Non-invasive continuous monitoring mattress for physiological indexes of newborns

Abstract

The invention discloses a non-invasive continuous monitoring mattress for physiological indexes of newborns, and relates to the technical field of intelligent mattresses. A memory foam layer is arranged on one side of the spring supporting layer, a fabric layer is arranged on the other side of the memory foam layer, the memory foam layer comprises a first sponge layer and a second sponge layer, and a flexible monitoring layer is arranged between the first sponge layer and the second sponge layer. Through a multi-sensor fusion architecture and by combining the advantages of the pressure sensor, the fiber grating sensor and the temperature sensor, cross verification and complementary monitoring of different vital signs are realized, the accuracy and reliability of monitoring data are improved, and meanwhile, through a layered signal processing algorithm, the accuracy and reliability of the monitoring data are improved. A complete processing chain is formed from original signal preprocessing, feature extraction and data fusion, the problem of mixed signal separation is effectively solved, and stable monitoring performance can still be kept under the interference conditions of neonatal turnover, movement and the like.

Description

Technical Field

[0001] This invention belongs to the field of smart mattress technology, and in particular relates to a non-invasive continuous monitoring mattress for newborn physiological indicators. Background Technology

[0002] Newborns (especially premature and low birth weight infants) have immature organ development and weak metabolic regulation ability, and their physiological indicators such as heart rate, respiration and blood oxygen saturation fluctuate significantly. They need to be continuously and dynamically monitored to prevent critical events such as asphyxia and apnea. Traditional monitoring methods have two major problems: (1) Limitations of invasive monitoring: ECG monitoring requires electrode pads, which can easily cause skin allergies or pressure sores (occurrence rate of about 15%); blood oxygen monitoring requires wire connection, and as the newborn’s body and limbs move, the wires can easily wrap around the body or limbs, causing injury; frequent operation interferes with the newborn’s sleep and affects their growth and development. (2) Bottlenecks of non-invasive technology: Existing optical sensors (such as PPG) are easily affected by skin color and motion artifacts, with an error rate as high as 20% in newborns; capacitive sensors have high requirements for skin fit, and the many skin folds of newborns lead to signal distortion; multi-parameter integrated monitoring equipment is large in size and cannot be seamlessly integrated with newborn care mattresses.

[0003] To address the aforementioned issues, this project proposes a non-invasive monitoring solution based on "flexible sensing, dynamic compensation, and multimodal fusion": 1. Flexible sensing layer: Biomimetic mechanical adaptation to newborn skin: Ultra-thin electrode design: Utilizing a nano-silver wire conductive film (thickness <50μm) with an elastic modulus of 0.2MPa, perfectly matching the newborn's skin and avoiding pressure-induced damage; Multimodal sensor integration: Optical sensor: Employing dual-wavelength LEDs (660nm / 940nm) combined with an adaptive filtering algorithm to reduce motion artifact interference; Capacitive sensor: Improving fit through a microstructured conductive layer (such as honeycomb polyurethane), increasing signal acquisition rate by 30%; Piezoelectric film sensor: Real-time monitoring of respiratory rate with an accuracy of ±2 breaths / minute. 2. Dynamic Compensation Algorithm: Addressing Neonatal Physiological Characteristics; Motion Artifact Suppression: Constructing a neonatal motion pattern library based on deep learning, separating effective signals from noise through time-frequency domain analysis; Skin Color Adaptive Calibration: Establishing a database of neonatal skin optical parameters, dynamically adjusting light source intensity and detection threshold; Multi-Parameter Cross-Validation: Achieving early warning of abnormal events (e.g., 10-second warning of apnea) through the physiological correlation between heart rate, respiration, and blood oxygenation. 3. Smart Mattress System: Deeply Integrated with Nursing Scenarios; Modular Design: Removable and replaceable sensor units adaptable to different sizes of neonatal beds; Wireless Data Transmission: Real-time push of physiological parameters to the nurse station via Bluetooth 5.0, supporting simultaneous display on multiple terminals; Pressure Distribution Visualization: Built-in 16×16 pressure sensor array, displaying neonatal position and pressure areas in real time, preventing pressure sores. Summary of the Invention

[0004] The purpose of this invention is to provide a non-invasive continuous monitoring mattress for neonatal physiological indicators. Through the combination of a multi-sensor fusion architecture and a hierarchical signal processing algorithm, it solves the problems of low monitoring accuracy, poor anti-interference ability, and insufficient adaptability in the prior art.

[0005] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution.

[0006] This invention relates to a non-invasive continuous monitoring mattress for neonatal physiological indicators, comprising:

[0007] A spring support layer is provided with a memory foam layer on one side and a fabric layer on the other side. The memory foam layer includes a first sponge layer and a second sponge layer. A flexible monitoring layer is provided between the first sponge layer and the second sponge layer. A monitoring unit is provided on the surface of the flexible monitoring layer. The monitoring unit includes a pressure sensor, a fiber optic grating sensor, and a temperature sensor.

[0008] The data processing and analysis unit includes a signal preprocessing module, a core separation and extraction module, and a feature detection and post-processing module.

[0009] The wireless transmission module transmits data with the remote monitoring center via a wireless communication protocol.

[0010] The invention is further configured such that the pressure sensor is made of cellulose composite piezoresistive material and is distributed in a 16×16 matrix in the middle of the flexible monitoring layer to capture pressure fluctuations in the chest and abdominal cavities caused by respiration and micro-pressure changes caused by heartbeat. The fiber optic grating sensor includes a grid-shaped optical fiber arranged along the flexible monitoring layer, which detects mattress deformation caused by body movement through Bragg wavelength shift, and helps to distinguish body movement events such as turning over and convulsions. The temperature sensor is integrated into the gap of the pressure sensor and uses a PDMS-encapsulated platinum resistance thermometer to compensate for the influence of environmental temperature drift on the pressure signal, and at the same time assists in assessing body temperature changes, and optimizes health assessment by combining other data.

[0011] The present invention is further configured such that the signal preprocessing module includes a wavelet transform module, a moving average filtering module, and a dynamic baseline correction module. The wavelet transform module eliminates high-frequency noise and low-frequency drift, and retains physiological signal characteristics of 0.1-100Hz. The moving average filtering module performs local averaging on the raw data of the pressure sensor array to reduce random noise interference. The dynamic baseline correction module establishes a baseline based on data from non-sleep periods and dynamically adjusts the signal reference during sleep.

[0012] The present invention is further configured such that the core separation and extraction module includes a Fourier transform module and a wavelet packet decomposition module. The Fourier transform module converts the pressure signal to the frequency domain and extracts the main frequency components of 0.2-0.5Hz (respiration) and 0.8-2Hz (heart rate). The wavelet packet decomposition module locates the energy proportion of a specific frequency band to identify abnormal spectral characteristics during sleep apnea.

[0013] The present invention is further configured such that the feature detection and post-processing module includes a pressure distribution mapping module, a hidden Markov module, and a weighted average module. The pressure distribution mapping module constructs a pressure heatmap using data from multi-region pressure sensors and fiber Bragg grating sensors, and determines postures such as supine and lateral lying based on geometric features. The hidden Markov module constructs a sleep stage transition probability model based on features such as body movement frequency and respiratory rate to distinguish between wakefulness, light sleep, deep sleep, and REM sleep. The weighted average module improves the accuracy of detection by weighting the pressure sensor, fiber Bragg grating sensor, and temperature sensor according to confidence level.

[0014] The invention is further configured such that the spring support layer adopts an independent pocket spring design, with each spring unit set independently to form an independent partition.

[0015] The invention is further configured such that the remote monitoring center includes a mobile APP and a cloud platform, the mobile APP is used to display sleep parameters in real time and generate sleep quality reports, and the cloud platform is used for long-term data storage and health risk warning.

[0016] The invention is further configured such that the flexible monitoring layer is fixedly connected to one side of the first sponge layer via Velcro.

[0017] The present invention has the following beneficial effects.

[0018] 1. This invention utilizes a multi-sensor fusion architecture, combining the advantages of pressure sensors, fiber Bragg grating sensors, and temperature sensors, to achieve cross-validation and complementary monitoring of different vital signs, thereby improving the accuracy and reliability of monitoring data. Simultaneously, through a hierarchical signal processing algorithm, a complete processing chain is formed from raw signal preprocessing to feature extraction and data fusion, effectively solving the problem of mixed signal separation and maintaining stable monitoring performance even under interference such as newborns turning over or moving.

[0019] 2. This invention, through adaptive baseline correction and personalized modeling, can adapt to the physiological characteristics and movement habits of different newborns, avoiding a one-size-fits-all threshold setting and improving the universality of the system. Through the structural design of the flexible monitoring layer and the independent zoned spring support layer, the mattress comfort is maintained while ensuring monitoring accuracy. The flexible sensor array conforms to the curves of the human body to avoid local pressure, and the independent pocket springs effectively isolate pressure transmission interference, so that the sensor data can more accurately reflect the physiological state of the newborn.

[0020] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below.

[0022] Figure 1 This is a three-dimensional image of a mattress for non-invasive continuous monitoring of neonatal physiological indicators.

[0023] Figure 2 This is a three-dimensional image of the flexible monitoring layer in a non-invasive continuous monitoring mattress for neonatal physiological indicators.

[0024] Figure 3 This is a schematic diagram of the testing process in a non-invasive continuous monitoring mattress for newborn physiological indicators.

[0025] Figure 4 This is a schematic diagram showing the function of the signal preprocessing module in a non-invasive continuous monitoring mattress for neonatal physiological indicators.

[0026] Figure 5 This is a schematic diagram showing the decomposition of the core separation and extraction module function in a non-invasive continuous monitoring mattress for neonatal physiological indicators.

[0027] Figure 6 This is a schematic diagram showing the function of the feature detection and post-processing module in a non-invasive continuous monitoring mattress for neonatal physiological indicators.

[0028] Figure 7 This is a schematic diagram showing the function of the detection unit in a non-invasive continuous monitoring mattress for neonatal physiological indicators.

[0029] In the attached diagram: 1. Spring support layer; 2. Memory foam layer; 21. First sponge layer; 22. Second sponge layer; 3. Flexible monitoring layer; 4. Fabric layer. Detailed Implementation

[0030] The technical solutions of the present invention will be described below with reference to the accompanying drawings. The described embodiments are only some embodiments of the present invention, and not all embodiments.

[0031] Example 1

[0032] Please see Figures 1-7 This invention relates to a non-invasive continuous monitoring mattress for newborn physiological indicators, comprising a spring support layer 1, a memory foam layer 2 on one side of the spring support layer 1, and a fabric layer 4 on the other side of the memory foam layer 2. The memory foam layer 2 includes a first sponge layer 21 and a second sponge layer 22. A flexible monitoring layer 3 is disposed between the first sponge layer 21 and the second sponge layer 22. A monitoring unit is disposed on the surface of the flexible monitoring layer 3, and the monitoring unit includes a pressure sensor, a fiber optic grating sensor, and a temperature sensor. The spring support layer 1 adopts an independent pocket spring design, with each spring unit set independently to form an independent zone. The flexible monitoring layer 3 is fixedly connected to one side of the first sponge layer 21 by Velcro.

[0033] Further details: The spring support layer 1 employs an independent pocket spring design, with each spring unit responding independently to pressure changes, providing a stable mechanical foundation for the upper sensor layer. It also matches the human body's curves, distributing pressure, avoiding localized compression, and improving long-term comfort. The memory foam layer 2 ensures comfort and provides a flat mounting surface for the flexible monitoring layer 3. The flexible monitoring layer 3 is fixed to the first sponge layer 21 via Velcro, facilitating disassembly and maintenance. The fabric layer 4 uses breathable and antibacterial fabric to ensure hygiene during prolonged use. The monitoring unit is integrated into the surface of the flexible monitoring layer 3, including a 16×16 matrix of 256 pressure sensors. It displays the newborn's position and pressure areas in real time, preventing pressure sores. It uses a cellulose composite pressure resistance material, possessing good flexibility and sensitivity, capable of capturing pressure fluctuations in the chest and abdominal cavities caused by respiration (amplitude approximately 0.5-2 Pa) and micro-pressure changes caused by heartbeat (amplitude approximately 0.1-0.5 Pa). The flexible monitoring layer 3 utilizes biomimetic mechanics and... The ultra-thin electrode design utilizes a nano-silver wire conductive film (thickness <50μm) with an elastic modulus of 0.2MPa, perfectly matching the newborn's skin and avoiding pressure damage. The sensor matrix integrates multiple modes: Optical sensors: employing dual-wavelength LEDs (660nm / 940nm) combined with an adaptive filtering algorithm to reduce motion artifact interference; Capacitive sensors: improving fit through a microstructured conductive layer (such as honeycomb polyurethane), increasing signal acquisition rate by 30%; Piezoelectric film sensors: real-time monitoring of respiratory rate with an accuracy of ±2 breaths / minute; a grid-shaped fiber optic grating sensor that penetrates the entire monitoring layer, detecting mattress deformation through Bragg wavelength shift with a sensitivity of 0.1mm; Temperature sensors distributed between pressure sensors, using PDMS-encapsulated platinum resistance thermometers with a temperature measurement accuracy of ±0.1℃, used for environmental temperature drift compensation and body temperature trend monitoring; the sensor units are detachable and replaceable, adaptable to different sizes of newborn beds.

[0034] Example 2

[0035] Please see Figures 1-7 Based on Example 1, the data processing and analysis unit includes a signal preprocessing module, a core separation and extraction module, and a feature detection and post-processing module. The signal preprocessing module includes a wavelet transform module, a moving average filtering module, and a dynamic baseline correction module. The wavelet transform module eliminates high-frequency noise and low-frequency drift, retaining physiological signal characteristics from 0.1-100Hz. The moving average filtering module performs local averaging on the raw data from the pressure sensor array to reduce random noise interference. The dynamic baseline correction module establishes a baseline based on data from non-sleep periods and dynamically adjusts the signal reference during sleep. The core separation and extraction module includes a Fourier transform module and a wavelet packet decomposition module. The Fourier transform module converts the pressure signal to the frequency domain, improving its performance. The system extracts the main frequency components of 0.2-0.5Hz (respiration) and 0.8-2Hz (heart rate). The wavelet packet decomposition module locates the energy proportion of specific frequency bands to identify abnormal spectral features during sleep apnea. The feature detection and post-processing module includes a pressure distribution mapping module, a hidden Markov module, and a weighted average module. The pressure distribution mapping module constructs a pressure heatmap using data from multi-region pressure sensors and fiber optic grating sensors, and combines geometric features to determine postures such as supine and lateral lying. The hidden Markov module constructs a sleep stage transition probability model based on features such as body movement frequency and respiratory frequency to distinguish between awake, light sleep, deep sleep, and REM sleep. The weighted average module improves the accuracy of detection by weighting the pressure sensor, fiber optic grating sensor, and temperature sensor according to confidence level.

[0036] Further details: The data processing and analysis unit adopts a hierarchical processing architecture. The signal preprocessing module first purifies the raw signal. The wavelet transform module uses the db4 wavelet basis to decompose the raw signal into 5 layers, retaining the 0.1-100Hz physiological signal frequency band (respiration 0.1-2Hz, heartbeat 0.8-2Hz, body movement 1-10Hz), filtering out high-frequency noise (>100Hz, such as environmental vibration) and low-frequency drift (<0.1Hz, such as respiratory baseline fluctuations), effectively separating noise and physiological signals. The raw data (sampling rate 100Hz) of the pressure sensor array is locally averaged with a window size of 5 to reduce random noise (such as sensor thermal noise) interference. The dynamic baseline correction module establishes a personalized baseline based on the data of the newborn 30 minutes before falling asleep and automatically updates it every 2 hours during sleep to adapt to the natural fluctuations of physiological parameters.

[0037] The core separation and extraction module employs a strategy combining frequency domain analysis and time-frequency analysis: the Fourier transform module performs spectral analysis on the preprocessed signal to extract the dominant respiratory frequency (0.2-0.5Hz) and dominant heart rate frequency (0.8-2Hz), with a spectral resolution of 0.01Hz; the wavelet packet decomposition module performs three-level wavelet packet decomposition on the deformation signal of the fiber optic grating sensor and the pressure sensor to locate the energy proportion in the 1-10Hz frequency band, identify the type and intensity of body movement signals, and simultaneously focus on monitoring the energy changes in the respiratory frequency band. When the energy remains below 20% of the baseline value for 10 seconds, it is marked as a suspected apnea event.

[0038] The feature detection and post-processing module achieves multi-source information fusion. A pressure heatmap is constructed based on a 16×16 pressure sensor array, generated every minute. The pressure difference between the left and right sides and the pressure ratio between the anterior and posterior zones are extracted as features and input into a CNN model (training set contains 2000 samples). This enables recognition of four sleeping positions: supine, left lateral, right lateral, and prone (accuracy exceeding 95%). The Hidden Markov Model (HMM) module, based on the transition probability model of four sleep stages, integrates body movement frequency, respiratory variability, and heart rate variability features to construct a state transition matrix (states include wakefulness, light sleep N1, light sleep N2, deep sleep N3, and REM). The optimal sleep stage sequence is decoded using the Viterbi algorithm. The weighted averaging module combines pressure sensor data (confidence 0.6) and fiber optic grating sensor data (confidence 0.3). The detection results from the temperature sensor (confidence level 0.1) are fused according to weights to improve the accuracy of abnormal event detection (such as apnea) (false alarm rate <4%). Simultaneously, the weights of each sensor are dynamically adjusted based on signal quality indicators. When the signal quality of a sensor deteriorates, its weight is automatically reduced to ensure the system can still function normally when some sensors fail. Furthermore, dynamic compensation algorithms are employed to address the physiological characteristics of newborns: motion artifact suppression: a newborn motion pattern database is built based on deep learning, and effective signals are separated from noise through time-frequency domain analysis; skin color adaptive calibration: a database of newborn skin optical parameters is established, and the light source intensity and detection threshold are dynamically adjusted; multi-parameter cross-validation: through the physiological correlation between heart rate, respiration, and blood oxygenation, early warning of abnormal events (such as a 10-second warning of apnea) is achieved.

[0039] Example 3

[0040] Please see Figures 1-7 Based on Embodiments 1 and 2, the wireless transmission module transmits data with the remote monitoring center through a wireless communication protocol. The remote monitoring center includes a mobile APP and a cloud platform. The mobile APP is used to display sleep parameters in real time and generate sleep quality reports, while the cloud platform is used for long-term data storage and health risk warnings.

[0041] Furthermore, the wireless transmission module utilizes Bluetooth Low Energy 5.0 technology, transmitting compressed feature data to the remote monitoring center (nurse station) every 5 minutes. The mobile app provides intuitive sleep quality reports, including sleep efficiency, deep sleep percentage, and respiratory event frequency. The cloud platform establishes a newborn parameter database (storing 30 days of historical data and using machine learning models to warn of risks such as sleep apnea syndrome and restless legs syndrome). When the cloud platform detects three consecutive apnea events or a persistently abnormal heart rate, it automatically sends a warning message to a pre-set contact.

[0042] The working principle of this invention is as follows: When a newborn lies on the mattress, the body pressure distribution is transmitted to the flexible monitoring layer 3 through the memory foam layer 2. The pressure sensor matrix captures micro-pressure changes in the chest and abdominal cavity areas, where the low-frequency component (0.2-0.5Hz) corresponds to respiratory movements, and the mid-frequency component (0.8-2Hz) corresponds to heartbeat vibrations. The fiber optic grating sensor monitors the overall deformation of the mattress and identifies large movements such as rolling over and body movements. The temperature sensor simultaneously monitors the ambient temperature to provide temperature drift compensation for the pressure signal.

[0043] In the signal preprocessing stage, wavelet transform removes high-frequency noise and low-frequency drift, moving average filtering smooths random interference, and dynamic baseline correction eliminates the influence of individual differences. In the core separation stage, Fourier transform extracts the frequency domain features of respiration and heart rate, and wavelet packet decomposition identifies abnormal patterns such as apnea. In the post-processing stage, multi-sensor data are fused according to confidence level, pressure distribution mapping identifies sleeping posture, and hidden Markov models analyze sleep structure. Simultaneously, based on dynamic compensation algorithms: motion artifact suppression addresses neonatal physiological characteristics by constructing a neonatal motion pattern database based on deep learning; skin color adaptive calibration establishes a database of neonatal skin optical parameters and dynamically adjusts light source intensity and detection thresholds; and multi-parameter cross-validation utilizes the physiological correlation between heart rate, respiration, and blood oxygenation to achieve early warning of abnormal events (such as a 10-second warning of apnea).

[0044] Finally, the processing results are transmitted to the nurse station via Bluetooth Low Energy 5.0: the mobile APP displays sleep parameters in real time and generates quality reports, the cloud platform establishes a newborn parameter database, analyzes long-term trends through machine learning models, and automatically sends warning information to the preset nurse station when three consecutive episodes of apnea or abnormal heart rate are detected, thus realizing seamless and continuous health monitoring and risk intervention.

[0045] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A non-invasive continuous monitoring mattress for neonatal physiological indicators, characterized in that, include: A spring support layer (1) is provided on one side of the spring support layer (1) and a fabric layer (4) is provided on the other side of the memory foam layer (2). The memory foam layer (2) includes a first sponge layer (21) and a second sponge layer (22). A flexible monitoring layer (3) is provided between the first sponge layer (21) and the second sponge layer (22). A monitoring unit is provided on the surface of the flexible monitoring layer (3). The monitoring unit includes a pressure sensor, a fiber optic grating sensor and a temperature sensor. The data processing and analysis unit includes a signal preprocessing module, a core separation and extraction module, and a feature detection and post-processing module. The wireless transmission module transmits data with the remote monitoring center via a wireless communication protocol.

2. The non-invasive continuous monitoring mattress for neonatal physiological indicators according to claim 1, characterized in that: The pressure sensor is made of cellulose composite piezoresistive material and is distributed in a 16×16 matrix in the middle of the flexible monitoring layer (3). The fiber optic grating sensor includes a grid-shaped optical fiber arranged along the flexible monitoring layer (3). The temperature sensor is integrated into the gap of the pressure sensor and is a PDMS-encapsulated platinum resistance thermometer.

3. The non-invasive continuous monitoring mattress for neonatal physiological indicators according to claim 1, characterized in that: The signal preprocessing module includes a wavelet transform module, a moving average filtering module, and a dynamic baseline correction module. The wavelet transform module eliminates high-frequency noise and low-frequency drift, and retains the physiological signal characteristics of 0.1-100Hz. The moving average filtering module performs local averaging on the raw data of the pressure sensor array to reduce random noise interference. The dynamic baseline correction module establishes a baseline based on data from non-sleep periods and dynamically adjusts the signal reference during sleep.

4. The non-invasive continuous monitoring mattress for neonatal physiological indicators according to claim 1, characterized in that: The core separation and extraction module includes a Fourier transform module and a wavelet packet decomposition module. The Fourier transform module converts the pressure signal to the frequency domain and extracts the main frequency components of 0.2-0.5Hz (respiration) and 0.8-2Hz (heart rate). The wavelet packet decomposition module locates the energy proportion of specific frequency bands to identify abnormal spectral characteristics during sleep apnea.

5. The non-invasive continuous monitoring mattress for neonatal physiological indicators according to claim 1, characterized in that: The feature detection and post-processing module includes a pressure distribution mapping module, a hidden Markov module, and a weighted average module. The pressure distribution mapping module constructs a pressure heatmap using data from multi-region pressure sensors and fiber Bragg grating sensors, and combines geometric features to determine postures such as supine and lateral lying. The hidden Markov module constructs a sleep stage transition probability model based on features such as body movement frequency and respiratory rate to distinguish between wakefulness, light sleep, deep sleep, and REM sleep. The weighted average module improves the accuracy of detection by weighting the pressure sensor, fiber Bragg grating sensor, and temperature sensor according to confidence level.

6. The non-invasive continuous monitoring mattress for neonatal physiological indicators according to claim 1, characterized in that: The spring support layer (1) adopts an independent pocket spring design, with each spring unit set independently to form an independent partition.

7. The non-invasive continuous monitoring mattress for neonatal physiological indicators according to claim 1, characterized in that: The remote monitoring center includes a mobile app and a cloud platform. The mobile app is used to display sleep parameters in real time and generate sleep quality reports, while the cloud platform is used for long-term data storage and health risk warnings.

8. The non-invasive continuous monitoring mattress for neonatal physiological indicators according to claim 1, characterized in that: The flexible monitoring layer (3) is fixedly connected to one side of the first sponge layer (21) via Velcro.

Images