Delivery early warning method and system based on conjoint analysis of uterine contraction and fetal heart

Abstract

The invention relates to a delivery early warning method and system based on uterine contraction and fetal heart joint analysis, and belongs to the technical field of medical data processing. The early warning method comprises the following steps: arranging a plurality of electrode type A-type sensors on the abdomen of a pregnant woman, collecting uterine electromyographic signals, extracting a plurality of uterine activity indexes, and constructing uterine contraction behavior vectors under a time sequence; at least one magnetic intensity type class B sensor is arranged, a plurality of reference positioners are supplemented, weak magnetic field signals generated by fetal heart activity are collected and analyzed, a plurality of fetal response indexes are obtained, and a fetal response vector is constructed. The system further constructs a joint state vector based on stimulation (uterine contraction) and response (fetal heart) signals, compares the joint state vector with a pre-established normal sample reference model, calculates a deviation degree score, assesses whether the fetus has abnormal response in each time window, and outputs premature delivery or delivery risk levels in a grading manner.

Description

Technical Field

[0001] This invention belongs to the field of medical data processing technology, and specifically relates to a method and system for early warning of labor based on the combined analysis of uterine contractions and fetal heart rate. Background Technology

[0002] In obstetric clinical monitoring, accurately identifying impending labor or the risk of preterm birth is crucial. Traditional monitoring methods rely heavily on manual observation and experience-based judgment of fetal heart rate and uterine contraction patterns, which suffers from problems such as delayed response, strong subjectivity in interpretation, and limited accuracy in anomaly identification. Especially in late pregnancy or high-risk groups for preterm birth, monitoring of a single indicator often fails to fully reveal the true physiological response of the fetus to uterine contractions. In recent years, with the development of new signal detection methods, the combined analysis of uterine contraction behavior and fetal heart rate changes has become an important direction for improving monitoring accuracy. However, there is still a lack of an efficient modeling method that can systematically integrate dual signal acquisition, feature analysis, and risk assessment. Therefore, there is an urgent need to construct an intelligent labor early warning technology solution that integrates multi-source time-series signals and enables stimulus-response coupling analysis to improve the ability to identify labor risks early and ensure maternal and infant safety.

[0003] A review of relevant publicly available technologies reveals several key technologies. One proposed technology, JP2014045918A, describes a fetal heart rate monitoring microphone that also incorporates electrodes capable of detecting electrical signals from the mother's abdomen. This allows for monitoring of the fetal heartbeat through a combination of acoustic and electrode signal analysis. Another technology, CN113171085A, proposes a small uterine contraction monitor that uses an ultrasonic contraction probe to detect uterine contractions while continuously recording relevant data for medical personnel to review. Finally, GB2505424A presents a uterine contraction monitor utilizing fiber optic sensors. This monitor places a detector with multiple optical fibers on the pregnant woman's abdomen and determines the state of uterine contractions by observing the changes in the fibers as the abdomen moves.

[0004] The above technical solutions all propose using various methods or systems to assess possible labor warning signals in pregnant women and fetuses, but there are few related technical solutions that combine the two to achieve more refined analysis and assessment.

[0005] The foregoing description of the background art is intended only to facilitate understanding of the invention. This description does not endorse or acknowledge any common general knowledge in the materials mentioned. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for early warning of labor based on joint analysis of uterine contractions and fetal heart rate, belonging to the field of medical data processing technology. The early warning method involves deploying multiple electrode-type Class A sensors on the pregnant woman's abdomen to collect uterine electromyography signals, extract multiple uterine activity indicators, and construct a time-series vector of uterine contraction behavior; setting at least one magnetic intensity-type Class B sensor, supplemented by multiple reference locators, to collect and analyze the weak magnetic field signals generated by fetal cardiac activity, obtain multiple fetal response indicators, and construct a fetal response vector. The system further constructs a joint state vector based on the stimulus (uterine contractions) and response (fetal heart rate) signals, compares it with a pre-established normal sample reference model, calculates a deviation score, thereby assessing whether the fetus exhibits abnormal responses within each time window, and outputting a graded risk level for preterm birth or delivery accordingly.

[0007] This invention adopts the following technical solution: a labor early warning method based on joint analysis of uterine contractions and fetal heart rate, the early warning method comprising the following steps:

[0008] S100: Multiple Class A sensors are installed in the abdomen of the target individual to collect uterine electromyography (EMG) signals of the target individual at multiple time series; time windows are constructed using the collected uterine EMG signals, and multiple Class A index values ​​are extracted in each time window to form a uterine contraction behavior vector U for multiple time windows;

[0009] S200: Simultaneously, one or more Class B sensors are installed on the abdomen of the target individual to collect fetal cardiac magnetic field signals at multiple time series, and further analyze them into multiple Class B indicators related to fetal heart rate and fetal movement to form a fetal response vector F;

[0010] S300: Based on the uterine contraction behavior vector U, perform pattern recognition on the uterine behavior corresponding to each time window and divide it into 4 activity modes, namely: resting period, weak contraction period, frequent non-labor contractions or labor contractions.

[0011] S400: Extract the uterine contraction behavior vector U and the fetal response vector F from the same time window other than the resting period, and construct the joint state vector Z of the two vectors;

[0012] S500: A reference set M of samples constructed based on multiple pre-established normal uterine contraction-fetal response samples. ref ;

[0013] S600: Set a time window T k The joint state vector Z (k) With reference set M ref The mean vector and covariance matrix are compared to calculate the deviation R. (k) ;

[0014] S700: Based on a preset risk threshold, assess the deviation R. (k) The level of risk that the target individual is facing.

[0015] Simultaneously, a labor early warning system based on joint analysis of uterine contractions and fetal heart rate is proposed. This early warning system applies the aforementioned labor early warning method based on joint analysis of uterine contractions and fetal heart rate. The early warning system includes:

[0016] The monitoring unit A is configured to collect uterine electromyography (EMG) signals of the target individual at multiple time series. Time windows are constructed using the collected EMG signals, and multiple Class A index values ​​are extracted in each time window to form a uterine contraction behavior vector U for multiple time windows.

[0017] The monitoring unit B is configured to collect magnetic field signals emitted by the fetal heart at multiple time series, and further analyze them into multiple Class B indicators related to fetal heart rate and fetal movement, forming a fetal response vector F;

[0018] The coupling analysis module is configured to concatenate the uterine contraction behavior vector extracted within each window with the fetal response vector to form a joint state vector Z; and evaluate the specified time window T. k The joint state vector Z (k) The degree of deviation from normal uterine contraction-fetal response is used to output a risk score R. (k) .

[0019] Preferably, the monitoring unit A includes:

[0020] Multiple Class A sensors; the Class A sensors are contact electrodes, which are set at multiple preset locations in the abdominal region of the target individual, and are used to collect uterine electromyography signals in a spatially distributed manner.

[0021] An electrical signal acquisition channel, connected to each of the Class A sensors, is used to perform preprocessing operations on the acquired electromyographic signals, including at least amplification, filtering, and analog-to-digital conversion, to output standardized time-series signal data.

[0022] The A-type processing unit, connected to the electrical signal acquisition channel, includes multiple analysis sub-units for analyzing the uterine activity characteristics of time-based uterine electromyography signals.

[0023] Preferably, the A-type calculation unit analyzes uterine activity characteristics and calculates multiple A-type indicators reflecting uterine activity, including at least: the onset time, location, range of influence, direction and speed of propagation of uterine contractions in the abdomen; and also includes the average uterine contraction amplitude expressed by the average amplitude of the uterine electromyography envelope based on time sequence, the spatial location and variance of the contraction peak, the degree of contraction diffusion, and the intensity and frequency distribution in the abdominal region; and further includes analyzing the uterine contraction intensity pattern based on time sequence, the propagation distance of the peak power, the change of power over time, and the change of frequency over time.

[0024] Preferably, monitoring unit B includes:

[0025] At least one Class B sensor, said Class B sensor being a magnetometer, is configured to acquire magnetic field signals generated by fetal cardiac activity in a low-frequency range; said Class B sensor is disposed on the abdominal surface of the target individual and close to the estimated location of the fetal heart.

[0026] A reference locator, disposed around the Class B sensor, is configured to generate multiple controllable magnetic field signals to provide the Class B sensor with reference positioning information on the relative orientation of the fetus, thereby enhancing the signal source localization and processing capabilities.

[0027] The control module controls multiple reference positioners to generate specified magnetic field signals;

[0028] The magnetic signal processing module, connected to the Class B sensor, is configured to perform bandpass filtering, notch filtering, noise suppression, and signal separation on the acquired raw magnetic field signal, extract multiple Class B indicators, and use the Class B indicators to construct a fetal response vector.

[0029] Preferably, the Class B indicators include at least: fetal heart rate, R-peak-R-peak interval sequence, and further analyze and classify: fetal heart rate variability, heart rate acceleration / deceleration events, beat-by-beat analysis parameters, and based on the position analysis of the fetal heart and the spatial model based on magnetic field calculation, estimate the fetal position and analyze the trend of fetal posture changes.

[0030] Preferably, the Class A sensor is installed by an adhesive method, directly contacting and closely adhering to the abdominal skin of the target individual.

[0031] Preferably, after the Class B sensor and the reference locator are fixed in relative position by the auxiliary component, the target individual wears the auxiliary component so that the Class B sensor and the reference locator fit as closely as possible to the target individual's abdominal skin.

[0032] The beneficial effects achieved by this invention are:

[0033] 1. The early warning method of this technical solution simultaneously collects uterine electromyography signals and fetal cardiac magnetic field signals, reflecting uterine contraction behavior and fetal physiological response status, respectively. By constructing a joint state vector of uterine contraction behavior vector and fetal response vector, and assessing its deviation, it effectively overcomes the lag and one-sidedness of traditional single monitoring methods in abnormal judgment, and significantly improves the accuracy of identifying the risk of premature birth or abnormal delivery.

[0034] 2. The early warning method of this technical solution establishes a statistical model based on a normal stimulus-response sample set, calculates the deviation between the actual state and the normal state under each time window, and realizes the quantitative judgment of abnormal coupling relationship.

[0035] 3. The Class A electrode sensor and Class B magnetic intensity sensor used in the early warning system of this technical solution are both non-invasive adhesive devices, suitable for pregnant women to wear and continuously monitor during the medium to long term before delivery. The overall system design fully considers the actual activity needs of pregnant women, allowing them to maintain a certain degree of physical activity during monitoring without affecting the stability of signal acquisition and the continuity of analysis. It is suitable for home monitoring or remote early warning scenarios, and helps to achieve more flexible and safer delivery management.

[0036] 4. The software and hardware components of the early warning system in this technical solution adopt a modular design. The working modules and components of the hardware component, as well as the instructions, parameters and algorithms of the software component, can be easily replaced and / or upgraded in the later stages, thereby reducing the construction and maintenance costs of this system. Attached Figure Description

[0037] The invention will be further understood from the following description taken in conjunction with the accompanying drawings. The components in the drawings are not necessarily drawn to scale, but rather the emphasis is on illustrating the principles of the embodiments. In different views, the same reference numerals designate corresponding parts.

[0038] Reference numerals: 100 - Coupling analysis module; 110 - Monitoring unit A; 210 - Monitoring unit B; 200 - Target individual; 111 - Class A sensor; 113 - Central reference electrode; 115 - Differential signal; 121 - Multi-channel amplifier; 123 - Filter; 125 - Analog-to-digital converter; 127 - Computation unit A; 130 - Coupling analysis module; 211 - Class B sensor; 213 - Reference locator; 215 - Base locator; 217 - Control module; 219 - Magnetic signal processing module; 700 - Computer system; 702 - Bus; 704 - Processor; 706 - Main memory; 708 - Read-only memory; 710 - Storage device; 712 - Display; 714 - Input device; 716 - Cursor control device; 718 - Network device;

[0039] Figure 1This is a flowchart of the steps of the early warning method described in this invention;

[0040] Figure 2 This is a schematic diagram of the early warning system described in this invention;

[0041] Figure 3 This is a schematic diagram of the arrangement of monitoring unit A in an embodiment of the present invention;

[0042] Figure 4 This is a schematic diagram of the data flow of monitoring unit A in an embodiment of the present invention;

[0043] Figure 5 This is a schematic diagram of the arrangement of monitoring unit B in an embodiment of the present invention;

[0044] Figure 6 This is a schematic diagram of the data flow of monitoring unit B in an embodiment of the present invention;

[0045] Figure 7 This is a schematic diagram of the computer system framework used in the coupling analysis module described in this embodiment of the invention. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to its embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention. Other systems, methods, and / or features of this embodiment will become apparent to those skilled in the art after reviewing the following detailed description. All such additional systems, methods, features, and advantages are intended to be included within this specification, within the scope of the invention, and protected by the appended claims. Further features of the disclosed embodiments are described in the following detailed description, and these features will become apparent from the following detailed description.

[0047] In the accompanying drawings of this invention, the same or similar reference numerals correspond to the same or similar components. In the description of this invention, it should be understood that if terms such as "upper," "lower," "left," and "right" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or component referred to must have a specific orientation. Because the invention is constructed and operated in a specific orientation, the terms describing positional relationships in the drawings are for illustrative purposes only and should not be construed as limiting this patent. Those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0048] Example 1: Exemplary, as shown in the appendix Figure 1 As shown, a labor early warning method based on joint analysis of uterine contractions and fetal heart rate is proposed. The early warning method includes the following steps:

[0049] S100: Multiple Class A sensors are installed in the abdomen of the target individual to collect uterine electromyography (EMG) signals of the target individual at multiple time series; time windows are constructed using the collected uterine EMG signals, and multiple Class A index values ​​are extracted in each time window to form a uterine contraction behavior vector U for multiple time windows;

[0050] S200: Simultaneously, one or more Class B sensors are installed on the abdomen of the target individual to collect fetal cardiac magnetic field signals at multiple time series, and further analyze them into multiple Class B indicators related to fetal heart rate and fetal movement to form a fetal response vector F;

[0051] S300: Based on the uterine contraction behavior vector U, perform pattern recognition on the uterine behavior corresponding to each time window and divide it into 4 activity modes, namely: resting period, weak contraction period, frequent non-labor contractions or labor contractions.

[0052] S400: Extract the uterine contraction behavior vector U and the fetal response vector F from the same time window other than the resting period, and construct the joint state vector Z of the two vectors;

[0053] S500: A reference set M of samples constructed based on multiple pre-established normal uterine contraction-fetal response samples. ref ;

[0054] S600: Set a time window T k The joint state vector Z (k) With reference set M ref The mean vector and covariance matrix are compared to calculate the deviation R. (k) ;

[0055] S700: Based on a preset risk threshold, assess the deviation R. (k) The level of risk that the target individual is facing.

[0056] Simultaneously, a labor early warning system based on joint analysis of uterine contractions and fetal heart rate is proposed, wherein the early warning system applies the aforementioned labor early warning method based on joint analysis of uterine contractions and fetal heart rate; as attached. Figure 2 As shown, the early warning system includes:

[0057] Monitoring unit 110 is configured to collect uterine electromyography (EMG) signals from target individuals 200 at multiple time series. Time windows are constructed using the collected EMG signals, and multiple Class A index values ​​are extracted in each time window to form a uterine contraction behavior vector U for multiple time windows.

[0058] Monitoring unit 210 is configured to collect magnetic field signals emitted by the fetal heart at multiple time series, and further analyze them into multiple Class B indicators related to fetal heart rate and fetal movement to form a fetal response vector F;

[0059] The coupling analysis module 100 is configured to concatenate the uterine contraction behavior vector extracted within each window with the fetal response vector to form a joint state vector Z; and evaluate the specified time window T. k The joint state vector Z (k) The degree of deviation from normal uterine contraction-fetal response is used to output a risk score R. (k) .

[0060] Preferably, the monitoring unit A includes:

[0061] Multiple Class A sensors; the Class A sensors are contact electrodes, which are set at multiple preset locations in the abdominal region of the target individual, and are used to collect uterine electromyography signals in a spatially distributed manner.

[0062] An electrical signal acquisition channel, connected to each of the Class A sensors, is used to perform preprocessing operations on the acquired electromyographic signals, including at least amplification, filtering, and analog-to-digital conversion, to output standardized time-series signal data.

[0063] The A-type processing unit, connected to the electrical signal acquisition channel, includes multiple analysis sub-units for analyzing the uterine activity characteristics of time-based uterine electromyography signals.

[0064] Preferably, the A-type calculation unit analyzes uterine activity characteristics and calculates multiple A-type indicators reflecting uterine activity, including at least: the onset time, location, range of influence, direction and speed of propagation of uterine contractions in the abdomen; and also includes the average uterine contraction amplitude expressed by the average amplitude of the uterine electromyography envelope based on time sequence, the spatial location and variance of the contraction peak, the degree of contraction diffusion, and the intensity and frequency distribution in the abdominal region; and further includes analyzing the uterine contraction intensity pattern based on time sequence, the propagation distance of the peak power, the change of power over time, and the change of frequency over time.

[0065] Preferably, monitoring unit B includes:

[0066] At least one Class B sensor, said Class B sensor being a magnetometer, is configured to acquire magnetic field signals generated by fetal cardiac activity in a low-frequency range; said Class B sensor is disposed on the abdominal surface of the target individual and close to the estimated location of the fetal heart.

[0067] A reference locator, disposed around the Class B sensor, is configured to generate multiple controllable magnetic field signals to provide the Class B sensor with reference positioning information on the relative orientation of the fetus, thereby enhancing the signal source localization and processing capabilities.

[0068] The control module controls multiple reference positioners to generate specified magnetic field signals;

[0069] The magnetic signal processing module, connected to the Class B sensor, is configured to perform bandpass filtering, notch filtering, noise suppression, and signal separation on the acquired raw magnetic field signal, extract multiple Class B indicators, and use the Class B indicators to construct a fetal response vector.

[0070] Preferably, the Class B indicators include at least: fetal heart rate, R-peak-R-peak interval sequence, and further analyze and classify: fetal heart rate variability, heart rate acceleration / deceleration events, beat-by-beat analysis parameters, and based on the position analysis of the fetal heart and the spatial model based on magnetic field calculation, estimate the fetal position and analyze the trend of fetal posture changes.

[0071] Preferably, the Class A sensor is installed by an adhesive method, directly contacting and closely adhering to the abdominal skin of the target individual.

[0072] Preferably, after the Class B sensor and the reference locator are fixed in relative position by the auxiliary component, the target individual wears the auxiliary component so that the Class B sensor and the reference locator fit as closely as possible to the target individual's abdominal skin.

[0073] Preferably, in an exemplary embodiment, both the Class A and Class B sensors are sensors that will not harm the health of the fetus and collect data in real time.

[0074] Preferably, the Class A sensor comprises at least four electrodes, and more preferably, six or more, for example, up to 20. Preferably, multiple electrodes are placed on the abdomen of the target individual and secured to a single location on a wearable or adhesive fabric such as an abdominal binder, cloth sleeve, or elastic fiber sleeve, thus fixing all electrodes firmly to the abdomen of the monitored individual while determining the relative positions of each electrode. Preferably, the tight fabric can be made of any lightweight, soft material suitable for medical use to ensure comfort and stability during wear.

[0075] Preferably, to improve electrode adhesion, the abdominal skin is gently rubbed with medical gel before electrode placement to reduce skin resistance.

[0076] The Class A sensor is configured to detect electromyographic signals generated by the uterus, specifically the bioelectrical signals caused by cell membrane depolarization and repolarization during the contraction of uterine smooth muscle cells. Uterine electromyographic signals are conducted to the body surface via tissue fluid, subcutaneous tissue, and skin, and are characterized by low frequency, low amplitude, and non-invasive detectability. When the Class A sensor is in full contact with the skin, it can sense the potential changes released in the subcutaneous uterine region during muscle activity. The collected electrical signals reflect the intensity, synchronicity, and propagation path of the electrical activity of the uterine smooth muscle group.

[0077] To enhance signal stability, the electrodes are preferably made of silver or silver chloride and used in conjunction with a specialized conductive gel to reduce skin-electrode contact resistance and improve the sensitivity and consistency of signal acquisition. (See attached image.) Figure 3 As shown, each of the Class A sensors 111 is connected to the amplifier in a unipolar connection manner. An electrode network is formed by multiple Class A sensors, with the central electrode serving as the central reference electrode 113. The central reference electrode 113 is positioned at the center of the abdomen as a reference point P. r A common-mode suppression lead is provided for the central reference electrode 113. For example, to accommodate individual differences, the electrode positions can be fine-tuned according to the location of the tire pressure gauge or fetal heart rate monitoring device during electrode placement. Preferably, the impedance of each Class A sensor 111 needs to be calibrated relative to the central reference electrode 113, which can be detected using an impedance testing device. If necessary, skin pretreatment can be repeated until the impedance decreases to below 10 kΩ.

[0078] For example, see attached Figure 3 As shown, eight Class A sensors 111 can be used for uterine electromyography (EMG) signal acquisition. The acquired EMG signals are processed by a high-resolution, low-noise multichannel amplifier 121. When configured as an electrode network with eight electrodes, all signals are connected to an eight-channel unipolar amplifier with a wide dynamic range, using a reference electrode as a reference. Preferably, the amplifier's 3dB bandwidth range is set between 0.05Hz and 10Hz to accommodate the low-frequency characteristics of the EMG signals. Subsequently, the acquired EMG signals are converted from analog to digital, preferably using an A / D acquisition card with 16-bit resolution to transmit them to a computer system.

[0079] Preferably, based on the aforementioned 8 channels of raw uterine electromyography (EMG) signals, 15 derived differential signals 115 are generated by performing pairwise differential operations on adjacent electrode channels to cancel out common components in the reference channels, thereby enhancing the identification of local uterine activity. Finally, the aforementioned 15 exemplary differential signals 115, combined with any two channels from the raw signal, constitute a local uterine contraction intensity channel covering 17 abdominal regions. The above is merely an illustrative description of the principle and basic composition of the electrode network. For greater coverage or higher spatial resolution, the number of Class A sensors can be increased and the number of acquisition channels further expanded.

[0080] As attached Figure 4 The diagram illustrates the processing path of uterine electromyography (EMG) signals after acquisition. The EMG signals acquired by the Class A sensor 111 are weak bioelectrical signals with extremely low amplitude. Therefore, the raw signal is first transmitted to a multi-channel amplifier 121 for amplitude enhancement, and then preprocessed by a filter 123 to effectively remove interference signals from non-target frequency bands. Since uterine EMG signals are mainly concentrated in the low-frequency range of 0.1Hz to 3Hz, preferably, the amplifier and sampling system used have corresponding low-frequency response capabilities to ensure signal integrity and timing stability.

[0081] Furthermore, filtering is necessary because the raw signals acquired by the sensor may contain non-target signals such as maternal electrocardiogram (ECG) signals, fetal ECG signals, electromyography (EMG) signals from other parts of the body, and environmental noise. Filtering can distinguish between multiple source components in the raw signal. In a preferred embodiment, bandpass filtering can be configured individually for each sensor channel to effectively extract signals within the target frequency range on the computing system.

[0082] Furthermore, the filtered signal enters the analog-to-digital converter 125 and is converted into a high-resolution digital signal; the digital signal is then input to the computation unit A 127. The computation unit A 127 is equipped with multiple sub-units for analyzing uterine activity characteristics, which are used to extract and analyze parameters from the pre-processed uterine electromyography signal, thereby identifying its key features in the time and spatial domains, providing a data foundation for subsequent contraction pattern recognition and labor early warning.

[0083] Furthermore, in an exemplary embodiment, the computation unit 127 may include the following analysis subunit:

[0084] First analysis subunit: configured to calculate the average contraction amplitude or frequency per unit time;

[0085] The second analysis subunit is configured to extract the peak intensity location and spatial variance of uterine contractions.

[0086] The third analysis subunit is configured to construct a contraction intensity distribution map based on spatiotemporal data to identify regular and abnormal contraction patterns.

[0087] In an exemplary embodiment, the Class B sensor is a magnetic field sensor. There is one or more Class B sensors, which are positioned within the electrode network composed of multiple Class A sensors, with the distance between the Class B sensor and its nearest neighbors being approximately equal. Preferably, the magnetic field sensor is a fluxgate magnetometer or giant spin resonance-based sensor, allowing for magnetic field measurements to be performed under normal temperature and unshielded conditions.

[0088] Preferably, the Class B sensor has a dynamic measurement range of 0.01 pT to 1 mT and a noise floor of no more than 50 fT / (Hz). 0.5 The frequency response range is 0.1Hz to 200Hz. Each magnetic sensor supports triaxial magnetic field measurement and can communicate with the electronic module via SPI or I2C bus.

[0089] In an exemplary embodiment, as shown in the appendix Figure 5 and attached Figure 6 As shown, the monitoring unit B also includes a reference locator. Multiple reference locators are used in conjunction with the Class B sensor to construct a spatial magnetic field coordinate system for the target individual's abdomen. Combined with the magnetic field inversion modeling method, the spatial positioning accuracy of the magnetic sensor and the accuracy of the fetal heart magnetic field source positioning are improved.

[0090] Specifically, the reference locator 213 is preferably a small AC magnetic coil, so that each reference locator can generate an alternating magnetic field source when an alternating current passes through it. During actual detection, the control module 217 provides power and sets control commands to control the direction and frequency of the current input to the reference locator, causing the reference locator to generate a periodically or continuously changing magnetic field signal with identifiable characteristics. This magnetic field signal can be captured by the Class B sensor 211 and further analyzed and verified by the magnetic signal processing module 219.

[0091] Furthermore, for ease of explanation, the monitoring principle of monitoring unit B is illustrated using the example of five reference locators. In actual implementations, the reference locators may have 8, 10, 12, or more data points. The implementation described herein is merely an example and not intended to be restrictive.

[0092] For example, see attached Figure 6As shown, four of the five reference locators 213 are evenly distributed around the pregnant woman's abdomen, forming a nearly coplanar distribution structure to define the X and Y axes in a two-dimensional plane. The fifth reference locator, serving as the baseline locator 215, is preferably positioned along the midline of the abdomen, specifically at the navel or below the sternum. Its function is to assist in determining the vertical direction (Z-axis), thereby forming a complete three-dimensional coordinate reference frame with the aforementioned coplanar beacons. In some embodiments, the reference locators can also be placed within clothing or abdominal binders worn by the target individual, such as the tightly woven fabric used to secure the Class A sensor, to ensure that the baseline locator remains unchanged during monitoring, thus providing a static reference for sensor calculation and forming a generally stable structure with the Class B sensor.

[0093] Furthermore, the signal transmission method of the reference locators can be flexible and variable in this technical solution, but it is necessary to ensure that the multiple reference locators are clearly distributed in space and that their signals do not interfere with each other. In a preferred embodiment, each reference locator transmits a continuous sinusoidal magnetic signal using a different carrier frequency. For example, the first reference locator uses frequency f1, the second reference locator uses frequency f2, and so on. By staggering the frequencies of the reference locators, the signals of the multiple reference locators can be separated in the frequency domain and are more easily identified.

[0094] In some alternative implementations, the reference locators can employ a time-division multiplexing approach, where each reference locator is activated sequentially according to a system-defined time slice. Only one reference locator is allowed to transmit at a time, while the others remain off. The system synchronously extracts the responses generated by each reference locator according to timestamps. In some implementations, the reference locators can employ spread spectrum modulation, where each beacon is loaded with an independent pseudo-random code to transmit a spread spectrum signal, achieving a satellite signal composition architecture similar to that of the Global Positioning System (GPS). This approach can still achieve highly recognizable signal matching in complex magnetic environments, contributing to improved system stability under dynamic monitoring conditions.

[0095] The positional relationship of the Class B sensor relative to any one of the reference locators with a known magnetic field strength can be approximated using a magnetic dipole model. That is, assuming the magnetic field generated by a reference locator can be considered as being excited by an ideal magnetic dipole, then at a distance of [missing information] from that magnetic dipole... At a given point in space, the magnetic field distribution satisfies the following calculation formula.

[0096]

[0097] In the above formula, δ0 is the free permeability, and p represents the equivalent magnetic dipole moment of a reference positioner. This represents the unit vector pointing from the reference locator position to the Class B sensor, where 's' is the distance between the reference locator and the Class B sensor. The position of the Class B sensor can be calculated using a three-dimensional coordinate reference frame, and the magnetic field value received by the Class B sensor... Since the values ​​are measurable, the position and orientation of the reference locator can be estimated using an inverse equation iterative method. By performing multiple measurements and solving the inverse equations, the 3D positioning of each reference locator within the 3D coordinate reference frame can be determined.

[0098] Furthermore, regarding the human heart, including that of a pregnant woman and her fetus, the electrical activity generated by the depolarization and repolarization of the myocardium during each heartbeat excites a corresponding time-varying magnetic field in the surrounding space. This magnetic field can be sensed and detected by the aforementioned Class B sensor placed on the pregnant woman's abdomen. Although the disturbance caused by the formation and changes in the magnetic field due to the heartbeat is weak, it can still be detected and further analyzed within the high-resolution three-dimensional coordinate reference frame constituted by the aforementioned Class B sensor and reference locator.

[0099] Specifically, the signals collected by Class B sensors are mixed magnetic field signals, including the magnetic fields of the fetal and maternal heartbeats (which have different frequencies), environmental power frequency noise (50 / 60Hz), motion artifacts (such as the pregnant woman's movement), and instrument noise.

[0100] The raw signal is first subjected to primary filtering via a bandpass filter and a notch filter. Preferably, the bandpass filter uses a model with a bandpass frequency range of 0.1–50 Hz; the notch filter uses a model with a center frequency of 50 / 60 Hz or its harmonics. Subsequently, the attitude and position changes of the magnetometer are corrected using data from the inertial measurement unit to further suppress interference caused by motion.

[0101] Furthermore, the monitoring unit B uses denoising and signal separation algorithms to separate the maternal heartbeat component and the fetal heartbeat component in the magnetic heartbeat signal. Relevant algorithms can include independent component analysis, principal component analysis, or source separation networks based on deep learning. Regarding the signal separation algorithm, since the fetal heartbeat frequency is significantly higher than the maternal heartbeat frequency, the two have good distinguishability in the frequency domain, providing a natural criterion for the separation algorithm. The fetal heartbeat magnetic signal after signal separation is the magnetic fluctuation induced by fetal myocardial depolarization with each beat. Its typical waveform includes the magnetic response of the R-peak, which can be further used for heart rate tracking and beat-by-beat analysis.

[0102] Once the fetal magnetic heart signal is separated, monitoring unit B can perform real-time tracking of fetal heartbeat parameters. Preferably, a phase-locked loop or Kalman filter structure can be used to predict the next R-peak occurrence time based on historical R-peak intervals or heart rate change trends, and window detection is performed in the magnetic signal to complete the process of heartbeat position analysis, heart rate analysis, and then heart rate variability analysis.

[0103] Furthermore, when conditions permit, such as when the target individual is stationary and calm, monitoring unit B can further estimate the spatial position and orientation of the fetal heart using the calculation method described in Equation 1 above. Essentially, after multiple magnetometers receive magnetic field disturbances generated by the fetal heart, they can calculate the source point position based on the magnetic dipole model. Based on the above settings, monitoring unit B can obtain several parameters related to the fetus, including: fetal heart rate, R-peak-R-peak interval sequence, and can further analyze and classify: fetal heart rate variability, heart rate acceleration / deceleration events, beat-by-beat analysis parameters, and based on the positional analysis of the fetal heart, can obtain information on the relative posture or orientation trend of the fetus and fetal position estimation based on the magnetic field model.

[0104] In the implementation of this technical solution, the Class A sensor adopts an electrode structure attached to the surface of the abdomen. It is mainly used to collect uterine electromyography signals and works based on detecting the potential difference of the skin surface. It does not have the ability to directly sense changes in magnetic fields.

[0105] When monitoring unit B is further introduced into the system, the low-frequency magnetic field signals emitted by the multiple reference locators will not interfere with the Class A sensors. This is because magnetic signals and electrical signals sensed by electrodes have fundamental differences in physical properties and conduction mechanisms. Magnetic field signals need to generate an induced current in a conductor before they can form a weak potential difference on the body surface. However, given the low conductivity of human tissue and the fact that the emitted magnetic field strength is at a safe low-frequency level, it is unlikely to generate interference signals that can be detected by electrodes under normal clinical use conditions.

[0106] Therefore, in the application of this technical solution, even when using a beacon device for magnetic sensing detection and an electrode network for electrical signal acquisition at the same time, stable coexistence with electromagnetic separation can be achieved. Different types of sensors in the system do not interfere with each other, ensuring accurate and synchronous acquisition of uterine contraction signals and fetal magnetophysiological signals.

[0107] Example 2: This example should be understood as including at least all the features of any of the foregoing examples, and further improving upon them.

[0108] In a further preferred embodiment, the physiological parameters of the target individual and the fetus calculated by the monitoring unit A and the monitoring unit B are input into the coupling analysis module. After the coupling analysis module summarizes the multiple parameters, it combines time series analysis and risk grading to output the preterm birth risk level of the target individual at different time series.

[0109] Preferably, the coupling analysis module can be a general-purpose computing device or a dedicated computing device. Data communication between the coupling analysis module and monitoring units A and B can be established via direct cable connection or wireless communication. Preferably, the wireless communication method includes, but is not limited to, Wi-Fi, Bluetooth, or ZigBee protocols.

[0110] The coupling analysis module statistically analyzes each time window T. k One or more of the following indicators are listed in (k = 1, 2, ..., K):

[0111] For monitoring unit A, each of its output indicators is designated as category A indicator X, and the i-th indicator is X. i On the T k The index value at time X i-k For example, Category A indicators may include: contraction frequency, contraction duration,

[0112] The amplitude or RMS energy of uterine contractions can be analyzed. Furthermore, by analyzing the location of uterine contractions, the following can be output: the rate of uterine contraction propagation, the coordinates of the pacemaker location, the local distribution of uterine electromyographic energy, and the main propagation direction of uterine contractions.

[0113] For monitoring unit B, each of its output indicators is designated as category B indicator Y, where the i-th indicator is...

[0114] Y i On the T k The index value at time Y i (k) For example, Class B indicators may include: fetal heart rate, R-peak-R-peak interval sequence, heart rate variability indicators (such as RMSSD, SDNN), and the number of accelerometer events N. acc Number of heart rate deceleration events N dec Dynamic indicators such as stamina and stamina.

[0115] Further, the following steps are performed to calculate the preterm birth risk index:

[0116] E100: Constructing a temporal model of uterine contraction behavior: Extracting uterine electromyography signals in segments and constructing a time window T. k Extracting category A indicators from each window to form the uterine contraction behavior vector U of the k-window. (k) :

[0117]

[0118] By setting rules, models, or training data, relevant technical personnel identify the uterine contraction behavior vector U for each time window. (k) The identified uterine contractions were categorized into one of the following activity patterns:

[0119] (1) Resting period;

[0120] (2) Weak contraction period;

[0121] (3) Frequent non-labor contractions;

[0122] (4) Labor contractions.

[0123] E200: Synchronously, during each contraction behavior window T except for (1) the resting period k Within the system, the aligned responses of class B indicators under the same sequence of the synchronous monitoring units are used to construct the fetal response vector F. (k) :

[0124]

[0125] E300: Combined with uterine contraction behavior vector U (k) and the fetal response vector F (k) Calculate the coupling strength and deviation of uterine contraction-fetal response, and calculate the single-window risk score R. (k) For example, the calculation steps include:

[0126] E310: Transform the uterine contraction behavior vector U (k) and the fetal response vector F (k) Construct the joint state vector Z (k) ,Right now:

[0127] Z (k) =[U (k) ,F (k) ]∈R m+n ;

[0128] E320: Collect historical sample data of uterine contraction behavior in conjunction with fetal heartbeat behavior, extract multiple "normal uterine contraction + normal fetal response" window samples from normal behavioral monitoring data, and construct a reference set M for the normal coupling pattern sample set. ref ,Right now:

[0129]

[0130] Z ref (i) This indicates that in the normal coupling mode sample set M refThe i-th normal sample data; among which each Z ref (i) All include the uterine contraction behavior vector U ref (i) and the fetal response vector F ref (i) ,Right now:

[0131]

[0132] E330: According to reference set M ref Calculate the mean vector μ ref The covariance matrix Σ ref .

[0133] E340: Transfer the joint state vector Z of the current window. (k) The coupling deviation R between the two is calculated by comparing them with a reference distribution. (k) :

[0134]

[0135] E350: Then, relevant technical personnel set risk classification thresholds τ1, τ2, etc., and correlate them with R. (k) The comparison is performed to determine the combined state of uterine contractions within time window k; an example could be:

[0136] R (k) <τ1, judged as low risk;

[0137] τ1≤R (k) <τ2, classified as medium risk;

[0138] R (k) If the value is ≥τ2, it is considered high risk.

[0139] Example 3: This example should be understood as including at least all the features of any of the foregoing examples, and further improving upon them.

[0140] For example, both monitoring unit A and monitoring unit B are equipped with a computing device capable of performing data processing. The computing device may be equipped with a microprocessor, a digital signal processor (DSP), or other microprocessor, as well as a memory for storing relevant programs and instructions.

[0141] Furthermore, this technical solution can be implemented in various situations. Monitoring unit A and monitoring unit B can be directly or remotely connected to the coupling analysis module 100. For example, this technical solution can operate in homes, hospitals, private residences, mobile facilities (e.g., ambulances or mobile medical vehicles), ships at sea, and remote areas where healthcare is not directly accessible.

[0142] Further details are attached. Figure 7 The following diagram illustrates the implementation of the computer system 700 used in the coupling analysis module 100; the computer system 700 can be applied to the data storage, calculation, and result output processes of each working module in the identification and judgment system.

[0143] For example, computer system 700 includes bus 702 or other communication mechanism for transmitting information, and one or more processors 704 coupled to bus 702 for processing information; processor 704 may be, for example, one or more general-purpose microprocessors.

[0144] The computer system 700 also includes a main memory 706, such as random access memory (RAM), cache and / or other dynamic storage devices, coupled to a bus 702 for storing information and instructions to be executed by the processor 704; the main memory 706 may also be used to store temporary variables or other intermediate information during the execution of instructions executed by the processor 704; when these instructions are stored in a storage medium accessible to the processor 704, the computer system 700 presents itself as a dedicated machine customized to perform the operations specified in the instructions;

[0145] The computer system 700 may also include a read-only memory (ROM) 708 or other static storage device coupled to the bus 702 for storing static information and instructions of the processor 704; wherein a storage device 710, such as a disk, optical disk or USB drive (flash drive), is coupled to the bus 702 for storing information and instructions.

[0146] Furthermore, the bus 702 may also include a display 712 for displaying various information, data, media, etc., and an input device 714 for allowing users of the computer system 700 to control, manipulate, and / or interact with the computer system 700.

[0147] A preferred method of interacting with the management system may be through a cursor control device 716, such as a computer mouse or a similar control / navigation mechanism;

[0148] Furthermore, the computer system 700 may also include a network device 718 coupled to the bus 702; wherein the network device 718 may include components such as wired network cards, wireless network cards, switching chips, routers, switches, etc.

[0149] Generally speaking, the terms “engine,” “component,” “system,” and “database” used in this article can refer to the logic embodied in hardware or firmware, or to a set of software instructions that may have entries and exit points, written in programming languages ​​such as Java, C, or C++; software components can be compiled and linked into executable programs and installed in dynamic link libraries, or can be written in interpreted programming languages ​​(such as BASIC, Perl, or Python); it should be understood that software components can be called from other components or from themselves, and / or can be called in response to detected events or interrupts;

[0150] Software components configured to execute on a computing device may be provided on a computer-readable medium, such as an optical disc, digital video disc, flash drive, magnetic disk, or any other tangible medium, or as a digital download (and may be initially stored in a compressed or installable format, requiring installation, decompression, or decryption prior to execution); such software code may be stored, in part or in whole, on a memory device executing the computing device; software instructions may be embedded in firmware, such as an EPROM; it should also be understood that hardware components may consist of connected logic units (e.g., gates and flip-flops), and / or may consist of programmable units (e.g., programmable gate arrays or processors);

[0151] The computer system 700 includes technologies described herein that can be implemented using custom hardwired logic, one or more ASICs or FPGAs, firmware and / or program logic, which, when combined with the computer system, enables the computer system 700 to become a dedicated computing device.

[0152] According to one or more embodiments, the techniques described herein are executed by a computer system 700 in response to a processor 704 executing one or more sequences of one or more instructions contained in main memory 706; such instructions may be read into main memory 706 from another storage medium such as storage device 710; execution of the sequence of instructions contained in main memory 706 causes processor 704 to perform the processing steps described herein; in alternative embodiments, hardwired circuitry may be used in place of or in combination with software instructions.

[0153] As used herein, the term "non-transitory medium" and similar terms refer to any medium that stores data and / or instructions that enable a machine to operate in a particular manner; such non-transitory medium may include non-volatile medium and / or volatile medium; non-volatile medium includes, for example, optical discs or magnetic disks, such as storage device 710; volatile medium includes dynamic memory, such as main memory 706.

[0154] Common forms of non-transitory media include, for example, floppy disks, hard disks, solid-state drives, magnetic tapes or any other magnetic data storage media, CD-ROMs, any other optical data storage media, any physical media with a hole pattern, RAM, PROM and EPROM, FLASH-EPROM, NVRAM, any other memory chips or cartridges and their network versions.

[0155] Non-transient media are different from transmission media, but can be used in conjunction with transmission media; transmission media participate in information transmission between non-transient media; for example, transmission media include coaxial cables, copper wires and optical fibers, including the wires that constitute bus 702; transmission media can also take the form of sound waves or light waves, such as radio waves and infrared data communication.

[0156] While this application has been described above with reference to various embodiments, it should be understood that many changes and modifications can be made without departing from the scope of this application. That is, the methods, systems, and devices discussed above are examples. Various configurations can be appropriately omitted, substituted, or added to various processes or components. For example, in alternative configurations, methods can be performed in a different order than those described, and / or various components can be added, omitted, and / or combined. Moreover, features described with respect to certain configurations can be combined in various other configurations, such as different aspects and elements of the configuration can be combined in a similar manner. Furthermore, the elements therein can be updated as the technology develops; that is, many elements are examples and do not limit the scope of this disclosure or the claims.

[0157] Specific details are provided in the specification to offer a thorough understanding of exemplary configurations, including implementations. However, configurations can be practiced without these specific details; for example, well-known circuits, processes, algorithms, structures, and techniques have been shown without unnecessary detail to avoid obscuring the configuration. This description provides only exemplary configurations and does not limit the scope, applicability, or configuration of the claims. Rather, the foregoing description of the configurations will provide those skilled in the art with an enabling description for implementing the described techniques. Various changes can be made to the function and arrangement of the elements without departing from the spirit or scope of this disclosure.

[0158] In summary, the above detailed description is intended to be illustrative rather than restrictive, and it should be understood that these embodiments are for illustrative purposes only and not for limiting the scope of protection of the invention. After reading the description of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent changes and modifications also fall within the scope defined by the claims of this invention.

Claims

1. A method for early warning of labor based on joint analysis of uterine contractions and fetal heart rate, characterized in that, The early warning method includes the following steps: S100: Multiple Class A sensors are installed in the abdomen of the target individual to collect uterine electromyography (EMG) signals of the target individual at multiple time series; time windows are constructed using the collected uterine EMG signals, and multiple Class A index values ​​are extracted in each time window to form a uterine contraction behavior vector U for multiple time windows; S200: Simultaneously, one or more Class B sensors are installed on the abdomen of the target individual to collect fetal cardiac magnetic field signals at multiple time series, and further analyze them into multiple Class B indicators related to fetal heart rate and fetal movement to form a fetal response vector F; S300: Based on the uterine contraction behavior vector U, perform pattern recognition on the uterine behavior corresponding to each time window and divide it into 4 activity modes, namely: resting period, weak contraction period, frequent non-labor contractions or labor contractions. S400: Extract the uterine contraction behavior vector U and the fetal response vector F from the same time window other than the resting period, and construct the joint state vector Z of the two vectors; S500: A reference set M of samples constructed based on multiple pre-established normal uterine contraction-fetal response samples. ref ; S600: Set a time window T k The joint state vector Z (k) With reference set M ref The mean vector and covariance matrix are compared to calculate the deviation R. (k) ; S700: Based on a preset risk threshold, assess the deviation R. (k) The level of risk that the target individual is facing.

2. A labor early warning system based on combined analysis of uterine contractions and fetal heart rate, characterized in that, The early warning system applies the labor early warning method based on combined analysis of uterine contractions and fetal heart rate as described in claim 1; the early warning system includes: The monitoring unit A is configured to collect uterine electromyography (EMG) signals of the target individual at multiple time series. Time windows are constructed using the collected EMG signals, and multiple Class A index values ​​are extracted in each time window to form a uterine contraction behavior vector U for multiple time windows. The monitoring unit B is configured to collect magnetic field signals emitted by the fetal heart at multiple time series, and further analyze them into multiple Class B indicators related to fetal heart rate and fetal movement, forming a fetal response vector F; The coupling analysis module is configured to concatenate the uterine contraction behavior vector extracted within each window with the fetal response vector to form a joint state vector Z; and evaluate the specified time window T. k The joint state vector Z (k) The degree of deviation from normal uterine contraction-fetal response is used to output a risk score R. (k) .

3. The early warning system as described in claim 2, characterized in that, The monitoring unit A includes: Multiple Class A sensors; the Class A sensors are contact electrodes, which are set at multiple preset locations in the abdominal region of the target individual, and are used to collect uterine electromyography signals in a spatially distributed manner. An electrical signal acquisition channel, connected to each of the Class A sensors, is used to perform preprocessing operations on the acquired electromyographic signals, including at least amplification, filtering, and analog-to-digital conversion, to output standardized time-series signal data. The A-type processing unit, connected to the electrical signal acquisition channel, includes multiple analysis sub-units for analyzing the uterine activity characteristics of time-based uterine electromyography signals.

4. The early warning system as described in claim 2, characterized in that, The aforementioned computational unit analyzes uterine activity characteristics and calculates multiple Class A indicators reflecting uterine activity, including at least: the onset time, location, extent of influence, direction and speed of propagation of uterine contractions in the abdomen; and also includes the average uterine contraction amplitude expressed by the average amplitude of the uterine electromyography envelope based on time sequence, the spatial location and variance of the contraction peak, the degree of contraction diffusion, and the intensity and frequency distribution in the abdominal region; and further includes analyzing the uterine contraction intensity pattern based on time sequence, the propagation distance of peak power, the change of power over time, and the change of frequency over time.

5. The early warning system as described in claim 2, characterized in that, Monitoring Unit B includes: At least one Class B sensor, said Class B sensor being a magnetometer, is configured to acquire magnetic field signals generated by fetal cardiac activity in a low-frequency range; said Class B sensor is disposed on the abdominal surface of the target individual and close to the estimated location of the fetal heart. A reference locator, disposed around the Class B sensor, is configured to generate multiple controllable magnetic field signals to provide the Class B sensor with reference positioning information on the relative orientation of the fetus, thereby enhancing the signal source localization and processing capabilities. The control module controls multiple reference positioners to generate specified magnetic field signals; The magnetic signal processing module, connected to the Class B sensor, is configured to perform bandpass filtering, notch filtering, noise suppression, and signal separation on the acquired raw magnetic field signal, extract multiple Class B indicators, and use the Class B indicators to construct a fetal response vector.

6. The early warning system as described in claim 2, characterized in that, The Class B indicators include at least: fetal heart rate and R-peak-R-peak interval sequence, and are further analyzed and classified into: fetal heart rate variability, heart rate acceleration / deceleration events, and beat-by-beat analysis parameters. Based on the position analysis of the fetal heart and the spatial model based on magnetic field calculation, the fetal position is estimated and the trend of fetal posture change is analyzed.

7. The early warning system as described in claim 2, characterized in that, The Class A sensor is installed by attaching itself to the target individual's abdominal skin.

8. The early warning system as described in claim 2, characterized in that, After the Class B sensor and the reference locator are fixed in relative position by the auxiliary component, the target individual wears the auxiliary component so that the Class B sensor and the reference locator fit as closely as possible to the target individual's abdominal skin.

Images