Target body temperature management method and system and medium

Abstract

The invention relates to a target body temperature management method and system and a medium, relates to the technical field of medical information, and is used for target body temperature management of coma patients after sudden cardiac arrest resuscitation. According to the scheme, physiological data including the body temperature, the electrocardiogram, the oxyhemoglobin saturation and the blood pressure of a patient are collected in real time, the physiological data are input into a target body temperature management model to obtain a body temperature management strategy, and target body temperature management of a patient object is achieved based on the body temperature management strategy. Wherein the target body temperature management neural network is obtained by performing experience playback training on neural network training based on reinforcement learning. Compared with the prior art, a more accurate target body temperature management strategy can be obtained based on the change of the physiological parameters inducting the body temperature change of the sudden cardiac arrest patient.

Description

Technical Field

[0001] The present invention relates to the field of medical information technology, and particularly relates to a method, a system and a medium for targeted temperature management. Background Art

[0002] After cardiopulmonary resuscitation of patients with cardiac arrest, although spontaneous circulation is restored, they are often in a coma state and targeted temperature management needs to be implemented. Targeted Temperature Management (TTM) has been proven to reduce nerve damage and improve the prognosis of these patients. The existing targeted temperature management methods mainly rely on temperature control devices (such as cooling blankets, temperature control instruments) and manual adjustment by medical staff, and there are problems such as response lag, inaccurate adjustment, and lack of personalized treatment plans. Summary of the Invention

[0003] The technical problem to be solved by this application is to provide a method, a system and a medium for targeted temperature management, which have the characteristics of obtaining a more accurate targeted temperature management strategy based on the changes in physiological parameters that are the inducements for the temperature changes of patients with cardiac arrest.

[0004] In a first aspect, in one embodiment, a method for targeted temperature management is provided for targeted temperature management of comatose patients after cardiac arrest resuscitation; the method includes: Collecting patient physiological data in real time, and inputting the physiological data into a targeted temperature management model to obtain a temperature management strategy, so as to achieve targeted temperature management of a patient object based on the temperature management strategy; the physiological data includes the temperature, electrocardiogram, blood oxygen saturation and blood pressure of the patient object; the targeted temperature management model is a targeted temperature management neural network obtained by training by combining the physiological data as input with a loss function, and the training method of this targeted temperature management neural network includes: Constructing and training the targeted temperature management neural network based on a reinforcement learning algorithm, including: Defining a state s(t) space based on physiological data as the environmental state; where t represents time; Defining an action a(t) space, and defining all possible control operations for adjusting the temperature that can be taken at each moment; Constructing a reward function R(s(t), a(t)), giving a positive reward when the collected temperature is within a first preset range of the target temperature, and giving a negative reward when the collected temperature is not within the first preset range of the target temperature; the first preset range includes the target temperature; Constructing a neural network infrastructure, including an input layer, a hidden layer and an output layer, where the neurons of the input layer correspond to the state space vectors of the physiological data, and the neurons of the output layer correspond to the action space vectors; Initialize the experience replay buffer for storing the experiences of the target body temperature management interacting with the physiological data environment; set the hyperparameters for training, including the discount factor, learning efficiency, and exploration rate, and perform a training loop based on the reinforcement learning algorithm to obtain the target body temperature management neural network.

[0005] In one embodiment, the state s(t) space defined based on physiological data is used as the environmental state, including: Among them, represents body temperature, represents electrocardiogram, represents blood oxygen saturation, represents blood pressure.

[0006] In one embodiment, the defined action a(t) space includes: Among them, represents increasing the power of the heating device, represents decreasing the power of the heating device, represents increasing the power of the cooling device, represents decreasing the power of the cooling device.

[0007] In one embodiment, the target body temperature is 34°C, and the first preset temperature range is [32°C, 36°C].

[0008] In one embodiment, when the collected body temperature is not within the first preset range of the target body temperature, a negative reward is given, including: when the collected body temperature is within the second preset range, a first negative reward is given, and when the collected body temperature is not within the second preset range, a second negative reward is given; the second negative reward is less than the first negative reward, and the second preset range includes a preset range below the lower limit of the first preset range and a preset range above the upper limit.

[0009] In one embodiment, when the target body temperature is 34°C and the first preset temperature range is [32°C, 36°C], the second preset temperature range is [31°C, 32°C) and (36°C, 37°C].

[0010] In a second aspect, in one embodiment, a target body temperature management system is provided for implementing the target body temperature management method described in any one of the above embodiments. The system includes a central processing unit, a data acquisition module, and a temperature control module; among them, The data acquisition module includes a temperature sensor, an electrocardiogram monitor, a blood oxygen saturation monitor, and a blood pressure monitor; the temperature sensor is used to collect the real-time body temperature of the patient object and report it to the central processing unit, the electrocardiogram monitor is used to collect the electrocardiogram of the patient object in real time and report it to the central processing unit, the blood oxygen saturation monitor is used to collect the blood oxygen saturation of the patient object in real time and report it to the central processing unit, and the blood pressure monitor is used to collect the blood pressure of the patient object in real time and report it to the central processing unit; The central processing unit is used to receive the physiological data of the patient object collected by the data acquisition module, obtain a body temperature management strategy through the target body temperature management model based on the physiological data, and control the temperature control module based on the obtained body temperature management strategy to achieve the target body temperature management of the patient object; The temperature control module includes a heating unit and a cooling unit, and the powers of the heating unit door and the cooling unit are adjustable.

[0011] In one embodiment, the system further includes an early warning module, which includes a low-temperature early warning unit and a high-temperature early warning unit. Low-temperature early warning is performed when the body temperature of the collected patient object is lower than the first preset temperature threshold, and high-temperature early warning is performed when the body temperature of the collected patient object is higher than the second preset temperature threshold; the first preset temperature threshold is less than the second preset temperature threshold.

[0012] In one embodiment, the system further includes a multi-dimensional data presentation module, which presents the physiological data of the patient object collected in real time in the form of icons and / or curves, and also includes presenting the current working states of the heating unit and the cooling unit.

[0013] In a third aspect, in one embodiment, a computer-readable storage medium is provided, and a program is stored in the medium, and the program can be loaded and executed by a processor to perform the target body temperature management method described in any one of the above embodiments.

[0014] The beneficial effects of the present invention are: Since the physiological parameters including body temperature, electrocardiogram, blood oxygen saturation, and blood pressure collected are input into the target body temperature management model to obtain a body temperature management strategy to achieve target body temperature management, it is possible to obtain a more accurate target body temperature management strategy based on the changes in the physiological parameters that are the inducements for the body temperature change of cardiac arrest patients. Since the experience replay training is carried out based on the reinforcement learning algorithm, and the neural network is constructed and trained to obtain the target body temperature management model, it is possible to solve the technical problems of insufficient training data and experience. Description of the Drawings

[0015] Figure 1 It is a schematic structural diagram of a target body temperature management system according to an embodiment of the present application; Figure 2It is a schematic flowchart of the target body temperature management method according to an embodiment of the present application.

[0016] In the figure, 01 is the central processing unit, 02 is the data acquisition module, 0201 is the temperature sensor, 0202 is the electrocardiogram monitor, 0203 is the blood oxygen saturation monitor, 0204 is the blood pressure monitor, 03 is the temperature control module, 0301 is the heating unit, and 0302 is the cooling unit. Specific embodiments

[0017] The present invention will be further described in detail below in conjunction with the accompanying drawings through specific embodiments. Similar elements in different embodiments are labeled with related similar element numbers. In the following embodiments, many details are described to enable a better understanding of the present application. However, those skilled in the art can easily recognize that some of the features can be omitted in different situations, or can be replaced by other elements, materials, and methods. In some cases, some operations related to the present application are not shown or described in the specification to avoid the core part of the present application being overwhelmed by excessive description. For those skilled in the art, it is not necessary to describe these related operations in detail, and they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0018] In addition, the features, operations, or characteristics described in the specification can be combined in any appropriate manner to form various embodiments. At the same time, the steps or actions in the method description can also be reordered or adjusted in an obvious manner by those skilled in the art. Therefore, the various sequences in the specification and drawings are only for clearly describing a certain embodiment and do not mean that they are the necessary sequences, unless it is stated that a certain sequence must be followed.

[0019] The serial numbers assigned to the components herein, such as "first", "second", etc., are only used to distinguish the described objects and do not have any sequential or technical meanings.

[0020] For the convenience of explaining the inventive concept of the present application, the following briefly describes the target hypothermia management technology.

[0021] In current target body temperature management, most attention is paid to the research on temperature control devices (including heating devices and cooling devices), and the reference basis for controlling the temperature control devices is only the body temperature of the patient collected for reference, and very few studies are conducted on other physiological parameters that affect temperature changes. However, the applicant found in the research that studying the physiological parameters that cause body temperature changes is more beneficial for target body temperature management and can obtain a more accurate target body temperature management plan.

[0022] In view of this, embodiments of the present application provide a target body temperature management method, system, and medium, which are applicable to the target body temperature management of comatose patients after cardiac arrest resuscitation. In this solution, on the one hand, physiological parameters of the patient, including body temperature, electrocardiogram, blood oxygen saturation, and blood pressure, are collected in real time, and the collected physiological parameters are input into the target body temperature management model to obtain a body temperature management strategy, so as to achieve target body temperature management. This enables a more accurate target body temperature management strategy to be obtained based on the changes in the physiological parameters (electrocardiogram, blood oxygen saturation, and blood pressure) that induce changes in the body temperature of cardiac arrest patients. On the other hand, based on the reinforcement learning algorithm, experience replay training is performed to construct and train a neural network to obtain the target body temperature management model, thereby solving the technical problems of insufficient training data and experience.

[0023] To facilitate a better understanding of the solution of the present application, the target body temperature management system of the embodiments of the present application will be introduced first below.

[0024] In an embodiment of the present application, a target body temperature management system is provided. Please refer to Figure 1 , which includes a central processing unit 01, a data acquisition module 02, and a temperature control module 03. Among them, the central processing unit 01 is used to receive the physiological data collected by the data acquisition module 02, and based on these physiological data, obtain a body temperature management strategy through the target body temperature management model, so as to control the temperature control module 03 based on the obtained body temperature management strategy to achieve target body temperature management of the patient object.

[0025] In one embodiment, the collected physiological data includes the body temperature, electrocardiogram, blood oxygen saturation, and blood pressure of the patient object. Correspondingly, for the data acquisition module 02, it may include a temperature sensor 0201, an electrocardiogram monitor 0202, a blood oxygen saturation monitor 0203, and a blood pressure monitor 0204. The temperature sensor 0201 is used to collect the real-time body temperature of the patient object and report it to the central processing unit 01, the electrocardiogram monitor 0202 is used to collect the electrocardiogram of the patient object in real time and report it to the central processing unit 01, the blood oxygen saturation monitor 0203 is used to collect the blood oxygen saturation of the patient object in real time and report it to the central processing unit 01, and the blood pressure monitor 0204 is used to collect the blood pressure of the patient object in real time and report it to the central processing unit 01.

[0026] The central processing unit 01 includes a processor, which can call the target management model program stored on the storage medium to obtain a body temperature management strategy, or send the received physiological data to other processors to obtain a body temperature management strategy from other processors.

[0027] The temperature control module 03 is used to perform temperature control based on the temperature management strategy given by the central processing unit 01. The temperature control module 03 includes a heating unit 0301 and a cooling unit 0302. The heating unit 0301 and the cooling unit 0302 can adopt existing conventional heating devices and cooling devices such as heating blankets, heating pads, cooling blankets, and cooling pads, which will not be elaborated here.

[0028] In one embodiment, both the heating unit 0301 and the cooling unit 0302 have adjustable power.

[0029] Those skilled in the art can understand that Figure 1 the hardware structure of the target body temperature management system shown in does not constitute a limitation on the target body temperature management system, and may include more or fewer components than shown, or combine some components, or have different component arrangements.

[0030] In one embodiment, the target body temperature management system further includes an early warning module. The early warning module includes a low-temperature early warning unit and a high-temperature early warning unit, which performs low-temperature early warning when the collected body temperature of the patient object is lower than the first preset temperature threshold, and performs high-temperature early warning when the collected body temperature of the patient object is higher than the second preset temperature threshold. The first preset temperature threshold is less than the second preset temperature threshold.

[0031] In one embodiment, the first preset temperature threshold can be 32 °C, and the second preset temperature threshold can be 36 °C. Thus, when the patient's body temperature deviates from the target range (32 °C - 36 °C), the system immediately gives an early warning and takes corresponding measures for adjustment. If the patient's body temperature is lower than 32 °C, the heating unit 0301 is started, and the heating intensity is automatically adjusted according to the rising speed of the body temperature. If the rising speed of the body temperature is greater than the preset first temperature rising speed threshold, the heating power of the heating unit 0301 can be reduced to avoid complications caused by too rapid a rise in body temperature. If the body temperature rises too slowly, lower than the preset second temperature rising speed threshold, the heating power of the heating unit 0301 can be increased to ensure the temperature control effect. The first temperature rising speed threshold is greater than the second temperature rising speed threshold. If the patient's body temperature is higher than 32 °C, the cooling unit 0302 is started, and the cooling intensity is automatically adjusted according to the falling speed of the body temperature. If the falling speed of the body temperature is greater than the preset first temperature falling speed threshold, the cooling power of the cooling unit 0302 can be reduced. If the body temperature falls too slowly, lower than the preset second temperature falling speed threshold, the cooling power of the cooling unit 0302 can be increased to ensure the temperature control effect. The first temperature falling speed threshold is greater than the second temperature falling speed threshold.

[0032] In one embodiment, the target temperature management system further includes a multi-dimensional data presentation module that presents the physiological data of the patient object collected in real time in the form of icons and / or curves, and also presents the current working states of the heating unit 0301 and the cooling unit 0302.

[0033] In one embodiment, medical staff can remotely monitor the patient's body temperature based on the multi-dimensional data presentation module.

[0034] In one embodiment of the present application, a target temperature management method is provided, which can be implemented based on the target temperature management system of any of the above embodiments. The method may include: collecting the patient's physiological data in real time, and inputting the collected physiological data into the target temperature management model to obtain a temperature management strategy, so as to implement the target temperature management of the patient object based on this temperature management strategy. Among them, the target temperature management model is a target temperature management neural network obtained by training with physiological data as the input and combining a loss function.

[0035] For patients with cardiac arrest, controlling the body temperature between 32°C and 36°C after the recovery of cardiac arrest can effectively reduce brain damage. The change in body temperature directly reflects the effect of temperature management. Therefore, it is the core of optimizing the target temperature management strategy. However, the applicant found in the research that the electrocardiogram records the electrical activity of the heart and reflects the health status of the heart. After cardiac arrest resuscitation, the heart function may not fully recover, and information such as heart rate and rhythm can help judge the status of circulatory recovery after resuscitation. Abnormal cardiac activity may affect the temperature management strategy. For example, rapid arrhythmia may lead to temperature management disorders. Oxygen supply is very important for maintaining normal body temperature. Insufficient oxygen may lead to difficulties in temperature regulation. Therefore, blood oxygen saturation also has a great influence on the development trend of body temperature. Blood pressure reflects the state of the human blood circulation system. If the blood pressure is too low, the body temperature cannot be effectively distributed due to insufficient blood circulation. Therefore, the regulation of body temperature will also be greatly affected.

[0036] Therefore, in the embodiment of the present application, based on physiological parameters including body temperature, electrocardiogram, blood oxygen saturation, and blood pressure, target temperature management is performed on the patient object. Since target temperature management can be performed based on the inducement of body temperature change, which is a potential trend management, the target temperature management of the patient object can be more accurately achieved.

[0037] However, the applicant found in the research that in the current target temperature management, few other physiological parameters that affect temperature change are studied. Therefore, it is difficult to obtain training samples of the correlation between physiological parameters including body temperature, electrocardiogram, blood oxygen saturation, and blood pressure and the temperature management strategy.

[0038] In view of this, in the embodiments of the present application, a reinforcement learning algorithm is used to improve data utilization and stabilize the training process, thereby overcoming the problem of insufficient training samples.

[0039] In one embodiment, please refer to Figure 2 , and a target body temperature management neural network is constructed and trained based on a reinforcement learning algorithm, including: Step S10, defining a state s(t) space based on physiological data as the environmental state.

[0040] The state space defines the environmental information that can be perceived at each moment. In one embodiment, step S10 can be expressed as: where t represents time, represents body temperature, represents electrocardiogram, represents blood oxygen saturation, represents blood pressure.

[0041] Step S20, defining an action a(t) space.

[0042] The action space defines all possible control operations for adjusting the temperature that can be taken at each moment to obtain a temperature control strategy. In one embodiment, the action space is defined based on increasing the power of the heating device, decreasing the power of the heating device, increasing the power of the cooling device, and decreasing the power of the cooling device. Then step S20 can be expressed as: where, represents increasing the power of the heating device, represents decreasing the power of the heating device, represents increasing the power of the cooling device, represents decreasing the power of the cooling device.

[0043] Step S30, constructing a reward function R(s(t), a(t)), giving a positive reward when the collected body temperature is within the first preset range of the target body temperature, and giving a negative reward when the collected body temperature is not within the first preset range of the target body temperature. Among them, the first preset range includes the target body temperature.

[0044] In one embodiment, the target body temperature is 34°C, and the first preset temperature range is [32°C, 36°C]. That is, in the first preset temperature range, 32°C is the lower limit of the first preset temperature range, and 36°C is the upper limit of the first preset temperature range.

[0045] The reward function is used to evaluate the effect of the current action, so as to reflect whether the current body temperature adjustment is effective. A greater reward can be given when the feedback body temperature is closer to the target body temperature. For example, the positive reward for the collected body temperature within the first preset range [32°C, 36°C] of the target body temperature can be divided into multiple positive rewards. For example, if the body temperature is within the range of [33.5°C, 34.5°C], a positive reward of 1 is given; if the body temperature is within the range of [33°C, 33.5°C) or (34.5°C, 35°C], a positive reward of 0.6 is given; if the body temperature is within the range of [32.5°C, 33°C) or (35°C, 35.5°C], a positive reward of 0.3 is given; if the body temperature is within the range of [32°C, 32.5°C) or (35.5°C, 36°C], a positive reward of 0.1 is given. In this way, a more precise feedback control mechanism is obtained. Specifically, it can be set based on actual requirements.

[0046] In one implementation, when the collected body temperature is not within the first preset range of the target body temperature, a negative reward is given, including: when the collected body temperature is within the second preset range, a first negative reward is given, and when the collected body temperature is not within the second preset range, a second negative reward is given. Among them, the second negative reward is less than the first negative reward, and the second preset range includes a preset range below the lower limit of the first preset range and a preset range above the upper limit.

[0047] In one embodiment, the second preset temperature range is [31°C, 32°C) and (36°C, 37°C], that is, if the collected body temperature is within the range of [31°C, 32°C) or (36°C, 37°C], a first negative reward is given, but if the collected body temperature is not within the second preset range, but less than 31°C or greater than 37°C, a second negative reward is given.

[0048] In one embodiment, the first negative reward can be set to -0.5, and the second negative reward can be set to -1.

[0049] Step S40, construct a neural network infrastructure, including an input layer, a hidden layer, and an output layer. Among them, the neurons in the input layer correspond to the state space vector of physiological data, and the neurons in the output layer correspond to the action space vector.

[0050] In one embodiment, the input layer has 4 neurons, corresponding to 4 dimensions of the state space vector, including body temperature, electrocardiogram, blood oxygen saturation, and blood pressure. The output layer has 4 neurons, equal to the number of actions, including increasing the power of the heating device, decreasing the power of the heating device, increasing the power of the cooling device, and decreasing the power of the cooling device. Each neuron outputs the value evaluation (Q value) of the corresponding action. Multiple hidden layers can be set between the input layer and the output layer. In one embodiment, two hidden layers can be set, and each hidden layer can include 64 neurons. The activation function can select the ReLU activation function.

[0051] Step S50: Initialize the experience replay buffer for storing the experiences of the target body temperature management interacting with the physiological data environment; set the hyperparameters for training, including the discount factor, learning efficiency, and exploration rate, and perform a training loop based on the reinforcement learning algorithm to obtain the target body temperature management neural network.

[0052] Discount factor ( ) It is used to balance the relationship between the current reward and future rewards. Since reinforcement learning needs to consider the long-term effects in the future, the discount factor can enable the system to balance the weights between the current decision and future decisions. The closer the discount factor is to 1, it indicates that the system pays more attention to future long-term rewards; the closer the discount factor is to 0, it indicates that the system pays more attention to the current reward.

[0053] If the body temperature management strategy not only considers the current body temperature adjustment but also the long-term body temperature stability, choosing a larger discount factor (such as 0.9) will help the system learn how to balance the short-term and long-term body temperature control effects. Those skilled in the art can understand that an appropriate discount factor can be selected so that considering the current body temperature control strategy may affect subsequent body temperature changes and make more rational decisions, thus avoiding the drastic fluctuations in body temperature caused by only focusing on short-term effects.

[0054] In one embodiment, to pay more attention to the long-term effect and maintain the long-term stability of the body temperature through smooth adjustment, the discount factor = 0.8. Thus, based on the inducement of the body temperature change, the target body temperature management can be carried out to achieve potential trend management, so that the target body temperature management of the patient object can be realized more precisely.

[0055] In one embodiment, the mean squared error loss function can be used as the loss function for training this neural network.

[0056] The target body temperature management neural network constructed and trained based on any of the above embodiments can, through the reinforcement learning algorithm, autonomously explore the body temperature control strategy, gradually reduce manual intervention, and provide an automation level. Through learning historical data and feedback, the system can avoid excessive fluctuations in body temperature, thereby reducing interference to the patient. Based on the obtained model, the power output of the temperature control device is dynamically adjusted according to the real-time monitored data to keep the body temperature of the patient object within the target range. It not only overcomes the defect of insufficient samples but also can more precisely achieve the target body temperature management of the patient object.

[0057] For the target body temperature management method based on any of the above embodiments, since it is based on physiological parameters including body temperature, electrocardiogram, blood oxygen saturation, and blood pressure collected and input into the target body temperature management model to obtain a body temperature management strategy for achieving target body temperature management, it is possible to obtain a more accurate target body temperature management strategy based on the changes in the physiological parameters that are the inducements for the body temperature changes in cardiac arrest patients. Since it is based on the reinforcement learning algorithm and a neural network is constructed and trained to obtain the target body temperature management model, it is possible to solve the technical problems of insufficient training data and experience.

[0058] In one embodiment, in order to accelerate the reinforcement learning process, on the one hand, historical data can be used for simulation training to accelerate the learning process of the algorithm. On the other hand, the management of multiple patients can be trained in parallel to improve the data utilization efficiency and shorten the algorithm training time. Through these technical means, the response speed and accuracy of the target body temperature management system can be further improved to ensure that the patient's body temperature always remains within the optimal range after successful resuscitation.

[0059] In one embodiment of the present application, a computer-readable storage medium is provided, and a program is stored on the storage medium. The stored program includes the method that can be loaded and processed by a processor in any of the above embodiments.

[0060] Those skilled in the art can understand that all or part of the functions of the above methods can be implemented in a hardware manner or in a computer program manner. When all or part of the functions in the above embodiments are implemented in a computer program manner, the program can be stored in a computer-readable storage medium. The storage medium may include: read-only memory, random access memory, magnetic disk, optical disk, hard disk, etc. The above functions can be implemented by a computer executing the program. For example, the program is stored in the memory of the device, and when the processor executes the program in the memory, the above all or part of the functions can be achieved. In addition, when all or part of the functions in the above embodiments are implemented in a computer program manner, the program can also be stored in a storage medium such as a server, another computer, magnetic disk, optical disk, flash drive, or mobile hard disk, downloaded or copied and saved to the memory of the local device, or the system of the local device is updated in version. When the processor executes the program in the memory, the above all or part of the functions in the embodiments can be achieved.

[0061] The above uses specific examples to elaborate on the present invention, which is only used to help understand the present invention and is not intended to limit the present invention. For those skilled in the technical field to which the present invention belongs, according to the idea of the present invention, several simple deductions, deformations, or substitutions can also be made.

Claims

1. A method for targeted temperature management, characterized in that, Target temperature management for comatose patients after cardiac arrest resuscitation; the method includes: Real-time collection of patient physiological data, and inputting the physiological data into a target temperature management model to obtain a temperature management strategy, so as to achieve the target temperature management of a patient object based on the temperature management strategy; the physiological data includes the body temperature, electrocardiogram, blood oxygen saturation, and blood pressure of the patient object; the target temperature management model is a target temperature management neural network obtained by training with the physiological data as input in combination with a loss function, and the training method of the target temperature management neural network includes: Constructing and training the target temperature management neural network based on a reinforcement learning algorithm, including: Defining a state s(t) space based on physiological data as the environmental state; where t represents time; Defining an action a(t) space, defining all possible control operations for adjusting the temperature that can be taken at each moment; Constructing a reward function R(s(t), a(t)), giving a positive reward when the collected body temperature is within the first preset range of the target temperature, and giving a negative reward when the collected body temperature is not within the first preset range of the target temperature; the first preset range includes the target temperature; Constructing a neural network infrastructure, including an input layer, a hidden layer, and an output layer, where the neurons in the input layer correspond to the state space vectors of physiological data, and the neurons in the output layer correspond to the action space vectors; Initializing an experience replay buffer for storing the experience of the interaction between target temperature management and the physiological data environment; setting the hyperparameters of training, including the discount factor, learning efficiency, and exploration rate, and performing training loops based on the reinforcement learning algorithm to obtain the target temperature management neural network.

2. The target body temperature management method according to claim 1, wherein The defining the state s(t) space based on physiological data as the environmental state includes: Among them, represents body temperature, represents electrocardiogram, represents blood oxygen saturation, represents blood pressure.

3. The target body temperature management method according to claim 1, wherein, The defining the action a(t) space includes: Among them, indicates increasing the power of the heating device, indicates decreasing the power of the heating device, indicates increasing the power of the cooling device, indicates decreasing the power of the cooling device.

4. The target body temperature management method according to claim 1, characterized in that The target temperature is 34°C, and the first preset temperature range is [32°C, 36°C].

5. The target body temperature management method according to claim 1 or 4, characterized in that, The giving a negative reward when the collected body temperature is not within the first preset range of the target temperature includes: giving a first negative reward when the collected body temperature is within the second preset range, and giving a second negative reward when the collected body temperature is not within the second preset range; the second negative reward is less than the first negative reward, and the second preset range includes a preset range below the lower limit of the first preset range and a preset range above the upper limit.

6. The target body temperature management method according to claim 5, wherein When the target temperature is 34°C and the first preset temperature range is [32°C, 36°C], the second preset temperature range is [31°C, 32°C) and (36°C, 37°C].

7. A target body temperature management system, characterized in that, For implementing the target temperature management method according to any one of claims 1 to 6, the system includes a central processing unit, a data collection module, and a temperature control module; where, The data acquisition module includes a temperature sensor, an electrocardiogram monitor, a blood oxygen saturation monitor, and a blood pressure monitor; the temperature sensor is used to collect the real-time body temperature of the patient object and report it to the central processing unit, the electrocardiogram monitor is used to collect the electrocardiogram of the patient object in real time and report it to the central processing unit, the blood oxygen saturation monitor is used to collect the blood oxygen saturation of the patient object in real time and report it to the central processing unit, and the blood pressure monitor is used to collect the blood pressure of the patient object in real time and report it to the central processing unit; The central processing unit is used to receive the physiological data of the patient object collected by the data acquisition module, obtain a body temperature management strategy through the target body temperature management model based on the physiological data, and control the temperature control module based on the obtained body temperature management strategy to achieve the target body temperature management of the patient object; The temperature control module includes a heating unit and a cooling unit, and the powers of the heating unit door and the cooling unit are adjustable.

8. The target body temperature management system according to claim 7, wherein The system further includes an early warning module, which includes a low-temperature early warning unit and a high-temperature early warning unit. Low-temperature early warning is carried out when the body temperature of the collected patient object is lower than the first preset temperature threshold, and high-temperature early warning is carried out when the body temperature of the collected patient object is higher than the second preset temperature threshold; The first preset temperature threshold is less than the second preset temperature threshold.

9. The target body temperature management system according to claim 7, wherein The system further includes a multi-dimensional data presentation module, which presents the physiological data of the patient object collected in real time in the form of icons and / or curves, and also includes presenting the current working states of the heating unit and the cooling unit.

10. A computer-readable storage medium, characterized in that, The medium stores a program that can be loaded and executed by a processor to perform the target body temperature management method according to any one of claims 1 to 6.

Images