Intelligent nursing system and nursing method

Abstract

The invention provides an intelligent nursing system and method, and relates to the field of medical instruments, and the intelligent nursing system comprises a nursing device, a temperature adjusting device, a monitoring device and a control processing device. The feeding device comprises an injection pump, an injection tube, an infusion tube and a simulation nipple. The monitoring device comprises a flow velocity monitoring unit, a pressure monitoring unit and a temperature monitoring unit. The control processing device comprises a machine learning module and a flow velocity control module, the machine learning module predicts a sucking demand through a neural network based on the sucking pressure monitored by the pressure sensor and generates a flow velocity correction factor, and the flow velocity control module adjusts the flow velocity of the injection pump based on the correction factor. The system further comprises a feeding efficiency evaluation module, a waveform analysis module, a feeding scheme configuration module, a data synchronization module and an alarm linkage module, wherein the data synchronization module and the alarm linkage module are linked with a hospital information system. According to the invention, precise feeding management of premature infants and swallowing dysfunction children is realized, and the feeding safety, efficiency and individual suitability are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical devices, and in particular to an intelligent breastfeeding system and a breastfeeding method. Background Art

[0002] For hospitalized newborns, feeding is a crucial component of medical care, not only impacting their growth and development but also directly impacting their neurological development, immune function, and long-term health outcomes. Especially for premature infants and those with swallowing disorders, the flow rate, temperature, pressure fluctuations, and suck-swallow-respiratory (SSR) coordination during feeding significantly impact feeding tolerance. Therefore, refined feeding control has become an important development direction in neonatal clinical care.

[0003] At present, the clinical practice mainly adopts traditional gravity drip, syringe pump infusion or manual feeding by nurses to supply liquid milk or nutrient solution. The gravity drip method relies on the height of the liquid level to adjust the flow rate, lacks precise control, and cannot respond to the sucking needs of the newborn in real time. This method makes it difficult to accurately adjust the feeding speed, which can easily cause feeding to be too fast or too slow, increasing the risk of feeding intolerance. The syringe pump infusion method infuses at a constant rate, but fails to adapt to the individualized feeding rhythm and cannot adjust the flow rate according to the real-time sucking intensity and frequency of the newborn, which may lead to feeding intolerance or aspiration risk. When the newborn's sucking ability temporarily decreases, the constant flow rate may cause excessive fluid intake, triggering apnea or hypoxia events.

[0004] Handfeeding is the most common method in clinical practice, relying on the caregiver's experience to adjust the bottle's tilt and gently tap the nipple. However, this procedure is subjective, difficult to standardize, and unable to quantify the newborn's sucking ability and feeding results. Caregivers rely on visual observation and experience to judge the feeding process, unable to accurately identify changes in sucking patterns or precisely adjust the liquid flow rate to match the newborn's swallowing ability. This experience-based feeding method leads to differences in operation among different caregivers, making it difficult to establish a unified feeding standard.

[0005] In addition, traditional feeding methods generally lack data recording and analysis functions, making it difficult for medical staff to optimize feeding plans based on long-term data. The lack of systematic data collection and analysis tools makes it impossible for medical staff to accurately track the specific parameters of each feeding, such as sucking intensity, sucking frequency, feeding duration and other key indicators. This lack of data leads to a lag in individualized feeding adjustments, making it difficult to predict and improve subsequent feeding plans based on historical feeding performance. Furthermore, manual feeding is labor-intensive for nursing staff. The newborn is in the incubator, and the nursing staff needs to bend over, stretch their arms to enter the incubator, hold the bottle in one hand, and support the newborn's back with the other hand. After feeding, they need to continue to maintain their position and pat the newborn's back to prevent the newborn from vomiting.

[0006] The limitations of existing technologies primarily lie in their lack of adaptability to newborn sucking patterns, making it impossible to precisely control the feeding process, manage it using data, and provide intelligent feedback. Furthermore, the common practice of manual feeding increases caregiver fatigue. With the development of intelligent medical equipment, the application of data monitoring, machine learning, and control algorithms in feeding management has gradually expanded, providing new technological approaches for improving feeding safety and efficiency. However, the market currently lacks intelligent systems that can comprehensively apply these advanced technologies to achieve personalized, precise feeding for newborns. Summary of the Invention

[0007] The purpose of the present invention is to provide an intelligent feeding system and feeding method that can monitor the sucking parameters of newborns in real time, dynamically adjust the feeding flow rate, quantitatively evaluate the feeding effect and support data tracking and analysis, thereby improving the feeding safety and efficiency of premature infants and children with swallowing dysfunction, reducing the labor intensity of nursing staff, and improving the level of clinical feeding management.

[0008] To achieve the above object, the present invention is implemented through the following technical solutions:

[0009] An intelligent breastfeeding system, comprising

[0010] A feeding device comprising a syringe pump, a syringe mounted on the syringe pump, an infusion tube connected to the syringe tube, and a simulated nipple connected to the infusion tube;

[0011] a temperature regulating device for regulating the temperature of the liquid in the infusion tube;

[0012] A monitoring device, comprising a flow rate monitoring unit, a pressure monitoring unit, and a temperature monitoring unit. The flow rate monitoring unit is used to monitor the flow rate of the liquid. The pressure monitoring unit includes a pressure sensor provided in the artificial nipple for monitoring the sucking pressure. The temperature monitoring unit is used to monitor the temperature of the liquid in the infusion tube.

[0013] a control processing device electrically connected to the feeding device, the temperature regulating device, and the monitoring device, wherein the control processing device is configured to dynamically adjust the flow rate of the injection pump based on information monitored by the monitoring device, wherein the control processing device includes: a machine learning module for predicting sucking demand based on the sucking pressure monitored by the pressure monitoring unit and generating a flow rate correction factor; and a flow rate control module for adjusting the flow rate of the injection pump based on the flow rate correction factor.

[0014] Furthermore: the machine learning module also includes: a pressure waveform prediction unit, used to predict future sucking pressure based on historical pressure data through a neural network; and a parameter optimization unit, used to optimize the PID control parameters of the injection pump.

[0015] Furthermore: the machine learning module also includes: a flow rate compensation unit, which is used to adjust the target flow rate based on the impact of temperature changes on liquid viscosity.

[0016] Furthermore, the control processing device further includes a feeding efficiency evaluation module for calculating a swallowing efficiency index based on the sucking pressure monitored by the pressure monitoring unit and predicting feeding tolerance.

[0017] Furthermore, the control processing device further includes: a waveform analysis module, which is used to perform waveform processing on the sucking pressure monitored by the pressure monitoring unit, identify the sucking pattern and calculate the sucking characteristic parameters.

[0018] Furthermore, the control processing device further includes a feeding plan configuration module for generating a personalized feeding plan based on the sucking pattern identified by the waveform analysis module and the sucking characteristic parameters calculated.

[0019] Furthermore, the control processing device further includes: a data synchronization module for synchronizing feeding data to a hospital information system; and an alarm linkage module for triggering an alarm when feeding is abnormal and adjusting the flow rate in conjunction with the monitoring equipment.

[0020] The present invention also provides a breastfeeding method using the above-mentioned intelligent breastfeeding system, comprising the following steps:

[0021] S1: monitoring the sucking pressure by the pressure monitoring unit;

[0022] S2: predicting sucking demand based on the monitored sucking pressure through the machine learning module and generating a flow rate correction factor;

[0023] S3: Adjusting the flow rate of the injection pump based on the flow rate correction factor through the flow rate control module.

[0024] Furthermore, step S2 specifically includes the following steps:

[0025] Predicting future sucking pressure based on historical pressure data using a neural network;

[0026] Optimizing and controlling the PID control parameters of the injection pump through reinforcement learning;

[0027] Adjust target flow rate based on the effect of temperature changes on liquid viscosity.

[0028] Furthermore, after step S3, the following steps are also included:

[0029] S4: calculating a swallowing efficiency index based on the sucking pressure monitored by the pressure monitoring unit, and predicting feeding tolerance;

[0030] S5: performing waveform processing on the sucking pressure monitored by the pressure monitoring unit, identifying the sucking pattern and calculating sucking characteristic parameters;

[0031] S6: Generate a personalized feeding plan based on the identified sucking pattern and the calculated sucking characteristic parameters.

[0032] Furthermore, after step S6, the following steps are also included:

[0033] S7: Synchronize feeding data to the hospital information system;

[0034] S8: Trigger an alarm when feeding is abnormal and adjust the flow rate in conjunction with the monitoring equipment.

[0035] Furthermore: in step S4, the method for calculating the swallowing efficiency index is:

[0036]

[0037] Among them E swallow is the swallowing efficiency index, P max,i is the sucking pressure peak sequence, T int,i is the inter-peak interval time, t swallow,i is the time series of swallowing events, N is the number of sucking cycles, and M is the number of swallowing events.

[0038] Furthermore, in step S5, a dynamic time warping algorithm is used to identify sucking patterns, wherein the sucking patterns include explosive sucking, rhythmic sucking, and disordered sucking.

[0039] Furthermore, in step S6, the recommended flow rate reference value is calculated based on the corrected gestational age weight factor, the average amplitude of the first suck, the pressure feedback delay time and the number of days of pulmonary surfactant administration.

[0040] Furthermore, in step S6, the flow rate adjustment range is controlled to be within 15% of the reference value, and the temperature compensation value is controlled to be within the range of ±0.5°C.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] First, this invention uses a pressure sensor embedded in a simulated nipple to monitor newborn sucking pressure in real time. Combined with a pressure waveform prediction model built using an LSTM neural network using a machine learning module, it proactively predicts future sucking needs and generates a flow rate correction factor, enabling precise dynamic adjustment of feeding flow rate. Furthermore, a PID control parameter matrix optimized through reinforcement learning adaptively responds to changes in sucking patterns, effectively reducing the adverse effects of flow rate fluctuations on swallowing coordination during feeding, significantly improving feeding safety and efficiency.

[0043] Second, the feeding effectiveness assessment module of the present invention calculates the swallowing efficiency index of suck-swallow-respiratory (SSR) coordination and combines it with the Cox proportional hazards regression model to predict feeding tolerance, providing medical staff with a scientific and quantitative assessment basis. The waveform analysis module uses a dynamic time warping algorithm to identify three types of sucking patterns: burst, rhythmic, and disordered. It calculates key characteristic parameters such as sucking frequency, peak negative pressure, and effective work. This allows feeding plans to be personalized based on objective data rather than subjective experience, improving the accuracy and consistency of clinical decision-making.

[0044] 3. The feeding plan configuration module of the present invention takes into account individual physiological parameters such as the correction gestational age weight factor, the average amplitude of the first suck, and the pressure feedback delay time. By strictly controlling the flow rate adjustment amplitude within 15% of the baseline value and the temperature compensation value within the range of ±0.5°C, a smooth transition of the feeding process is ensured, effectively reducing the occurrence of feeding intolerance and improving the feeding adaptability of premature infants and children with swallowing dysfunction.

[0045] 4. The present invention realizes data synchronization and abnormal alarm linkage with the hospital information system. It can automatically trigger a nutrition department consultation reminder when the feeding residual amount is abnormal, and work with the monitoring equipment to adjust the flow rate in real time when the blood oxygen saturation decreases. It builds a full-range feeding management closed loop from data collection, intelligent analysis to safety warning, changes the limitations of traditional feeding methods such as the lack of data recording and analysis, and provides traceable data support for the long-term optimization of individual feeding plans.

[0046] 5. The present invention deeply integrates advanced pressure sensing technology, machine learning algorithms and medical information systems to create a systematic solution that can monitor, accurately analyze and intelligently adjust the feeding process in real time. It effectively solves the shortcomings of traditional feeding methods in accuracy, adaptability and data utilization, and provides safer, more efficient and more personalized feeding management for premature infants and children with swallowing dysfunction, while reducing the labor intensity of nursing staff, and has important clinical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A schematic diagram of a system architecture of an intelligent breastfeeding system in an embodiment;

[0048] Figure 2 A schematic structural diagram of a feeding device according to an embodiment;

[0049] Figure 3 It is a schematic diagram of the dynamic flow rate control technology architecture;

[0050] Figure 4 This is a schematic diagram of the feeding effectiveness assessment flow chart;

[0051] Figure 5A schematic diagram of a flow chart of a feeding method of an intelligent feeding system in an embodiment;

[0052] Figure 6 A schematic diagram of a flow chart of a feeding method of an intelligent feeding system in another embodiment;

[0053] In the picture:

[0054] 1. Feeding device; 101. Syringe pump; 102. Syringe tube; 103. Infusion tube; 104. Simulated pacifier; 2. Temperature regulating device; 3. Monitoring device; 301. Flow rate monitoring unit; 302. Pressure monitoring unit; 303. Temperature monitoring unit; 4. Control processing device; 401. Machine learning module; 402. Flow rate control module; 403. Control panel; 404. Display. DETAILED DESCRIPTION

[0055] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0056] In the description of the present invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0057] The present invention provides an intelligent breastfeeding system. Figure 1 As shown, the system mainly includes a feeding device 1, a temperature regulating device 2, a monitoring device 3 and a control processing device 4.

[0058] like Figure 2 As shown, the feeding device 1 includes a syringe pump 101, a syringe 102 mounted on the syringe pump 101, an infusion tube 103 connected to the syringe 102, and a simulated nipple 104 connected to the infusion tube 103. The syringe pump 101 has a fixed slot at the top for mounting the syringe 102. The inner cavity of the syringe 102 is used to hold liquid milk or nutrient solution. The simulated nipple 104 simulates the shape of a breastfeeding nipple to facilitate sucking by the newborn.

[0059] Temperature control device 2 includes a constant-temperature insulated box, a temperature sensor, and a heating element. The infusion tube 103 passes through the insulated box, which is equipped with a temperature sensor and heating element to maintain a constant temperature of the liquid in the infusion tube 103. The insulated box provides a stable temperature environment, ensuring that the liquid milk or nutrient solution remains at an appropriate temperature, improving feeding comfort for the newborn.

[0060] The monitoring device 3 includes a flow rate monitoring unit 301, a pressure monitoring unit 302, and a temperature monitoring unit 303. The flow rate monitoring unit 301 monitors the liquid's descent rate in real time via a liquid level sensor within the syringe 102 or a flowmeter within the infusion tube 103. The pressure monitoring unit 302, comprised of a pressure sensor within the simulated nipple 104, monitors changes in oral pressure during sucking. The temperature monitoring unit 303 monitors the liquid temperature within the infusion tube 103 in real time via a temperature sensor, and provides feedback to the heating element via a microprocessor for closed-loop control.

[0061] The control processing device 4 is electrically connected to the feeding device 1, the temperature adjustment device 2, and the monitoring device 3. Its front end integrates a control panel 403 and a display 404. The control panel 403 has a built-in microprocessor, and the display 404 is a touch screen that supports menu selection and data input. The control processing device 4 is configured to dynamically adjust the flow rate of the syringe pump 101 based on the information monitored by the monitoring device 3. The control processing device 4 includes a machine learning module 401 and a flow rate control module 402.

[0062] The machine learning module 401 is used to predict the sucking demand based on the sucking pressure monitored by the pressure monitoring unit 302 and generate a flow rate correction factor. This module includes a pressure waveform prediction unit, a parameter optimization unit, and a flow rate compensation unit. Figure 3 The figure shows the technical architecture of dynamic flow rate control.

[0063] The pressure waveform prediction unit predicts the future sucking pressure based on the historical pressure data through the LSTM neural network. Specifically, the input time window length is set to five seconds, the output time window length is set to three seconds, and P is set to t Represents the pressure value at time t, and the input sequence {P t-5 ,P t-4 ,P t-3 ,P t-2 ,P t-1}, the LSTM network calculates the hidden state h through a recursive gating mechanism t , output prediction sequence {P t+1 ,P t+2 ,P t+3 Pressure waveform prediction is used to assist in real-time flow rate adjustment by constructing a flow rate correction factor δV t, so that the flow rate correction model is aligned with the expected sucking demand, which is expressed as:

[0064]

[0065] where α is the sucking response adjustment factor, which is calibrated based on the individual's historical sucking pattern to adapt to the feeding needs of different newborns.

[0066] The parameter optimization unit is used to optimize the PID control parameters of the injection pump 101. This unit uses a reinforcement learning framework to dynamically optimize the PID control parameter matrix. The reward function is constructed as the weighted sum of the integral of the absolute value of the flow rate deviation and the standard deviation of the ambient temperature. The optimization goal is to minimize the fluctuation of the liquid flow rate and improve the flow stability. The reinforcement learning state space includes the current flow rate deviation, the pressure prediction value and the ambient temperature, and the action space is the PID parameter adjustment step size. The strategy is updated using a gradient-based optimization method to ensure the stability of the parameter adjustment. The reward function is set as:

[0067]

[0068] Where β is the flow rate deviation weight, γ is the temperature fluctuation factor, V 实际,i is the actual flow rate at time i, V 目标,i is the preset target flow rate, σ T is the temperature standard deviation. During the reinforcement learning training process, the policy network continuously adjusts the PID parameters to maximize the reward function value.

[0069] The flow rate compensation unit is used to adjust the target flow rate based on the effect of temperature changes on liquid viscosity. This unit establishes a flow rate compensation database to record the influence coefficient of liquid viscosity changes on flow rate under different temperature gradients, ensuring that temperature fluctuations do not significantly affect feeding stability. The main mechanism by which liquid viscosity affects flow rate is that increased viscosity leads to increased flow resistance, thereby reducing flow rate. The compensation database experimentally determines the flow rate correction factor under different temperature gradients. The system dynamically adjusts the target flow rate according to the current temperature state to maintain a stable feeding process. In actual applications, the database is combined with the reinforcement learning module to fine-tune parameters to ensure stability in complex environments.

[0070] The flow rate control module 402 is used to adjust the flow rate of the syringe pump 101 based on the flow rate correction factor. This module uses a PID algorithm to adjust the extrusion frequency of the syringe pump 101, achieving adaptive flow rate regulation. The control panel 403 displays parameters such as the raw milk volume, current flow rate, delivered milk volume, remaining milk volume, liquid temperature, and feeding countdown in real time, making it easy for medical staff to monitor the feeding process.

[0071] The control processing device 4 further includes a feeding efficiency evaluation module for calculating a swallowing efficiency index based on the sucking pressure monitored by the pressure monitoring unit 302 and predicting feeding tolerance. Figure 4 The figure below shows a flow chart for feeding effectiveness assessment. This module evaluates the coordination of sucking, swallowing, and breathing in newborns through a comprehensive analysis of pressure sensor data, flow rate trends, and ambient temperature fluctuations, and optimizes feeding plans based on individual feeding tolerance.

[0072] The sucking-swallowing-breathing coordination analysis unit extracts the characteristic values of sucking cycle, sucking amplitude and pause duration based on the pressure sensor data, and calculates the swallowing efficiency index in combination with the flow rate curve. The peak sequence of the sucking pressure waveform is set as P max,i , the inter-peak interval is T int,i , the swallowing time is t swallow,i , the swallowing efficiency index is defined as:

[0073]

[0074] Where N is the number of sucking cycles and M is the number of swallowing events. The swallowing efficiency index quantifies the coordination of a newborn's swallowing ability and sucking rhythm. A higher value indicates a better match between swallowing movements and sucking rhythm, which helps improve feeding efficiency.

[0075] Feeding tolerance prediction was performed using the Cox proportional hazards regression method to assess the risk of feeding tolerance. The input parameters included the growth rate of single feeding duration, the change rate of residual milk volume, and the frequency of temperature fluctuations. The regression model was as follows:

[0076] h(t)=h0(t)exp(β1R time +β2R milk +β3F temp )

[0077] Where h(t) represents the instantaneous risk of feeding intolerance in the newborn at time t, h0(t) is the baseline risk function, and R time R is the growth rate of single feeding time, milk is the residual milk volume change rate, F temp is the frequency of temperature fluctuations, and β1, β2, and β3 are parameters to be estimated. This model predicts the probability of intolerance during individual feeding through regression analysis of historical data, supporting early intervention strategies.

[0078] The control processing device 4 also includes a waveform analysis module for processing the sucking pressure monitored by the pressure monitoring unit 302, identifying the sucking pattern and calculating the sucking characteristic parameters. This module accurately quantifies the sucking characteristics of the newborn based on high-frequency data sampling, pattern recognition and interactive analysis technology. The real-time waveform rendering engine reconstructs the pressure waveform at a sampling rate of 100Hz and uses an adaptive smoothing algorithm to eliminate mechanical vibration noise to ensure high-precision signal analysis of the sucking process. The original pressure signal is set to P t , the weighted sliding mean filtering method is introduced to optimize the signal, and the smoothed pressure value P′ t The calculation formula is as follows:

[0079]

[0080] where w i is the adaptive weight, k is the smoothing window radius, is the normalization factor. This method dynamically adjusts the weight distribution so that the signal can effectively suppress high-frequency noise while maintaining edge sharpness.

[0081] Pattern recognition uses the dynamic time warping (DTW) algorithm to identify sucking patterns, which include burst sucking, rhythmic sucking, and disordered sucking. Burst sucking is characterized by a high-frequency sequence of negative pressure peaks in a short period of time, rhythmic sucking presents periodic negative pressure fluctuations, and disordered sucking is characterized by irregular changes in sucking intervals and amplitudes. Let the sucking pressure time series be {P1, P2, ..., P n}, the standard pattern sequence is {Q1,Q2,…,Q m}, DTW distance is defined as:

[0082] D(i,j)=|P i -Q j |+min(D(i-1,j),D(i,j-1),D(i-1,j-1))

[0083] Where D(i,j) represents the matching error at the current moment. Minimizing this distance can achieve optimal pattern matching, and then automatically classify the sucking type, providing a basis for subsequent feeding strategy optimization.

[0084] The interactive analysis tool allows medical staff to define a waveform interval on the touch screen, and the system automatically calculates the sucking frequency, negative pressure peak, and effective work parameters within that interval. The sucking frequency is determined by the number of negative pressure cycles per unit time, the negative pressure peak is extracted from the local minimum of the waveform, and the effective work is calculated as the integral of the sucking negative pressure, which is defined as:

[0085]

[0086] Where t1 and t2 are the start and end times of the analysis interval, Δt is the sampling time interval, and the calculation results reflect the energy consumption level of the newborn's sucking.

[0087] The control and processing device 4 also includes a feeding plan configuration module, which generates a personalized feeding plan based on the sucking patterns identified and the sucking characteristic parameters calculated by the waveform analysis module. This module constructs a three-dimensional feeding plan matrix, using the premature infant classification, swallowing dysfunction level, and nutrient solution type as variables. This generates a total of 72 basic feeding plans, each with corresponding reference values for initial flow rate, feeding interval, and temperature settings.

[0088] The personalized parameter generation algorithm further adjusts the recommended flow rate reference value V based on the basic solution base , to adapt to individual physiological characteristics, the calculation formula is:

[0089]

[0090] Where W gestational To correct for gestational age, the weight factor increases with gestational age to reflect the stronger sucking ability of more mature infants; W gestational The average amplitude of the first suck is extracted by the dynamic waveform analysis system to measure the individual sucking strength; response Represents the pressure feedback delay time. A larger value indicates that the newborn is less adaptable to flow rate changes and the initial flow rate needs to be reduced. maturity The number of days of pulmonary surfactant administration was logarithmically transformed to smooth its effect, ensuring that the developmental state of respiratory function in premature infants is included in the calculation. This formula is designed to combine individual physiological parameters to provide an appropriate initial flow rate for different neonates, optimize feeding efficiency, and reduce the risk of aspiration.

[0091] The system uses a progressive adjustment mechanism to control the flow rate adjustment range within 15% of the baseline value to ensure the adaptability of the newborn and prevent discomfort caused by sudden changes. The temperature compensation value is limited to ±0.5℃ to ensure the stability of the liquid temperature and avoid changes in fluid viscosity due to temperature fluctuations, which may affect the flow rate and sucking coordination. Set the flow rate correction value ΔV during the current feeding cycle t The calculation is as follows:

[0092]

[0093] Where V target is the recommended flow rate, V actual is the current real-time flow rate, dV actual / dt is the flow rate change rate, α and β are adjustment coefficients, the former is used to correct the current flow rate deviation, and the latter is used to suppress the mutation trend to ensure a smooth transition during the feeding process.

[0094] The control and processing device 4 also includes a data synchronization module and an alarm linkage module. The data synchronization module is used to synchronize feeding data to the hospital information system. This module implements deep integration of the hospital information system (HIS) based on the HL7FHIR standard, ensuring standardized storage, real-time synchronization, and intelligent linkage control of feeding data. The data synchronization interface uses the FHIRObservation resource type to store feeding records, where the effectiveDateTime field records the feeding time, and the valueQuantity field contains core data such as the single feeding amount, flow rate, and feeding duration. It also supports extended fields to store temperature curves and sucking pressure waveform characteristic parameters.

[0095] The alarm linkage module is used to trigger an alarm when feeding is abnormal and to adjust the flow rate in conjunction with the monitoring equipment. The intelligent alarm linkage mechanism triggers a consultation reminder from the nutrition department based on the trend of feeding residual amount changes. The system continuously monitors the residual amount R t And calculate the mean of the changes in the three feedings:

[0096]

[0097] When ΔR>5ml, the system automatically generates an electronic medical record warning sign, notifies medical staff to adjust the feeding plan, and pushes consultation recommendations to the HIS notification center.

[0098] Device collaborative control uses the IEEE 11073 device interconnection protocol to synchronize the feeding system and the monitor, ensuring dynamic monitoring of vital signs during feeding. When the blood oxygen saturation (SpO2) data falls below the safety threshold of 88% and persists for 10 seconds, the system immediately executes the flow rate adjustment strategy and calculates the flow rate correction value ΔV SpO2 :

[0099] ΔV SpO2 =-k(88-SpO2 min )

[0100] Where SpO2 min is the currently monitored lowest blood oxygen value, and k is the flow rate attenuation coefficient; a negative value indicates a decrease in flow rate. The system calculates the corrected flow rate based on this formula and uses PID control to adjust the operating parameters of syringe pump 101 in real time, gradually reducing the flow rate to within 50%, thereby reducing the respiratory burden on the newborn during feeding.

[0101] In another embodiment, the present invention further provides a feeding method using the above-mentioned intelligent feeding system, the specific steps of which are as follows:

[0102] S1: Monitor sucking pressure through the pressure monitoring unit 302. The pressure monitoring unit 302 uses the micro pressure sensor embedded in the artificial nipple 104 to monitor the changes in the oral pressure of the newborn during sucking and collects pressure data in real time.

[0103] S2: The machine learning module 401 predicts sucking demand based on the monitored sucking pressure and generates a flow rate correction factor. This includes: using an LSTM neural network to predict future sucking pressure based on historical pressure data; optimizing the PID control parameters of the syringe pump 101 through reinforcement learning; and adjusting the target flow rate based on the effect of temperature changes on liquid viscosity.

[0104] S3: The flow rate control module 402 adjusts the flow rate of the syringe pump 101 based on the flow rate correction factor. The flow rate control module 402 adjusts the extrusion frequency of the syringe pump 101 through the PID algorithm according to the flow rate correction factor generated by the machine learning module 401 to achieve adaptive flow rate regulation.

[0105] In another embodiment, the method further includes the following steps: S4: calculating a swallowing efficiency index based on the sucking pressure monitored by the pressure monitoring unit 302 and predicting feeding tolerance; S5: performing waveform processing on the sucking pressure monitored by the pressure monitoring unit 302, identifying a sucking pattern and calculating sucking characteristic parameters; S6: generating a personalized feeding plan based on the identified sucking pattern and the calculated sucking characteristic parameters.

[0106] In another embodiment, the method further includes the following steps: S7: synchronizing feeding data to a hospital information system; S8: triggering an alarm in the event of feeding anomalies and adjusting the flow rate in conjunction with monitoring equipment. When the mean change in residual feeding volume exceeds a threshold, a nutrition consultation reminder is triggered; and when blood oxygen saturation remains below a safety threshold for a certain period of time, a flow rate adjustment strategy is implemented to ensure feeding safety.

[0107] Through the above-mentioned system and method, the present invention realizes precise feeding management for newborns, especially premature infants and children with swallowing dysfunction, improves the safety, efficiency and individual adaptability of feeding, and provides an innovative intelligent feeding solution for clinical practice.

[0108] The above embodiments are intended only to illustrate the technical concepts and features of the present invention. Their purpose is to enable those skilled in the art to understand the contents of the present invention and implement them accordingly. They are not intended to limit the scope of protection of the present invention. Any equivalent changes or modifications made in accordance with the spirit of the present invention are intended to be covered by the scope of protection of the present invention.

Claims

1. An intelligent breastfeeding system, characterized in that: include A feeding device comprising a syringe pump, a syringe mounted on the syringe pump, an infusion tube connected to the syringe tube, and a simulated nipple connected to the infusion tube; a temperature regulating device for regulating the temperature of the liquid in the infusion tube; A monitoring device, comprising a flow rate monitoring unit, a pressure monitoring unit, and a temperature monitoring unit. The flow rate monitoring unit is used to monitor the flow rate of the liquid. The pressure monitoring unit includes a pressure sensor provided in the artificial nipple for monitoring the sucking pressure. The temperature monitoring unit is used to monitor the temperature of the liquid in the infusion tube. a control processing device electrically connected to the feeding device, the temperature regulating device, and the monitoring device, the control processing device being configured to dynamically adjust the flow rate of the syringe pump based on information monitored by the monitoring device, wherein the control processing device comprises: a machine learning module for predicting sucking demand based on the sucking pressure monitored by the pressure monitoring unit and generating a flow rate correction factor; and a flow rate control module for adjusting the flow rate of the injection pump based on the flow rate correction factor.

2. The intelligent breastfeeding system according to claim 1, characterized in that: The machine learning module includes: a pressure waveform prediction unit for predicting future sucking pressure based on historical pressure data through a neural network; and a parameter optimization unit for optimizing the PID control parameters of the injection pump.

3. The intelligent breastfeeding system according to claim 2, characterized in that: The machine learning module also includes: a flow rate compensation unit for adjusting the target flow rate based on the effect of temperature changes on liquid viscosity.

4. The intelligent breastfeeding system according to claim 1, characterized in that: The control processing device further includes: a feeding efficiency evaluation module, configured to calculate a swallowing efficiency index based on the sucking pressure monitored by the pressure monitoring unit, and predict feeding tolerance.

5. The intelligent breastfeeding system according to claim 4, characterized in that: The control processing device further includes: a waveform analysis module, configured to perform waveform processing on the sucking pressure monitored by the pressure monitoring unit, identify the sucking pattern, and calculate sucking characteristic parameters.

6. The intelligent breastfeeding system according to claim 5, characterized in that: The control processing device further includes a feeding plan configuration module for generating a personalized feeding plan based on the sucking pattern identified by the waveform analysis module and the sucking characteristic parameters calculated.

7. The intelligent breastfeeding system according to claim 6, characterized in that: The control processing device further includes: a data synchronization module for synchronizing feeding data to a hospital information system; and an alarm linkage module for triggering an alarm when feeding is abnormal and adjusting the flow rate in conjunction with the monitoring equipment.

8. A breastfeeding method using the intelligent breastfeeding system according to any one of claims 1 to 7, characterized in that: The following steps are involved: S1: monitoring the sucking pressure by the pressure monitoring unit; S2: predicting sucking demand based on the monitored sucking pressure through the machine learning module and generating a flow rate correction factor; S3: Adjusting the flow rate of the injection pump based on the flow rate correction factor through the flow rate control module.

9. The breastfeeding method according to claim 8, wherein: The step S2 specifically includes the following steps: Predicting future sucking pressure based on historical pressure data using a neural network; Optimizing and controlling the PID control parameters of the injection pump through reinforcement learning; Adjust target flow rate based on the effect of temperature changes on liquid viscosity.

10. The breastfeeding method according to claim 8, characterized in that: After step S3, the following steps are also included: S4: calculating a swallowing efficiency index based on the sucking pressure monitored by the pressure monitoring unit, and predicting feeding tolerance; S5: performing waveform processing on the sucking pressure monitored by the pressure monitoring unit, identifying the sucking pattern and calculating sucking characteristic parameters; S6: Generate a personalized feeding plan based on the identified sucking pattern and the calculated sucking characteristic parameters.

11. The breastfeeding method according to claim 10, wherein: After step S6, the following steps are also included: S7: Synchronize feeding data to the hospital information system; S8: Trigger an alarm when feeding is abnormal and adjust the flow rate in conjunction with the monitoring equipment.

12. The breastfeeding method according to claim 10, wherein: In step S4, the method for calculating the swallowing efficiency index is: Among them E swallow is the swallowing efficiency index, P max,i is the sucking pressure peak sequence, T int,i is the inter-peak interval time, t swallow,i is the time series of swallowing events, N is the number of sucking cycles, and M is the number of swallowing events.

13. The breastfeeding method according to claim 10, wherein: In step S5, a dynamic time warping algorithm is used to identify sucking patterns, where the sucking patterns include explosive sucking, rhythmic sucking, and disordered sucking.

14. The breastfeeding method according to claim 10, wherein: In step S6, the recommended flow rate reference value is calculated based on the corrected gestational age weight factor, the average amplitude of the first suck, the pressure feedback delay time, and the number of days of pulmonary surfactant administration.

15. The breastfeeding method according to claim 10, wherein: In step S6, the flow rate adjustment range is controlled to be within 15% of the reference value, and the temperature compensation value is controlled to be within the range of ±0.5°C.

Images