy inference system 45 is applied for making an inference by fuzzy and logical operations, during which the probability of individual signals with maximums falling into a given processible signal group is determined, and the obtained result is applied for prescribing new strategies to be stored in the knowledge base module 47, in the course of which the fuzzy inference rules and inference data are modified corresponding to the results. Defuzzification performed utilizing the defuzzification interface 49 involves converting the output signal of the inference system into numerical signals, based on which it can be determined whether a useful heart sound signal has been found. Once the majority of useful heart signals have been found, the values (the temporal locations of the peak amplitudes) are stored in the input/output MEM3 memory 50 as the result of defuzzification. If the fetal heart sound signal (and its recurrences) could not be found based on the evaluation provided by the fuzzy inference system, then this measurement result is considered as uninterpretable, and the search is restarted.

In the course of evaluation a second output signal is generated by the fuzzy inference system 45, which second output signal is proportional to the amplitude of the fetal heart sound signal and to the noise level caused by other signals considered as disturbance signals from the aspect of the measurement. This second signal is stored in the control MEM2 memory 60, and is also utilized for controlling the amplifier 22 and the filter 25 in a sense that, acting as a control loop, it adjusts the gain of the amplifier to achieve the maximum possible value of the fetal heart sound signal and sets the filter parameters (bandwidth, slope) such that the lowest possible noise level is obtained.

The device can be connected to external systems via the modem 80 connected to the input/output MEM3 memory 50, while further data, by way of example, data describing fetal movements, can be entered simultaneously with the measurement via an external data input 90.

Fig. 4 illustrates the block diagram of an embodiment of the device according to Fig. 3, implemented, by way of example, in part utilizing a microprocessor. The microprocessor can optionally be implemented as a mobile phone, e.g. as a so- called smartphone capable of data communication and display. The structure of the microprocessor includes any such circuit components and programming capabilities that provide for AID conversion, programmable digital filtering, intermediate data storage, programmable implementation of autocorrelation and of the components of the fuzzy expert system, as well as for writing and fetching data to and from memory units, and connection to a modem.

The device has an acoustic sensor 10, to which the amplifier 22 of the programmable amplifier and filter unit 20 and - connected in series therewith - the filter 23 are connected. The output of the filter 23 is connected to the input of the A/D converter of the microprocessor.

The microprocessor further comprises the digital filter, the programmable units adapted for performing autocorrelation and the modules of the fuzzy expert system, as well as the intermediate and control memories and the memory adapted for storing measurement results and other data. Further data related to measurements can be entered to the memory through an external data input 90, and the microprocessor can be interfaced to data traffic systems by the modem 80. The acoustic display unit 70 is also connected to the microprocessor.

Fig. 5 shows a cross sectional view of the schematic structural arrangement of the acoustic sensor 10. This configuration has maximum sensitivity in the frequency range of fetal heart sounds. The acoustic sensor 10 has a housing 11 comprising an internal space divided into a first chamber 13, a second chamber 16, and a third chamber 19. The chamber 13 is connected with the chamber 16 by a sound conduction opening 14 disposed in the separator wall situated between them. At its side situated opposite the sound conduction opening 14 the chamber 13 is open, while the chamber 13 has a rigid side wall, an edge 12 being formed in the side wall. During use, the edge 12 of the chamber 13 is supported against the maternal abdominal wall, the chamber 13 being closed by the skin surface enclosed by the edge 12. The acoustic sensor 10 is placed on the maternal abdominal wall in a biased (pre-loaded) state, which pressure is generated by a flexible belt, and thereby the enclosed portion of the skin surface acts as a membrane. The external configuration of the chamber 13 and the internal configuration of the chamber 16, together with the acoustic coupling between the chambers provides that the acoustic sensor 10 has the desired frequency characteristics. An electromechanical acoustic transducer 17 is disposed in the chamber 19, with the sensor membrane being disposed in the wall of the chamber 16 situated opposite the sound conduction opening 14. The chamber 16 has rigid side walls. Proper acoustic coupling between the maternal abdominal wall and the electromechanical acoustic transducer 17 is ensured by the volume of the chambers 13 and 16 and the dimensions of the sound conduction opening 14.

A respective low-diameter pressure equalization opening 15, 18, opening to the outside air, is disposed in the side wall of the chamber 13 and in the side wall of the chamber 19. These openings are adapted for reducing the damping effect of the air mass brought in motion by the oscillations of the enclosed skin surface. At the same time, the opening 15 also functions as a high-pass filter adapted for damping maternal heart sounds. The opening 18 is adapted for reducing the effects of the air pocket formed behind the electromechanical acoustic transducer 17, and provides protection against background noises through compensation.

Thereby the acoustic sensor 10 transduces and provides for further processing a relatively acoustically pre-filtered signal.

In Fig. 6 a configuration of a conceivable system for examining fetal heart function is shown. The system allows for performing fetal heart sound examining at home, primarily utilizing the device 100, for the verification of the measurements by a physician, and for storing and archiving measurement data and other information. The system comprises one or more devices 100, an interfacing modem 80 for each device 100, as well as a central computer 105 adapted for evaluation and archiving, and one or more computers 107 adapted to be used by treating physicians and/or mobile phones 109 capable of data communication and display, the units being interconnected through an internet network via a data connection carrying bidirectional data traffic. When the system is in use, the results measured by the device 100 are sent to a medical evaluation and archiving centre via the modem 80, where the data are evaluated, stored and archived. Equipped with an internet connection, the treating physician may also perform measurement evaluation and send information to the mother and to the central computer using his/her own computer 107 or mobile phone 109. Thereby detection (measurement), display and measurement evaluation can be temporally and spatially separated from one another.

The most important advantage of the method according to the invention, the device carrying out the method and the system implemented applying the device is that, compared to known devices comprising acoustic sensors, it allows for examining and evaluation of fetal heart sounds in an earlier stage of pregnancy (before the last trimester), and, due to the appropriate configuration of the sensor and to the signal processing solution provided by the series connection of the autocorrelation unit and the fuzzy expert system it allows for more accurate and higher-sensitivity fetal heart sound monitoring compared to known devices. A further advantage related to that is that it is capable of carrying out tests without medical supervision, at home, or in telemedicine.

Another advantage of the method and device according to the invention is that - in addition to measuring fetal heart rate - symptoms related to other cardiac conditions can also be detected.

List of reference numerals

10 acoustic sensor

11 housing

12 edge

13 (first) chamber

14 sound conduction opening (connecting/coupling sound conduction opening)

15 opening (pressure-equalizing bores for damping)

16 chamber (second)

17 electromechanical acoustic transducer (acoustic sensor)

18 opening (equalization), pressure equalization bore

19 chamber (third), balancing 20 programmable amplifier and filter unit

22 amplifier (analogue, adjustable gain)

23 filter (analogue)

24 A/D converter

25 filter (programmable digital)

26 memory (intermediate MEM 1 )

30 autocorrelation unit

31 delay circuit

33 multiplier circuit

35 integrator

40 fuzzy expert system (logical)

41 parameter normalization unit

43 fuzzification unit

45 fuzzy inference system (logical)

47 knowledge base module

49 defuzzification interface (module)

50 memory (input/output MEM3)

60 memory (control MEM2)

70 acoustic display unit

80 modem

90 external data input

100 device

105 central computer (medical evaluation and archiving)

107 computer (treating physician's)

109 mobile phone (treating physician's)

Claims

1. A method for determining fetal heart sounds by passive sensing, comprising the steps of pre-filtering, amplifying, digitizing, digitally filtering a phonocardiographic signal obtained by an acoustic sensor from a maternal abdominal wall,

storing in intermediate memory, and performing autocorrelation in a time window of a predetermined size,

characterised by

determining local maximums of the signals obtained as a result of autocorrelation, determining time position of the local maximums and determining variation of time position of the local maximums,

and, applying those as input parameters or signals of a fuzzy expert system, classifying them in probability groups by means of a fuzzy rule set of a decision logic stored in a knowledge base and of biologically expectable data, and evaluating the results,

modifying the fuzzy rule set utilizing the result of the evaluation,

defuzzifying the result of the evaluation,

and storing numeric results of defuzzification in an input and output memory and in a control memory for further processing.

2. The method according to claim 1,

characterised by

adjusting the gain of the amplifier and the parameters of the digital filter by means of the values stored in the control memory, to achieve a maximum amplitude for the fetal heart rate signal and a minimum for disturbance signals.

3. The method according to claim 1 or 2,

characterised by

selecting, applying the knowledge base and inference system of the fuzzy expert system and based on the data stored in the knowledge base, the rules applicable for making inferences, and based on that, determining, utilizing fuzzy logic, the temporal distance between the signals and thereby the values of fetal heart rate (FHR).

4. The method according to any of claims 1 to 3,

c h a r a c t e r i s e d b y

transposing the digitized sound signal derived from fetal heart sounds to a higher frequency and audibly displaying the transposed sound.

5. The method according to any of claims 1 to 4,

c h a r a c t e r i s e d b y

displaying the fetal heart rate values stored in memory, assigning further data to them, processing, archiving, and locally or remotely storing the fetal heart rate values.

6. A device for carrying out the method according to any of claims 1 to 5, comprising an acoustic sensor (10) adapted to be placed on a maternal abdominal wall and adapted for converting acoustic signals of fetal heart function to electric signals, a programmable amplifier and filter unit (20) comprising

a programmable analogue amplifier (22) connected to the acoustic sensor ( 0), an analogue filter (23) connected to the output of the analogue amplifier

(22), an A/D converter (24) connected to the output of the analogue filter (23), a digital filter (25) connected to the output of the A/D converter (24), and an intermediate memory (26) connected to the output of the digital filter (25), and an autocorrelation unit (30) having a delay circuit (31), a multiplier circuit (33) and an integrator (35), wherein one of the inputs of the multiplier circuit (33) and the input of the delay circuit (31) are interconnected and are connected to the output of the programmable amplifier and filter unit (20), the output of the delay circuit (31) being connected to the other input of the multiplier circuit (33), and the output of the multiplier circuit (33) being connected to the input of the integrator (35), wherein the output of the autocorrelation unit (30) is constituted by the output of the integrator (35), the device further comprising

a fuzzy expert system (40) connected to the output of the autocorrelation unit, a parameter normalization unit (41) being an input unit of the fuzzy expert system (40), the output of the parameter normalization unit (41 ) being connected to the input of a fuzzification unit (43), the output of the fuzzification unit (43) being connected to the input of a fuzzy inference system (45), a knowledge base module (47) being connected via bidirectional connection to the fuzzy inference system (45), wherein a database and a rule set are stored in the knowledge base module (47), the output of the fuzzy inference system (45) is connected to the input of a defuzzification interface (49), wherein the outputs of the fuzzy expert system (40) are constituted by the outputs of the defuzzification interface (49),

a first output of the defuzzification interface (49) is connected to the input of an input and output memory (50) storing fetal heart sound data, a second output of the defuzzification interface (49) being connected to a control memory (60) storing data applied for gain and filtering control, and

the output of the control memory (60) is connected to a gain control input of the amplifier (22) and to a filter programming input of the digital filter (25),

a modem (80) being connected to the input and output memory (50), and the input and output memory (50) is equipped with an external data input (90), and the output of the programmable amplifier and filter unit (20) is connected to a further input of the input and output memory (50) and to the input of the acoustic display unit (70) adapted for the audible display of fetal heart sounds.

7. The device according to claim 6,

c h a r a c t e r i s e d b y

having a microprocessor circuit equipped with an input A/D converter, the circuit comprising the A/D converter (24), the digital filter (25) and the intermediate memory (26) of the programmable amplifier and filter unit (20), and also the autocorrelation unit (30), the fuzzy expert system (40), as well as the control memory (60) and the input and output memory (50).

8. The device according to claim 6 or claim 7,

c h a r a c t e r i s e d i n t h a t the acoustic sensor (10) comprises a housing (1 1) having an internal space divided into an open first chamber (13) and closed second and third chambers (16, 19), a sound conduction opening (14) disposed in a separator wall between the first and second chambers (13, 16), an electromechanical acoustic transducer (17) disposed in the third chamber (19), the sensor membrane of the electromechanical acoustic transducer (17) being disposed in a wall situated between the second chamber (16) and the third chamber (19), the first chamber (13) having an edge (12) adapted to fit against a maternal abdominal wall, wherein the first chamber (13) is closed in a position where it is fitted against the maternal abdominal wall, and wherein a respective low-diameter pressure equalization opening (15, 18), opening to the outside air, is disposed in the side wall of each of the first chamber (13) and of the third chamber (19).

9. A system for examining fetal heart function, particularly implemented utilizing the device according to any of claims 6 to 8,

c h a r a c t e r i s e d b y

comprising one or more devices (100), a modem (80) connected to the output of the device(s), a central computer (105) adapted for medical evaluation and archiving, and one or more computers (107) adapted to be used by treating physicians and/or mobile phones (109) capable of data communication and display, the units being interconnected through an internet network via a data connection carrying bidirectional data traffic.