Intelligent monitoring and early warning blood sampling and pulse pressing belt, system and method

Abstract

The invention relates to the technical field of blood sampling and pulse pressing belt monitoring, and discloses a blood sampling and pulse pressing belt, system and method for intelligent monitoring and early warning. The system comprises a data acquisition unit, a state analysis unit, an abnormity judgment unit, a control execution unit, a communication feedback unit and a self-checking calibration unit. The data acquisition unit is used for acquiring pressure distribution, contact area and blood flow velocity signals in real time; the state analysis unit generates a deviation coefficient; an anomaly judgment unit constructs a dynamic fusion matrix input state classification network to output an anomaly probability; the control execution unit sends a corresponding instruction according to the abnormal probability; the communication feedback unit transmits instructions and data. The blood sampling and pulse pressing belt is applied to the system and is provided with various sensors and adjusting modules. The blood sampling method comprises the steps of collecting signals, calculating coefficients, constructing a matrix and triggering operation according to the abnormal probability. Multi-parameter real-time monitoring, intelligent regulation and early warning in the blood sampling process are achieved, blood sampling safety, accuracy and efficiency are improved, and patient experience is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of blood collection tourniquet monitoring, and specifically to an intelligent monitoring and warning blood collection tourniquet, system and method. Background Art

[0002] During the process of medical blood collection, the tourniquet, as a commonly used tool, its usage effect directly affects the smooth progress of blood collection and the patient's experience. There are many problems in the use of traditional blood collection tourniquets, which are difficult to meet the high requirements of modern medicine for the safety, accuracy and comfort of blood collection.

[0003] From the aspect of pressure control, traditional tourniquets usually rely on the experience of medical staff to adjust the tightness. This results in a lack of precision in pressure control during actual operation. If the tourniquet is too tight, it will not only cause obvious pain and discomfort to the patient, increase the patient's psychological burden, but also may impede blood circulation due to excessive pressure, causing local tissue hypoxia, affecting blood components, and further interfering with the quality of the blood collection sample, leading to deviation of the test results. On the contrary, if the tourniquet is too loose, it cannot effectively block venous blood flow, making it difficult for veins to be fully filled, increasing the difficulty of blood collection, possibly requiring multiple punctures, further aggravating the patient's pain, while also reducing the blood collection efficiency and wasting medical resources.

[0004] In the monitoring of the blood collection state, the traditional methods are very limited. Medical staff mainly judge the blood collection situation by observing the patient's body surface characteristics and simply asking the patient's feelings, lacking real-time and precise monitoring of key parameters such as the pressure distribution, contact area and blood flow velocity of the tourniquet. This monitoring method is highly subjective and cannot detect potential problems in a timely and accurate manner. For example, in some cases, even if the patient has no obvious discomfort, abnormal local pressure or blood flow velocity fluctuations of the tourniquet may affect the quality of blood collection, and these problems are difficult to detect through traditional monitoring methods, easily leading to blood collection failure or inaccurate test results, delaying the diagnosis of the disease.

[0005] With the continuous progress of medical technology and the increasing demand for the quality of medical services, the intelligent upgrade of blood collection tourniquets has become particularly urgent. An intelligent blood collection tourniquet system can not only achieve real-time and precise monitoring of various parameters during blood collection, but also make adjustments and give warnings in a timely manner based on the monitored data, effectively solving the problems existing in traditional tourniquets, improving the safety, accuracy and efficiency of blood collection, and providing reliable blood sample support for clinical diagnosis. However, although there are some blood collection tourniquets on the market that claim to have intelligent functions at present, most of the products have imperfect functions and cannot achieve comprehensive and precise monitoring and intelligent control of multiple parameters such as pressure, contact area and blood flow velocity, making it difficult to meet the actual clinical needs. Therefore, it is of great practical significance to develop an intelligent monitoring and warning blood collection tourniquet system that can effectively solve the above problems. Summary of the Invention

[0006] The object of the present invention is to provide an intelligent monitoring and warning blood collection tourniquet, system and method to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solution: An intelligent monitoring and warning blood collection tourniquet system, the system includes: A data acquisition unit for real-time acquisition of the pressure distribution signal, contact area signal and blood flow velocity signal of the tourniquet during blood collection; A state analysis unit for dynamically matching the pressure distribution signal, contact area signal and blood flow velocity signal with preset tightness parameters and blood flow expectation parameters through an adaptive adjustment model to generate a pressure deviation coefficient, a contact deviation coefficient and a blood flow deviation coefficient; An abnormality determination unit for constructing a dynamic fusion matrix according to the pressure deviation coefficient, contact deviation coefficient and blood flow deviation coefficient, inputting it into a pre-trained state classification network, and outputting the probability of abnormal blood collection state; A control execution unit for generating a tightness adjustment instruction, an emergency release instruction or an alarm trigger instruction when the probability of abnormal blood collection state is greater than or equal to a preset threshold; A communication feedback unit for transmitting the tightness adjustment instruction, emergency release instruction or alarm trigger instruction to the tourniquet terminal and the background server; Among them, the execution steps of the state analysis unit include: Performing spatio-temporal decomposition on the pressure distribution signal to extract pressure time-domain features and spatial distribution features; Based on the pressure time-domain features and spatial distribution features, combined with preset tightness parameters, calculating a dynamic pressure deviation coefficient through a sliding window algorithm; Fusing the time series of the contact area signal and the blood flow velocity signal to generate a contact-blood flow coupling coefficient as an input parameter for the contact deviation coefficient and the blood flow deviation coefficient.

[0008] Preferably, the execution steps of the data acquisition unit include: Collecting the pressure distribution signal through a distributed pressure sensor array embedded in the tourniquet; Obtaining the contact area signal through an infrared optical sensor; Real-time monitoring of the blood flow velocity signal through a Doppler blood flow meter; Normalizing the pressure distribution signal, contact area signal and blood flow velocity signal into a data stream with a unified sampling frequency.

[0009] Preferably, the steps of constructing a dynamic fusion matrix include: Normalize the dynamic pressure deviation coefficient, contact deviation coefficient, and blood flow deviation coefficient into a three-dimensional vector; Generate the fusion matrix eigenvalue according to the weight ratio of each component in the three-dimensional vector; Suppress the noise of the fusion matrix eigenvalue through the Kalman filter algorithm to obtain an optimized dynamic fusion matrix.

[0010] Preferably, the construction steps of the state classification network include: Collect the pressure distribution signal in the historical blood collection data , contact area signal , and blood flow velocity signal , and label the corresponding blood collection state label ; Extract the time-domain feature vector of the pressure distribution signal and the spatial distribution feature matrix , where is the pressure value at the time point, is the pressure variance, is the sampling window length; is the partial derivative of the pressure value in the axis direction, is the partial derivative of the pressure value in the axis direction, is the matrix dimension parameter, representing the grid layout resolution of the distributed pressure sensor array on the surface of the tourniquet; Construct the contact-blood flow coupling coefficient calculation function:

[0011] Among them, is the contact-blood flow coupling coefficient, are the weight factors of the contact area deviation and the blood flow velocity deviation respectively, is the historical contact area signal, referring to the measured area value of the contact area between the tourniquet and the skin in the time series, dynamically captured by the infrared optical sensor, is the preset contact area threshold, is the historical blood flow velocity signal, referring to the measured value of the blood flow rate in the target blood vessel during blood collection, collected by the Doppler blood flow meter in the pulsed wave mode, is the blood flow velocity expected value, is the normalization operator; Using , and to form a training sample set Using Using the supervised data to train the weight parameters of the convolutional long short-term memory hybrid neural network to generate the state classification network.

[0012] Preferably, the execution steps of the control execution unit further include: When generating the tightness adjustment instruction, according to the priority sorting of the pressure deviation coefficients, call the airbag hierarchical adjustment module of the tourniquet to adjust the airbag pressure in gradients; When generating the emergency release instruction, trigger the electromagnetic lock quick release mechanism and reset the initial state of the tourniquet after release; When generating the alarm trigger instruction, start the audible and visual alarm device and send an emergency status code to the background server.

[0013] Preferably, the generation steps of the priority sorting include: Calculate the weight factors according to the contribution degrees of the components in the dynamic fusion matrix; Combine the weight factors with the real-time blood collection stage parameters to generate the priority sequence of the pressure deviation coefficients through the fuzzy logic algorithm.

[0014] Preferably, the execution steps of the communication feedback unit include: Transmit instructions to the tourniquet terminal through the low-power Bluetooth module; Establish an encrypted communication link with the background server through the 5G narrowband Internet of Things protocol and upload the blood collection status data and exception logs in real time.

[0015] Preferably, the system further includes: A self-check and calibration unit, used to trigger a self-check process through the tourniquet terminal before blood collection starts, including: Traverse the output consistency of the distributed pressure sensor array, and eliminate abnormal sensor nodes; Calibrate the baseline parameters of the infrared optical sensor to ensure the accuracy of the contact area signal; Verify the signal stability of the Doppler blood flow meter, generate a self-check report and feedback it to the background server.

[0016] Preferably, the present invention further includes an intelligent monitoring and warning blood collection tourniquet, applied to the above-mentioned blood collection tourniquet system, and the tourniquet includes: An elastic ring-shaped main body, the surface is covered with an antibacterial coating, and a distributed pressure sensor array, an infrared optical sensor and a Doppler blood flow meter connected to the data acquisition unit are embedded inside; An airbag hierarchical adjustment module electrically connected to the control execution unit, including 6 independent air chambers distributed along the circumferential direction of the ring, and each air chamber realizes gradient pressure adjustment through a micro air pump and a solenoid valve; An electromagnetic latch integrated at the end of the elastic ring-shaped body triggers a spring-type quick separation mechanism in response to the emergency release instruction; A detachable communication module, which is built with a low-power Bluetooth chip and a 5G NB-IoT antenna array that match the communication feedback unit; Among them, the node density of the distributed pressure sensor array is 4 sensing units per square centimeter, and the spatial resolution matches the dimension requirements of the spatial distribution feature matrix in the state analysis unit.

[0017] Preferably, the present invention further includes an intelligent monitoring and warning blood collection method, which is applied to the above-mentioned blood collection tourniquet system. The method includes: Real-time collection of the pressure distribution signal, contact area signal, and blood flow velocity signal of the tourniquet; Calculating the dynamic pressure deviation coefficient, contact deviation coefficient, and blood flow deviation coefficient through an adaptive adjustment model; Constructing a dynamic fusion matrix and inputting it into the state classification network to output the probability of abnormal blood collection status; Triggering tightness adjustment, emergency release, or alarm operations according to the abnormal probability; Among them, the construction steps of the adaptive adjustment model include: Based on different blood collection stages, dynamically adjusting the expected threshold of the tightness parameter; Using a multi-objective optimization algorithm to balance the constraint conditions of pressure uniformity, blood flow stability, and contact fit.

[0018] Compared with the prior art, the beneficial effects of the present invention are: In terms of blood collection safety and patient experience, the system obtains the pressure distribution signal, contact area signal, and blood flow velocity signal of the tourniquet in real time through the data acquisition unit, providing a basis for precise control. The state analysis unit dynamically matches these signals with preset parameters to generate deviation coefficients, and the abnormal determination unit accurately judges the probability of abnormal blood collection status based on this. Once the abnormal probability is greater than the preset threshold, the control execution unit quickly responds. When generating a tightness adjustment instruction, the airbag hierarchical adjustment module is called according to the priority ranking of the pressure deviation coefficient, and the airbag pressure is adjusted in gradients to avoid discomfort and harm caused to the patient by the tourniquet being too tight or too loose, greatly improving the comfort of the patient during blood collection. The emergency release instruction can trigger the quick release mechanism of the electromagnetic latch, and in case of an emergency, quickly release the restraint of the tourniquet on the patient's limb, ensuring the safety of the patient and effectively avoiding serious consequences such as limb injuries caused by improper use of the tourniquet.

[0019] From the perspective of blood collection quality and accuracy, the precise monitoring and regulation of various parameters by the system play a crucial role. By performing spatio-temporal decomposition on the pressure distribution signal and calculating the dynamic pressure deviation coefficient in combination with the preset tightness parameter, it ensures uniform pressure of the tourniquet and avoids abnormal local pressure from affecting blood flow and collection. The contact-blood flow coupling coefficient generated by fusing the contact area signal and the blood flow velocity signal, as the input parameter of the contact deviation coefficient and the blood flow deviation coefficient, can comprehensively consider various factors during the blood collection process. In this way, it can effectively maintain stable blood flow, ensure that the quality of the collected blood sample is not affected, provide accurate and reliable samples for subsequent medical tests, improve the accuracy of diagnosis, and reduce misdiagnosis and missed diagnosis cases caused by sample problems.

[0020] In terms of improving blood collection efficiency, the system also performs excellently. The real-time monitoring and intelligent regulation functions enable medical staff to avoid repeatedly adjusting the tourniquet based on experience, reducing the preparation time before blood collection. Once an abnormal situation occurs, the system can respond quickly and handle it, avoiding interruptions in the operation and re-puncturing during the blood collection process, thus greatly shortening the overall blood collection time, improving blood collection efficiency, optimizing the utilization of medical resources, enabling medical staff to serve more patients within the same time, and enhancing the overall level of medical services.

[0021] In addition, the self-check and calibration unit in this application comprehensively detects the equipment before blood collection starts, ensuring the normal operation of each sensor, and improving the reliability and stability of the equipment. The communication feedback unit transmits instructions to the tourniquet terminal through a low-power Bluetooth module and establishes an encrypted communication link with the background server through the 5G narrowband Internet of Things protocol, uploading the blood collection status data and abnormal logs in real time, facilitating medical staff to view and analyze the blood collection situation at any time, realizing the informatization management of medical data, and helping to improve the management level and quality traceability ability of medical services. The intelligent monitoring and warning blood collection tourniquet system, tourniquet, and blood collection method solve the problems existing in the traditional blood collection method from multiple dimensions, bringing a safer, more efficient, and more accurate solution to the medical blood collection field, and having broad application prospects and significant social and economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 is the working principle diagram of the blood collection tourniquet system described in the present invention; Figure 2 is the working principle diagram of the data acquisition unit; Figure 3 is the working principle diagram of constructing the dynamic fusion matrix; Figure 4 is the working principle diagram of the communication feedback unit. DETAILED DESCRIPTION OF THE INVENTION

[0023] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0024] Please refer to Figures 1-4 , the present invention provides a blood pressure cuff system for intelligent monitoring and early warning, and the system is composed of a data acquisition unit, a state analysis unit, an anomaly determination unit, a control execution unit, and a communication feedback unit. The specific implementation scheme is as follows: Data acquisition unit: Real-time acquisition of the pressure distribution signal, contact area signal, and blood flow velocity signal of the blood pressure cuff during the blood collection process.

[0025] State analysis unit: Through an adaptive adjustment model, dynamically match the pressure distribution signal, contact area signal, and blood flow velocity signal with preset tightness parameters and blood flow expectation parameters to generate a pressure deviation coefficient, a contact deviation coefficient, and a blood flow deviation coefficient.

[0026] Anomaly determination unit: Construct a dynamic fusion matrix based on the pressure deviation coefficient, contact deviation coefficient, and blood flow deviation coefficient, input it into a pre-trained state classification network, and output the probability of abnormal blood collection status.

[0027] Control execution unit: When the probability of abnormal blood collection status is greater than or equal to a preset threshold, generate a tightness adjustment instruction, an emergency release instruction, or an alarm trigger instruction.

[0028] Communication feedback unit: Transmit the tightness adjustment instruction, emergency release instruction, or alarm trigger instruction to the blood pressure cuff terminal and the background server.

[0029] The execution steps of the state analysis unit include: Perform spatio-temporal decomposition on the pressure distribution signal, and extract the pressure time-domain characteristics and spatial distribution characteristics; Based on the pressure time-domain characteristics and spatial distribution characteristics, combined with the preset tightness parameters, calculate the dynamic pressure deviation coefficient through a sliding window algorithm; Fuse the time series of the contact area signal and the blood flow velocity signal to generate a contact-blood flow coupling coefficient as the input parameter of the contact deviation coefficient and the blood flow deviation coefficient.

[0030] The following further illustrates the present invention with reference to Embodiments 1 to 5: Embodiment 1: This embodiment details how the data acquisition unit realizes the acquisition of the pressure distribution signal, contact area signal, and blood flow velocity signal, ensuring the accuracy and reliability of the acquired data, and providing a basis for the subsequent analysis and decision-making of the system.

[0031] The data acquisition unit obtains signals through the following specific methods: Acquisition of pressure distribution signals: The tourniquet is embedded with a distributed pressure sensor array. These sensors are closely distributed and can accurately sense the pressure changes at the contact part between the tourniquet and the human body during blood collection. For example, in an actual application scenario, when the tourniquet is wound around the arm, the distributed pressure sensor array can collect the pressure values at different positions on the arm in real time to form a pressure distribution signal. Each sensor node can quickly respond to pressure changes and has high sensitivity, capable of capturing minute pressure fluctuations to ensure that the collected pressure distribution signal truly reflects the actual situation.

[0032] Obtaining contact area signals: Infrared optical sensors are used to obtain contact area signals. The infrared optical sensors emit infrared light. When the light irradiates the contact area between the tourniquet and the skin, part of the light is reflected back, and the sensor calculates the area of the contact area by receiving the reflected light. In the design, the installation position and angle of the infrared optical sensors are carefully adjusted to ensure that the actual contact area between the tourniquet and the skin can be accurately measured. At the same time, advanced signal processing algorithms are used to filter and denoise the signals collected by the sensors to improve the accuracy of the contact area signals.

[0033] Monitoring blood flow velocity signals: A Doppler blood flow meter is used to monitor blood flow velocity signals in real time. The Doppler blood flow meter utilizes the Doppler effect. By emitting ultrasonic waves into the human blood vessels and receiving the ultrasonic signals reflected by the blood flow in the vessels, through signal processing and analysis, the blood flow velocity in the target blood vessel is obtained. In actual use, the probe of the Doppler blood flow meter is integrated with the tourniquet and can be parameter-adjusted according to different blood collection sites and blood vessel conditions to ensure accurate measurement of the blood flow velocity.

[0034] Signal standardization processing: To facilitate subsequent unified processing and analysis of the collected signals, the pressure distribution signals, contact area signals, and blood flow velocity signals are standardized into data streams with a unified sampling frequency. During the signal standardization process, digital signal processing techniques are used to perform sampling rate conversion and synchronization processing on signals with different frequencies to ensure that all signals are consistent in time for subsequent fusion and analysis.

[0035] Example 2: The steps for constructing a dynamic fusion matrix include: Normalize the dynamic pressure deviation coefficient, contact deviation coefficient, and blood flow deviation coefficient into three-dimensional vectors. Through normalization, coefficients with different dimensions and value ranges are transformed into comparable vector forms for subsequent fusion calculations. For example, normalization formulas are used to map each coefficient to the interval [0, 1] so that they can be compared and fused on the same scale.

[0036] Generate the eigenvalues of the fusion matrix according to the weight ratios of the components in the three-dimensional vector. The determination of the weight ratios is based on a large amount of experimental data and data analysis. By evaluating the importance of each coefficient under different blood collection states, the weight of each component in the fusion matrix is determined. For example, in some blood collection scenarios, the pressure deviation coefficient has a greater impact on the blood collection state, so a higher weight is assigned to it; while in other scenarios, the blood flow deviation coefficient may be more critical, and its weight is correspondingly increased.

[0037] Suppress the noise of the eigenvalues of the fusion matrix through the Kalman filter algorithm to obtain an optimized dynamic fusion matrix. The Kalman filter algorithm can utilize the historical information of the signal and the current measurement value to effectively estimate and remove the noise, improving the accuracy and stability of the dynamic fusion matrix. In practical applications, the collected signals are often interfered by various noises. Through the processing of the Kalman filter algorithm, the quality of the dynamic fusion matrix can be significantly improved, providing more reliable data for subsequent state classification.

[0038] This embodiment also includes the training process of the state classification network, and the specific steps include: Widely collect pressure distribution signals from a large amount of historical blood collection data , contact area signals , and blood flow velocity signals . These data cover different patients, different blood collection environments, and different blood collection stages to ensure the diversity and comprehensiveness of the training data. At the same time, each set of data is labeled with the corresponding blood collection state label , clarifying whether the set of data represents a normal blood collection state or a certain abnormal blood collection state.

[0039] For the pressure distribution signal, extract the time-domain feature vector . Among them, is the pressure value at the time point, is the pressure variance, reflecting the pressure fluctuation situation, is the sampling window length, used to limit the time range for calculating the time-domain features, and is the maximum pressure value within the sampling window.

[0040] Through these parameters, the characteristics of the pressure distribution signal are comprehensively characterized from the time dimension.

[0041] Extract the spatial distribution feature matrix of the pressure distribution signal .

[0042] Here, is the partial derivative of the pressure value in the axis direction, is the pressure value in the The partial derivative in the axial direction is the matrix dimension parameter, which matches the grid layout resolution of the distributed pressure sensor array on the surface of the tourniquet, so as to describe the change of pressure distribution from the spatial dimension.

[0043] Construct the coupling coefficient calculation: The function is as follows:

[0044] Among them, is the coupling coefficient, are the weight factors of the contact area deviation and the blood flow velocity deviation respectively, which are used to adjust the relative importance of the two factors in the coupling coefficient. is the historical contact area signal, which is the measured area value of the contact area between the tourniquet and the skin in the time series, and is dynamically captured by the infrared optical sensor; is the preset contact area threshold, which is used as the standard to measure whether the contact area is normal. is the historical blood flow velocity signal, which is the measured value of the blood flow rate in the target blood vessel during blood collection, and is collected by the Doppler blood flow meter in the pulse wave mode; is the expected blood flow velocity value, which represents the expected blood flow velocity under normal blood collection conditions. is the normalization operator, which enables the calculation of each deviation under a unified scale, and finally obtains the coupling coefficient that comprehensively reflects the deviation of the contact area and the blood flow velocity .

[0045] The extracted time-domain feature vector , the spatial distribution feature matrix and the coupling coefficient are combined into a training sample set , and the labeled blood collection status label is used as the supervised data. A convolutional long short-term memory hybrid neural network is used for training. During the training process, by continuously adjusting the weight parameters of the network, the network learns the mapping relationship between the input features and the blood collection status label.

[0046] The convolutional neural network (CNN) part uses the convolutional layer to automatically extract the local features in the input data. By sliding the convolutional kernel on the data for convolution, it captures the local patterns of the pressure distribution, contact area, and blood flow velocity signals in space and time, such as the concentrated area of pressure, the change trend of the contact area, etc. The pooling layer then downsamples the output of the convolutional layer, reducing the data volume while retaining the main features and reducing the computational complexity.

[0047] The long short-term memory network (LSTM) can process data with time series characteristics, effectively memorize the change information of signals at different times during the blood collection process, and avoid the gradient disappearance or gradient explosion problems in long-term series learning. It can learn the dependencies between signals at different blood collection stages, such as the impact of pressure changes at a certain stage on subsequent blood flow velocity, so as to better understand the dynamic changes of the blood collection process.

[0048] Use a portion of historical data that was not involved in the training as a test set to evaluate the trained state classification network. By calculating indicators such as accuracy, recall, and F1 value, the accuracy and reliability of the network's prediction of the probability of abnormal blood sampling status can be judged. If the evaluation results do not meet expectations, analyze possible problems, such as overfitting or underfitting of the model.

[0049] If overfitting occurs, it may be that the model is too complex and has overlearned the noise and details in the training data. In this case, you can use regularization methods, such as L1 and L2 regularization, to constrain the network weights to prevent them from being too large; or reduce the number of layers and neurons in the network to reduce the complexity of the model; or increase the amount of training data so that the model can learn more generalizable features.

[0050] If there is an underfitting problem, it means that the model has not fully learned the features and patterns in the data. You can try to increase the complexity of the network, such as adding more convolutional layers or LSTM layers; adjust training parameters, such as learning rate, number of iterations, etc., to give the model more opportunities to learn data features; perform more effective feature engineering on the input data to mine more valuable features.

[0051] After evaluation and optimization, the trained state classification network is deployed to the abnormality judgment unit of the blood sampling pulse belt system of intelligent monitoring and early warning, which is used to predict the abnormal probability of blood sampling state in real time. In the actual operation of the system, as new blood sampling data continues to accumulate, the data is regularly re-collected and sorted, and the state classification network is updated and trained according to the above training steps, so that the network can adapt to the changes in blood sampling state under different circumstances and continue to maintain good prediction performance.

[0052] Embodiment 3: This embodiment describes in detail how the control execution unit performs corresponding operations according to different instruction types after receiving the abnormal probability output by the abnormal determination unit to ensure the safety and smooth progress of the blood collection process.

[0053] The control execution unit includes the following specific operations during execution: When generating a tightening / loosening adjustment instruction, according to the priority sorting of the pressure deviation coefficient, the airbag hierarchical adjustment module of the tourniquet is called to adjust the airbag pressure in gradients. For example, during actual blood collection, if the pressure deviation coefficient indicates that the tourniquet is too tight and it is determined that this pressure deviation problem is relatively serious according to the priority sorting, the control execution unit will send an instruction to the airbag hierarchical adjustment module. The airbag hierarchical adjustment module includes 6 independent air chambers distributed along the circumferential direction. Each air chamber realizes gradient pressure adjustment through a micro air pump and an electromagnetic valve. The control execution unit precisely controls the micro air pump and the electromagnetic valve according to the degree of pressure deviation, gradually reducing the airbag pressure to restore the tightness of the tourniquet to an appropriate range, avoiding discomfort to the patient or affecting the blood collection effect due to the tourniquet being too tight.

[0054] When generating an emergency release instruction, trigger the rapid release mechanism of the electromagnetic lock and reset the initial state of the tourniquet after release. In case of sudden emergencies, such as the patient experiencing severe discomfort or the device malfunctioning, the control execution unit will immediately trigger the rapid release mechanism of the electromagnetic lock. The electromagnetic lock is integrated at the end of the elastic ring-shaped main body. In response to the emergency release instruction, it triggers the spring-type rapid separation mechanism to quickly loosen the tourniquet and ensure the safety of the patient. After releasing the tourniquet, the control execution unit will automatically reset the initial state of the tourniquet, including closing relevant sensors, resetting the airbag pressure, etc., for the next use.

[0055] When generating an alarm trigger instruction, start the audible and visual alarm device and send an emergency status code to the background server. Once the probability of abnormal blood collection status reaches the preset threshold and it is determined that an alarm is needed, the control execution unit will simultaneously start the audible and visual alarm device on the tourniquet terminal. For example, emit a loud alarm sound and flashing lights to attract the attention of medical staff. At the same time, send an emergency status code to the background server through the communication feedback unit. After receiving the code, the background server can record the abnormal situation in a timely manner and notify the relevant medical staff for handling, realizing real-time monitoring and emergency response during the blood collection process.

[0056] Example 4: The operation steps of the communication feedback unit include: Communicate with the tourniquet terminal: The communication feedback unit transmits instructions to the tourniquet terminal through a low-power Bluetooth module. The low-power Bluetooth module has advantages such as low power consumption and stable connection, and can reduce the energy consumption of the device while ensuring the accuracy of data transmission. In actual applications, the tightening / loosening adjustment instructions, emergency release instructions, etc. generated by the control execution unit are quickly transmitted to the tourniquet terminal through the low-power Bluetooth module to ensure that the tourniquet can respond to the control commands of the system in a timely manner.

[0057] Communication with the background server: Establish an encrypted communication link with the background server through the 5G narrowband Internet of Things protocol, and upload blood collection status data and exception logs in real time. The 5G narrowband Internet of Things protocol features high speed, low latency, wide coverage, etc., and can meet the system's requirements for data transmission. During the blood collection process, the communication feedback unit packs data such as the pressure distribution signal, contact area signal, blood flow velocity signal, and the abnormal probability of the blood collection status output by the abnormal determination unit, and uploads it to the background server through the encrypted communication link. At the same time, when an abnormal situation occurs in the system, the communication feedback unit will upload the exception log together, facilitating the background management personnel to monitor and analyze the system operation status.

[0058] The operation steps of the self-check and calibration unit include: Before the blood collection starts, trigger the self-check process through the tourniquet terminal, traverse the output consistency of the distributed pressure sensor array, and eliminate abnormal sensor nodes. The self-check program will sequentially send detection signals to each node of the distributed pressure sensor array, obtain its output value, and compare it with the preset standard value. If the output value of a certain sensor node exceeds the allowable error range, the system will determine that node as an abnormal node and eliminate it to ensure the accuracy of the subsequent collected pressure distribution signal.

[0059] Calibrate the baseline parameters of the infrared optical sensor to ensure the accuracy of the contact area signal. During the long-term use of the infrared optical sensor, its baseline parameters may drift, affecting the measurement accuracy of the contact area signal. Therefore, during the self-check process, the baseline parameters of the infrared optical sensor are calibrated through specific calibration devices and algorithms to restore it to the best working state, ensuring that the collected contact area signal is true and reliable.

[0060] Verify the signal stability of the Doppler blood flow meter, generate a self-check report, and feedback it to the background server. During the self-check, the Doppler blood flow meter will simulate the actual working state, emit ultrasonic waves and receive reflected signals, and evaluate its stability through signal analysis and processing. If the signal stability of the Doppler blood flow meter meets the requirements, the system will generate a self-check report and upload the report to the background server through the communication feedback unit; if there are problems, the system will prompt the operator to perform corresponding maintenance and adjustments.

[0061] Embodiment 5: This embodiment mainly describes the hardware structure composition of the intelligent monitoring and warning blood collection tourniquet, demonstrating how it collaborates with each unit of the system to achieve the intelligent monitoring and warning functions during the blood collection process.

[0062] The intelligent monitoring and warning blood collection tourniquet is applied to the system of the present invention, and its hardware structure is as follows: Elastic Ring-shaped Body: The surface of the elastic ring-shaped body is covered with an antibacterial coating, which can effectively prevent bacteria from growing and ensure the hygienic safety of patients during blood collection. Inside, a distributed pressure sensor array, an infrared optical sensor, and a Doppler blood flow meter connected to the data acquisition unit are embedded. The node density of the distributed pressure sensor array is 4 sensing units per square centimeter, and the spatial resolution matches the dimension requirements of the spatial distribution feature matrix in the state analysis unit to ensure accurate acquisition of the pressure distribution signal. The infrared optical sensor and the Doppler blood flow meter are also closely connected to the data acquisition unit to collect the contact area signal and the blood flow velocity signal in real time.

[0063] Airbag Grading Adjustment Module: The airbag grading adjustment module electrically connected to the control execution unit includes 6 independent air chambers distributed along the circumferential direction of the ring. Each air chamber realizes gradient pressure adjustment through a micro air pump and an electromagnetic valve. According to the tightening adjustment instruction sent by the control execution unit, the airbag pressure is accurately adjusted to achieve automatic adjustment of the tightness of the tourniquet. For example, during blood collection, when it is necessary to increase the pressure of the tourniquet, the control execution unit will control the micro air pump of the corresponding air chamber to inflate the air chamber, and adjust the inflation volume and inflation speed through the electromagnetic valve to achieve precise control of the airbag pressure.

[0064] Electromagnetic Lock: The electromagnetic lock integrated at the end of the elastic ring-shaped body triggers a spring-type quick separation mechanism in response to an emergency release instruction. When the control execution unit receives an emergency release instruction, the electromagnetic lock will act quickly, triggering the spring-type quick separation mechanism to quickly loosen the tourniquet and ensure the safety of the patient. The design of the electromagnetic lock ensures reliable release of the tourniquet in case of an emergency, while maintaining the stable connection of the tourniquet during normal use.

[0065] Detachable Communication Module: The detachable communication module is built-in with a low-power Bluetooth chip and a 5G NB-IoT antenna array that match the communication feedback unit. The low-power Bluetooth chip realizes short-range communication with the tourniquet terminal, and the 5G NB-IoT antenna array is used to establish a remote communication link with the background server. The design of the detachable communication module facilitates users to replace or upgrade when needed to ensure the normal operation of the communication function.

[0066] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "including", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0067] Although embodiments of the present invention have been shown and described, those of ordinary skill in the art will appreciate that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent monitoring and warning blood pressure cuff system, characterized in that, Including: A data acquisition unit, configured to obtain in real time the pressure distribution signal, contact area signal, and blood flow velocity signal of the tourniquet during blood collection; A status analysis unit, configured to dynamically match the pressure distribution signal, contact area signal, and blood flow velocity signal with preset tightness parameters and blood flow expectation parameters through an adaptive adjustment model, and generate a pressure deviation coefficient, a contact deviation coefficient, and a blood flow deviation coefficient; An abnormality determination unit, configured to construct a dynamic fusion matrix based on the pressure deviation coefficient, contact deviation coefficient, and blood flow deviation coefficient, input it into a pre-trained status classification network, and output the probability of abnormal blood collection status; A control execution unit, configured to generate a tightness adjustment instruction, an emergency release instruction, or an alarm trigger instruction when the probability of abnormal blood collection status is greater than or equal to a preset threshold; A communication feedback unit, configured to transmit the tightness adjustment instruction, emergency release instruction, or alarm trigger instruction to the tourniquet terminal and the background server; Among them, the execution steps of the status analysis unit include: Performing spatio-temporal decomposition on the pressure distribution signal, and extracting pressure time-domain features and spatial distribution features; Based on the pressure time-domain features and spatial distribution features, and in combination with preset tightness parameters, calculating a dynamic pressure deviation coefficient through a sliding window algorithm; Fusing the time series of the contact area signal and the blood flow velocity signal to generate a contact-blood flow coupling coefficient, as an input parameter for the contact deviation coefficient and the blood flow deviation coefficient.

2. The system according to claim 1, wherein The execution steps of the data acquisition unit include: Collecting the pressure distribution signal through a distributed pressure sensor array embedded in the tourniquet; Obtaining the contact area signal through an infrared optical sensor; Real-time monitoring the blood flow velocity signal through a Doppler blood flow meter; Normalizing the pressure distribution signal, contact area signal, and blood flow velocity signal into data streams with a unified sampling frequency.

3. The system according to claim 2, wherein The steps of constructing a dynamic fusion matrix include: Normalizing the dynamic pressure deviation coefficient, contact deviation coefficient, and blood flow deviation coefficient into three-dimensional vectors; Generating fusion matrix eigenvalues according to the weight ratio of each component in the three-dimensional vector; Performing noise suppression on the fusion matrix eigenvalues through a Kalman filter algorithm to obtain an optimized dynamic fusion matrix.

4. The system according to claim 2, wherein The construction steps of the status classification network include: Collect the pressure distribution signal, contact area signal, and blood flow velocity signal in the historical blood collection data, and label the corresponding blood collection status label , contact area signal and blood flow velocity signal , and label the corresponding blood collection status label ; Extract the time-domain feature vector of the pressure distribution signal and the spatial distribution feature matrix , where is the pressure value at the time point, is the pressure variance, is the sampling window length; is the partial derivative of the pressure value in the axis direction, is the partial derivative of the pressure value in the axis direction, is the matrix dimension parameter, representing the grid layout resolution of the distributed pressure sensor array on the surface of the blood pressure cuff; Constructing a contact-blood flow coupling coefficient calculation function: Among them, is the contact-blood flow coupling coefficient, are the weight factors of the contact area deviation and the blood flow velocity deviation respectively, is the historical contact area signal, which refers to the measured area value of the contact area between the finger pressure cuff and the skin in the time series, and is dynamically captured by an infrared optical sensor, is the preset contact area threshold, is the historical blood flow velocity signal, which refers to the measured value of the blood flow rate in the target blood vessel during blood collection, and is collected by a Doppler blood flow meter in pulse wave mode, is the expected value of blood flow velocity, is the normalization operator; Using , and to form a training sample set , using as supervision data to train the weight parameters of a convolutional long short-term memory hybrid neural network, generating the state classification network.

5. The system according to claim 1, wherein The execution steps of the control execution unit further include: When generating the tightness adjustment instruction, according to the priority ranking of the pressure deviation coefficient, calling the airbag hierarchical adjustment module of the tourniquet to adjust the airbag pressure in gradients; When generating the emergency release instruction, triggering the electromagnetic lock quick release mechanism, and resetting the initial state of the tourniquet after release; When generating the alarm trigger instruction, starting the audible and visual alarm device, and sending an emergency status code to the background server.

6. The system according to claim 5, wherein The steps of generating the priority ranking include: Calculating a weight factor according to the contribution degree of each component in the dynamic fusion matrix; Combining the weight factor with real-time blood collection stage parameters, and generating a priority sequence of the pressure deviation coefficient through a fuzzy logic algorithm.

7. The system according to claim 1, wherein The execution steps of the communication feedback unit include: Transmitting instructions to the tourniquet terminal through a low-power Bluetooth module; Establish an encrypted communication link with the background server through the 5G narrowband Internet of Things protocol, and upload blood collection status data and abnormal logs in real time.

8. The system according to claim 1, characterized in that, It also includes: A self-check and calibration unit, which is used to trigger a self-check process through the tourniquet terminal before blood collection starts, including: Traversing the output consistency of the distributed pressure sensor array, and eliminating abnormal sensor nodes; Calibrating the baseline parameters of the infrared optical sensor to ensure the accuracy of the contact area signal; Verifying the signal stability of the Doppler blood flow meter, generating a self-check report and feeding it back to the background server.

9. An intelligent monitoring and warning blood pressure cuff, characterized in that, Applied to the system according to any one of claims 1-8, including: An elastic ring-shaped body, with an antibacterial coating on the surface, and a distributed pressure sensor array, an infrared optical sensor, and a Doppler blood flow meter connected to the data acquisition unit are embedded inside; An airbag grading adjustment module electrically connected to the control execution unit, which includes 6 independent air chambers distributed along the circumferential direction. Each air chamber realizes gradient pressure adjustment through a micro air pump and a solenoid valve; An electromagnetic lock integrated at the end of the elastic ring-shaped body, which triggers a spring-type quick separation mechanism in response to the emergency release instruction; A detachable communication module, which is built-in with a low-power Bluetooth chip and a 5G NB-IoT antenna array that matches the communication feedback unit; Among them, the node density of the distributed pressure sensor array is 4 sensing units per square centimeter, and the spatial resolution matches the dimension requirements of the spatial distribution feature matrix in the state analysis unit.

10. An intelligent monitoring and early warning blood collection method, characterized in that, For the system according to any one of claims 1-8, the method includes: Real-time collecting the pressure distribution signal, contact area signal and blood flow velocity signal of the tourniquet; Calculating the dynamic pressure deviation coefficient, contact deviation coefficient and blood flow deviation coefficient through an adaptive adjustment model; Constructing a dynamic fusion matrix and inputting it into the state classification network to output the abnormal probability of the blood collection state; Triggering tightness adjustment, emergency release or alarm operations according to the abnormal probability; Among them, the construction steps of the adaptive adjustment model include: Based on different blood collection stages, dynamically adjusting the expected threshold of the tightness parameter; Adopting a multi-objective optimization algorithm to balance the constraint conditions of pressure uniformity, blood flow stability and contact fit.

Images