A personalized cold risk prediction and assessment method and system based on deep learning

Abstract

The present application relates to the field of cold risk prediction technology, and provides a personalized cold risk prediction and assessment method and system based on deep learning. The method includes: sequentially cleaning and formatting the received characteristic parameters of the target individual and the environmental parameters of the target individual's location; based on a pre-built thermal regulation model, making a preliminary prediction of the target individual's skin temperature, and determining the target section in the thermal regulation model that requires key parameter adjustment based on the preliminary prediction results and the formatted characteristic parameters; wherein the key parameters include: metabolic rate, set point temperature, heat capacity, or one or more; based on a deep learning algorithm, cyclically adjusting the key parameters of the target section based on the formatted data; based on the thermal regulation model after the key parameters are adjusted, making a secondary prediction of the target individual's skin temperature, and determining the target individual's cold stress risk based on the secondary prediction results, and feeding back to the target individual.

Description

Technical Field

[0001] The present application relates to the technical field of cold risk prediction, and in particular to a personalized cold risk prediction and evaluation method, system, computer-readable storage medium and electronic device based on deep learning. Background Art

[0002] As the duration, frequency and intensity of low temperature climate phenomena such as cold waves, rain, snow and freezing continue to increase, emergency rescue personnel, engineering rescue personnel, special operations personnel, etc. will inevitably face extreme low temperature and freezing environments, which will not only affect work efficiency and quality, but also cause local or systemic injuries to the operators, and even cause casualties.

[0003] As an effective tool for predicting the physiological response of target individuals in extreme environments, thermal regulation models are widely used to assess heat stress and cold risk. However, most of the current thermal regulation models are developed based on Stolwijk, and the basic data used, such as the set-point temperature, the distribution of metabolic rate coefficients for each individual segment, and the body heat capacity value of each segment, are all based on the group level. There is a lack of processing, analysis and prediction of data differences between individuals, and the prediction results lack effectiveness and pertinence for different individuals.

[0004] In recent years, some analysis and prediction models that consider personalized factors such as age, gender, height, and physical signs have made individual-level analysis and prediction more accurate to a certain extent. However, due to the huge differences between individuals, especially in extremely cold environments, the differences in parts such as the trunk and limbs are greater, and accurate predictions at the individual level cannot be achieved through these analysis and prediction models.

[0005] Therefore, there is an urgent need to provide a technical solution to the above-mentioned deficiencies in the prior art. Summary of the invention

[0006] The purpose of this application is to provide a personalized cold risk prediction and assessment method and system based on deep learning to solve or alleviate the problems existing in the above-mentioned prior art.

[0007] In order to achieve the above objectives, this application provides the following technical solutions:

[0008] The present application provides a personalized cold risk prediction and assessment method based on deep learning, including: step S101, cleaning and formatting the received characteristic parameters of the target individual and the environmental parameters of the location of the target individual in sequence; step S102, based on a pre-constructed thermal regulation model, making a preliminary prediction of the skin temperature of the target individual, and determining the target section in the thermal regulation model that requires key parameter adjustment based on the preliminary prediction result and the formatted characteristic parameters; wherein the key parameters include: metabolic rate, set point temperature, heat capacity, or one or more; step S103, based on a deep learning algorithm, cyclically adjusting the key parameters of the target section according to the formatted data; step S104, based on the thermal regulation model after the key parameters are adjusted, making a secondary prediction of the skin temperature of the target individual, determining the cold stress risk of the target individual based on the secondary prediction result, and feeding back to the target individual

[0009] Preferably, in step S101, the data cleaning includes sequentially performing outlier detection, missing value processing and noise filtering on the characteristic parameters and the environmental parameters; the formatting includes sequentially performing standardization processing, feature engineering and data reconstruction on the characteristic parameters and the environmental parameters after the data cleaning.

[0010] Preferably, in step S102: based on the pre-constructed thermal regulation model, a preliminary prediction is made on the skin temperature of the target individual, and according to the error evaluation between the preliminary prediction result and the formatted characteristic parameter, the target section in the thermal regulation model where the key parameter adjustment is required is determined; wherein the thermal regulation model is:

[0011]

[0012] In the formula, is the serial number of the segment part for segmenting the target individual, ; The hierarchical sequence number for each section. ; For the The first section Heat capacity of the layer; For the The first section Temperature of the layer; The time during which the target individual is exposed to the cold environment; For the The first section Heat generation of the layer; For the The first section The blood in the layer exchanges heat;

[0013] Respectively represent The first section Layer, Layer and Conductive heat exchange between other layers of the sections; For the The first section The breathing heat transfer of the layer; The target individuals are The radiation heat, convection heat and evaporation heat exchange between each section and the external environment.

[0014] Preferably, in step S103, based on a genetic algorithm, the key parameters that need to be adjusted in the target section are adjusted cyclically and individually according to preset priorities until the error between the prediction result of the thermal regulation model and the formatted characteristic parameters is less than a preset temperature threshold.

[0015] Preferably, step S104 includes: performing a secondary prediction of the skin temperature of the target individual based on the thermal regulation model after personalized adjustment of the key parameters; in response to the wind chill temperature of the target individual being less than a preset wind chill temperature, issuing a wind chill temperature warning to the target individual; in response to the skin temperature of the target individual predicted by the thermal regulation model being less than or equal to a preset frostbite risk index, issuing a frostbite risk warning to the target individual.

[0016] Preferably, in step S104, according to the formatted environmental parameters of the location of the target individual, the formula:

[0017]

[0018] Calculate the wind chill temperature ;

[0019] in, is the air temperature, is the wind speed at the location of the target individual, .

[0020] An embodiment of the present application also provides a personalized cold risk prediction and assessment method based on deep learning, including: based on the configured wearable device, obtaining the characteristic parameters and environmental parameters of the target individual, sending them to the cloud, and receiving the cold stress risk feedback from the cloud; wherein the cloud performs personalized adjustment of key parameters of a pre-constructed thermal regulation model according to the characteristic parameters and environmental parameters, and determines the cold stress risk of the target individual based on the thermal regulation model after the key parameters are adjusted.

[0021] The embodiment of the present application also provides a personalized cold risk prediction and assessment system based on deep learning, including: a preprocessing unit, configured to perform data cleaning and formatting on the received characteristic parameters of the target individual and the environmental parameters of the location of the target individual in sequence; a target segment determination unit, configured to make a preliminary prediction of the skin temperature of the target individual based on a pre-constructed thermal regulation model, and determine the target segment in the thermal regulation model that requires key parameter adjustment based on the preliminary prediction result and the formatted characteristic parameters; wherein the key parameters include: metabolic rate, set point temperature, and heat capacity. One or more; a parameter adjustment unit, configured to cyclically adjust the key parameters of the target segment according to the formatted data based on a deep learning algorithm; a prediction feedback unit, configured to make a secondary prediction of the skin temperature of the target individual based on the thermal regulation model after the key parameters are adjusted, and determine the cold stress risk of the target individual based on the secondary prediction result, and feed back to the target individual.

[0022] An embodiment of the present application also provides a computer-readable storage medium on which a computer program is stored, wherein the computer program is any of the deep learning-based personalized cold risk prediction and assessment methods described above.

[0023] An embodiment of the present application also provides an electronic device, comprising: a memory, a processor, and a program stored in the memory and executable on the processor, wherein when the processor executes the program, the personalized cold risk prediction and assessment method based on deep learning as described above is implemented.

[0024] Beneficial effects:

[0025] The personalized cold risk prediction and assessment method based on deep learning provided in the embodiment of the present application first clarifies and formats the characteristic parameters of the received target individual and the environmental parameters of the target individual's location in sequence; then, based on the pre-constructed thermal regulation model, the skin temperature of the target individual is preliminarily predicted, and according to the preliminary prediction results and the actual monitoring values ​​(formatted characteristic parameters), the target segment in the thermal regulation model where key parameters need to be adjusted is determined; then, based on the deep learning algorithm, one or more parameters of the metabolic rate, set-point temperature, and heat capacity of the target segment are cyclically adjusted according to the formatted data; finally, based on the thermal regulation model after the key parameters are adjusted, the skin temperature of the target individual is secondary predicted, and the cold stress risk of the target individual is determined according to the secondary prediction results, and the cold stress risk is fed back to the target individual.

[0026] Therefore, through the data of the target individual, the deep learning algorithm is used to adjust and optimize the key parameters of the thermal regulation model, and a personalized thermal regulation model suitable for the target individual is generated. The skin temperature of the target individual is personalized predicted through the personalized thermal regulation model, providing a refined and accurate early warning at the individual level; and the key parameters of the personalized thermal regulation model can be dynamically adjusted and optimized as time and the environment change, thereby ensuring the accuracy and effectiveness of the prediction and preventing people from suffering from cold injuries. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The drawings constituting part of the present application are used to provide a further understanding of the present application. The exemplary embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. Among them:

[0028] Figure 1 A flowchart of a personalized cold risk prediction and assessment method based on deep learning provided according to some embodiments of the present application;

[0029] Figure 2 A schematic diagram of the structure of a personalized cold risk prediction and assessment system based on deep learning according to some embodiments of the present application;

[0030] Figure 3 A schematic diagram of the structure of an electronic device provided according to some embodiments of the present application;

[0031] Figure 4 A hardware structure diagram of an electronic device provided according to some embodiments of the present application. DETAILED DESCRIPTION

[0032] The present application will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments. Each example is provided by way of explanation of the present application and does not limit the present application. In fact, it will be clear to those skilled in the art that modifications and variations may be made in the present application without departing from the scope or spirit of the present application. For example, a feature shown or described as a part of an embodiment may be used in another embodiment to produce yet another embodiment. Therefore, it is desired that the present application includes such modifications and variations within the scope of the appended claims and their equivalents.

[0033] At present, although the thermal regulation models that take personalized factors into consideration can more accurately predict the skin temperature at the individual level to a certain extent, they are unable to accurately and effectively achieve individual-level predictions in extreme environments due to the differences between different individuals.

[0034] Based on this, this application proposes a personalized cold risk prediction and assessment method based on deep learning. Through the characteristic parameters (physiological data) of the target individual, the environmental parameters of the target individual's location, etc., the deep learning algorithm is used to personalize the key parameters in the thermal regulation model, and generate a cold risk assessment model suitable for the individual level. It provides individual-level refined formulation of outdoor work organization strategies and individual protection strategies, effectively preventing workers from suffering from cold injuries, in order to further improve outdoor work efficiency.

[0035] like Figure 1 As shown, the personalized cold risk prediction and evaluation method based on deep learning includes:

[0036] Step S101: Clean and format the received characteristic parameters of the target individual and the environmental parameters of the location where the target individual is located.

[0037] In this application, the characteristic parameters of the target individual are mainly the physiological parameters of the target individual, which are collected through the wearable device possessed by the target individual. Specifically, the heart rate and wrist skin temperature of the target individual are detected in real time through the smart watch worn by the target individual; the skin temperature of various parts of the target individual's body is monitored in real time through the skin temperature sensor.

[0038] Among them, heart rate monitoring is achieved through the built-in heart rate sensor in the smart watch and is completed through the broadband capacitance pulse wave technology; wrist skin temperature monitoring is achieved through the temperature sensor in the smart watch that is in contact with the skin. The skin temperature sensor used to monitor the skin temperature of various parts of the target individual's body can be a smart fiber attached to different skin parts of the target individual or embedded in the clothes worn by the target individual.

[0039] The environmental parameters of the target individual's location are collected in real time through a portable monitoring device (with integrated temperature sensor, humidity sensor, and wind speed sensor) carried by the target individual, and the environmental parameters are collected and transmitted in real time through a built-in power supply and wireless transmission function. Here, the portable monitoring device can be a portable meteorological monitor, a multifunctional mobile environmental monitor, a portable meteorological monitoring station, etc.

[0040] After the characteristic parameters of the target individual and the environmental parameters of the target individual's location are transmitted to the cloud, the received physiological parameters (i.e., characteristic parameters) and environmental parameters of the target individual are cleaned and formatted in the cloud. Specifically, first, the cloud server stores the received data in the database, and then performs outlier detection, missing value processing, and noise filtering on the characteristic parameters and environmental parameters in turn; finally, the characteristic parameters and environmental parameters after data cleaning are standardized, feature engineered, and reconstructed in turn, and the data is formatted into a standard form.

[0041] The cloud server stores the received data in the database, and then performs outlier detection, missing value processing, and noise filtering on the characteristic parameters and environmental parameters in turn. When performing outlier detection, first, for skin temperature and heart rate, refer to medical standards to define the abnormal range; for environmental parameters, the normal range can be determined based on historical data; then, statistical methods (such as box plots) are used to detect outliers for each parameter; then, the outliers are evaluated to decide whether to delete or retain them, where outliers caused by errors are deleted and outliers reflecting important and atypical situations are retained.

[0042] When processing missing values, first, identify the missing pattern. For example, if a sensor often loses data under certain conditions, there may be a problem with the device. Then, select a processing strategy based on the data type and analysis objectives. For key parameters (such as heart rate), consider using interpolation methods to fill in missing values. For environmental parameters, time series analysis methods can be used.

[0043] In the noise filtering process, first, the noise is identified by detecting short-term fluctuations that are significantly different from the average level; then, the identified data is processed. For example, for skin temperature and heart rate data, a moving average or median filter can be used to smooth the data.

[0044] Finally, the characteristic parameters and environmental parameters after data cleaning are standardized and reconstructed in turn, and the data are formatted into a standard form to ensure that the data is stored in a format suitable for analysis.

[0045] Step S102: Based on the pre-built thermal regulation model, a preliminary prediction is made on the skin temperature of the target individual, and according to the preliminary prediction result and the formatted characteristic parameters, a target section in the thermal regulation model where key parameters need to be adjusted is determined.

[0046] In this application, according to the physiological structure of the human body, the human body is divided into multiple sections to establish a thermal regulation model of the human body. Specifically, the human body is divided into 22 sections, including the head, face, right upper arm, left upper arm, right lower arm, left lower arm, right hand, left hand, left finger, right finger, chest, right shoulder, left shoulder, abdomen, right buttocks, left buttocks, right thigh, left thigh, right calf, left calf, right foot, and left foot. Each section is divided into skin layer, muscle layer, fat layer, and core layer. The thermal regulation models of each section and layer of the human body are as follows:

[0047]

[0048] in, is the serial number of the divided section. ; The hierarchical sequence number for each section. ; For the The first section Heat capacity of the layer; For the The first section Temperature of the layer; The time the target individual is exposed to the cold environment; For the The first section The heat production of each layer mainly includes: basal metabolic rate , Shivering Heat Production Rate and heat production rate ; For the The first section The blood in the layer exchanges heat; Respectively represent The first section Layer, Layer and Conductive heat exchange between other layers of the sections; For the The first section The breathing heat transfer of the layer; The target individuals are The radiation heat, convection heat and evaporation heat exchange between each section and the external environment.

[0049] In this application, based on the pre-built thermal regulation model, the skin temperature of the target individual is preliminarily predicted, and the target segment in the thermal regulation model that needs key parameter adjustment is determined based on the error evaluation between the preliminary prediction result and the actual monitoring value (formatted characteristic parameter). Specifically, the original parameters (i.e., initial values) of the thermal regulation model (as shown in Table 1) are used for preliminary prediction, and the predicted values ​​are compared with the actual monitoring values ​​to identify the segment where the average error is greater than 2°C. This segment is used as the target segment that needs parameter adjustment and optimization, and its metabolic rate (basal metabolic rate )、Set point temperature( )、heat capacity( ) to adjust and optimize one or more of them. Table 1 is as follows:

[0050] Table 1 Original parameters of thermal regulation model

[0051] parameter Initial Value Basal metabolic rate of muscle layer: Mmu (W / m3) 510 Basal metabolic rate of fat layer: Mfa (W / m3) 62 Basal metabolic rate of the skin layer: Msk (W / m3) 320 Thermal conductivity of the muscle layer: Kmu (Wm-1.k-1) 0.42 Thermal conductivity of fat layer: Kfa (Wm-1.k-1) 0.25 Thermal conductivity of the skin layer: Ksk (Wm-1.k-1) 0.35 Density of the muscle layer: ρmu (Kg.m-3) 1048 Density of fat layer: ρfa (Kg.m-3) 865 Density of the skin layer: ρsk (Kg.m-3) 1070 Heat capacity of core layer: Cbr (J.kg-1.K-1) 3750 Heat capacity of fat layer: Cfa (J.kg-1.K-1) 2100 Heat capacity of skin layer: Csk (J.kg-1.K-1) 3550

[0052] Step S103: Based on the deep learning algorithm, key parameters of the target segment are cyclically adjusted according to the formatted data.

[0053] In this application, the segment with an average error greater than 2°C is selected to adjust and optimize its key parameters. Specifically, based on the genetic algorithm, the metabolic rate (basal metabolic rate) that needs to be adjusted in the target segment is optimized according to the preset priority. )、Set point temperature( )、heat capacity( ) are adjusted cyclically and individually until the error between the predicted result of the thermal regulation model and the actual monitored value is less than the preset temperature threshold.

[0054] In the present application, when the key parameters are adjusted by genetic algorithm, the genetic algorithm population consists of 200 individuals, each of which has 12 genes representing active and passive parameters in the human body thermal regulation model, as shown in Table 2 below.

[0055] Table 2 Active and passive parameters in the thermal regulation model

[0056] parameter Initial Value Minimum Maximum Mmu (W / m3) 510 385 1023 Mfa (W / m3) 62 40 108 Msk (W / m3) 320 257 684 Kmu (Wm-1.k-1) 0.42 0.19 0.64 Kfa (Wm-1.k-1) 0.25 0.14 0.50 Ksk (Wm-1.k-1) 0.35 0.16 0.96 ρmu (Kg.m-3) 1048 1040 1090 ρfa (Kg.m-3) 865 850 920 ρsk (Kg.m-3) 1070 1000 1109 Cbr (J.kg-1.K-1) 3750 3583 3850 Cfa (J.kg-1.K-1) 2100 1900 3800 Csk (J.kg-1.K-1) 3550 3391 3800

[0057] In the process of adjusting and optimizing the key parameters of the target section, the objective function used is as follows:

[0058]

[0059] in, The first The prediction results of each segment; The target individual The actual monitoring value (characteristic parameter) of each section, is the number of segments of the target individual, .

[0060] Specifically, when adjusting key parameters through genetic algorithms, first, the population is initialized and individuals are defined. Each individual is a parameter set containing specific values. These parameters include basal metabolic rate , and heat capacity ( Then, a group of individuals are randomly generated as the initial population, and the parameter set of each individual (basal metabolic rate , and heat capacity ( )) are all selected randomly.

[0061] Next, based on the objective function MAE, the squares of the differences between the predicted value and the actual value of each segment are summed, and then the average is taken to evaluate the performance of each individual; and based on the fitness of the individual, the individuals with better performance are selected from the current population as the "parents" of the next generation in the genetic algorithm.

[0062] Select the basal metabolic rate from the "parent" individual , and heat capacity ( ) and exchange these subsets to produce offspring; randomly change some parameters in individuals (for example, increase or decrease the basal metabolic rate , adjust the heat capacity ( ) to introduce new genetic variation. Then, a new generation of population is generated through selection, crossover and mutation operations. Repeat the above steps until the number of iterations or fitness threshold is met, and select the individual with the highest fitness, whose corresponding basal metabolic rate , and heat capacity ( ) is the key parameter after adjustment and optimization. The corresponding key parameter after adjustment and optimization is applied in the thermal regulation model to minimize the objective function goal.

[0063] Step S104: Based on the thermal regulation model after the key parameters are adjusted, a secondary prediction is made on the skin temperature of the target individual, and the cold stress risk of the target individual is determined according to the result of the secondary prediction, and the result is fed back to the target individual.

[0064] Specifically, first obtain the environmental parameters of the target individual's location based on the environmental monitoring instrument or weather forecast, calculate the wind chill temperature, and when the wind chill temperature is less than the preset wind chill temperature, issue a wind chill temperature warning to the target individual. According to the formula:

[0065]

[0066] Calculating Wind Chill Temperature .in, is the air temperature, is the wind speed at the location of the target individual, .

[0067] The risk of target individuals is divided into 4 risk levels according to the wind chill temperature, and wind chill temperature warnings are issued to target individuals, as shown in Table 3 below.

[0068] Table 3 Wind chill temperature assessment risk classification

[0069]

[0070] In this application, local cold injury is divided into freezing cold injury and non-freezing cold injury. When the local skin temperature is lower than When the skin temperature of the target individual predicted by the thermal regulation model is lower than the preset frostbite risk index, a frostbite risk warning is issued to the target individual. Specifically, the target individual is divided into cold exposure risk levels according to the skin temperature to issue a frostbite risk warning to the target individual. The cold exposure risk level classification is shown in Table 4 below:

[0071] Table 4 Classification of cold exposure risk levels

[0072] Skin temperature Cold exposure risk Influence <-0.25 Very high Freezing frostbite -0.25~0 high Non-freezing frostbite 0~5 middle Non-freezing frostbite 5~10 Low Non-freezing frostbite 10~15 Lower Pain 15~20 none Cold discomfort >20 none No risk of cold exposure

[0073] In this application, based on the wearable device configured for the target individual, the characteristic parameters and environmental parameters of the target individual are obtained, sent to the cloud, and the cold stress risk feedback from the cloud is received. The cloud makes personalized adjustments to the key parameters of the pre-built thermal regulation model based on the characteristic parameters and environmental parameters, and determines the cold stress risk (i.e., cold risk) of the target individual based on the thermal regulation model after the key parameters are adjusted.

[0074] Therefore, the physiological parameters and environmental data of the human body are collected through the matching wearable devices, and the thermal regulation model is personalized based on the genetic algorithm, so that the differences of each individual (such as body mass index, age, gender, health status, etc.) can be more accurately considered to predict the skin temperature of various parts of the body of the target individual when working outdoors, so as to issue personalized risk warnings to the target individual, so as to better fit the actual situation of the target individual and provide a scientific and quantitative health risk assessment for outdoor work, so as to take appropriate preventive measures, such as appropriate rest time, warming measures and necessary medical preparations, to minimize the health threats brought by cold environments.

[0075] like Figure 2 As shown, the embodiment of the present application also provides a personalized cold risk prediction and evaluation system based on deep learning, the system comprising:

[0076] A preprocessing unit is configured to sequentially clean and format the received characteristic parameters of the target individual and the environmental parameters of the location where the target individual is located;

[0077] The target segment determination unit is configured to make a preliminary prediction of the skin temperature of the target individual based on the remaining constructed thermal regulation model, and determine the target segment in the thermal regulation model where key parameters need to be adjusted according to the preliminary prediction result and the formatted characteristic parameters. The key parameters include: metabolic rate, set point temperature, and heat capacity. One or more.

[0078] The parameter adjustment unit is configured to cyclically adjust key parameters of the target segment according to the formatted data based on a deep learning algorithm.

[0079] The prediction feedback unit is configured to perform a secondary prediction of the skin temperature of the target individual based on the thermal regulation model after the key parameters are adjusted, and determine the cold stress risk of the target individual according to the secondary prediction result, and feed back to the target individual.

[0080] The personalized cold risk prediction and assessment system based on deep learning provided in the embodiments of the present application can implement the steps and processes of the personalized cold risk prediction and assessment method based on deep learning described in any of the above embodiments, and achieve the same technical effects, which will not be repeated here one by one.

[0081] Figure 3 Schematic diagram of the structure of an electronic device provided according to some embodiments of the present application; Figure 3 As shown, the electronic device includes:

[0082] One or more processors 301;

[0083] The computer-readable medium can be configured to store one or more programs 302. When one or more processors 301 execute one or more programs 302, the following steps are implemented: the received characteristic parameters of the target individual and the environmental parameters of the target individual's location are cleaned and formatted in sequence; based on a pre-built thermal regulation model, a preliminary prediction is made on the skin temperature of the target individual, and based on the preliminary prediction results and the formatted characteristic parameters, a target section in the thermal regulation model where key parameters need to be adjusted is determined; wherein the key parameters include: metabolic rate, set point temperature, and heat capacity. One or more; based on a deep learning algorithm, the key parameters of the target section are cyclically adjusted according to the formatted data; based on the thermal regulation model after the key parameters are adjusted, a secondary prediction is made on the skin temperature of the target individual, and the cold stress risk of the target individual is determined based on the secondary prediction results, and the cold stress risk of the target individual is fed back to the target individual.

[0084] Figure 4 The hardware structure of the electronic device provided according to some embodiments of the present application; Figure 4 As shown, the hardware structure of the electronic device may include: a processor 401 , a communication interface 402 , a computer-readable medium 403 and a communication bus 404 .

[0085] The processor 401 , the communication interface 402 , and the computer-readable storage medium 403 communicate with each other via a communication bus 404 .

[0086] Optionally, the communication interface 402 may be an interface of a communication module, such as an interface of a GSM module.

[0087] Among them, the processor 401 can be specifically configured to: perform data cleaning and formatting on the received characteristic parameters of the target individual and the environmental parameters of the target individual's location in sequence; based on a pre-constructed thermal regulation model, make a preliminary prediction of the skin temperature of the target individual, and determine the target section in the thermal regulation model where key parameters need to be adjusted according to the preliminary prediction results and the formatted characteristic parameters; wherein the key parameters include: metabolic rate, set point temperature, and heat capacity. One or more; based on a deep learning algorithm, cyclically adjust the key parameters of the target section according to the formatted data; based on the thermal regulation model after the key parameters are adjusted, make a secondary prediction of the skin temperature of the target individual, determine the cold stress risk of the target individual according to the secondary prediction results, and feed back to the target individual.

[0088] The processor 401 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc. It may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. It may implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0089] The electronic device of the embodiment of the present application exists in various forms, including but not limited to:

[0090] (1) Mobile communication devices: These devices are characterized by having mobile communication functions and their main purpose is to provide voice and data communications. These terminals include: smart phones (e.g. iPhone), multimedia phones, functional phones, and low-end phones.

[0091] (2) Ultra-mobile personal computer devices: These devices fall into the category of personal computers, have computing and processing functions, and generally also have mobile Internet access features. These terminals include: PDA, MID and UMPC devices, such as iPad.

[0092] (3) Portable entertainment devices: These devices can display and play multimedia content. They include audio and video players (e.g. iPod), handheld game consoles, e-books, smart toys and portable car navigation devices.

[0093] (4) Server: A device that provides computing services. The server consists of a processor, hard disk, memory, system bus, etc. The server is similar to a general computer architecture, but because it needs to provide highly reliable services, it has higher requirements in terms of processing power, stability, reliability, security, scalability, and manageability.

[0094] (5) Other electronic devices with data interaction functions.

[0095] It should be pointed out that, according to the needs of implementation, the various components / steps described in the embodiments of the present application can be split into more components / steps, or two or more components / steps or partial operations of components / steps can be combined into new components / steps to achieve the purpose of the embodiments of the present application.

[0096] The above-mentioned method according to the embodiment of the present application can be implemented in hardware, firmware, or implemented as software or computer code that can be stored in a recording medium (such as a CD ROM, RAM, floppy disk, hard disk or magneto-optical disk), or implemented as a computer code originally stored in a remote recording medium or a non-temporary machine storage medium downloaded through a network and to be stored in a local recording medium, so that the method described herein can be stored in such software processing on a recording medium using a general-purpose computer, a dedicated processor or programmable or dedicated hardware (such as an ASIC or FPGA). It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component (e.g., RAM, ROM, flash memory, etc.) that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor or hardware, the personalized cold risk prediction and assessment method based on deep learning described herein is implemented. In addition, when a general-purpose computer accesses the code for implementing the method shown herein, the execution of the code converts the general-purpose computer into a dedicated computer for executing the method shown herein.

[0097] Those of ordinary skill in the art will appreciate that the units and method steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and constraints involved in the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the embodiments of the present application.

[0098] It should be noted that the various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device and system embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

[0099] The device and system embodiments described above are merely illustrative, wherein the units not shown as separate may or may not be physically separated, and the units not shown as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative work.

[0100] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A personalized cold risk prediction and assessment method based on deep learning, characterized in that: include: Step S101: cleaning and formatting the received characteristic parameters of the target individual and the environmental parameters of the location of the target individual in turn in the cloud; Step S102: Based on the pre-built thermal regulation model, a preliminary prediction is made on the skin temperature of the target individual, and according to the error evaluation between the preliminary prediction result and the formatted characteristic parameters, the average error greater than The target section of the thermal regulation model where key parameters need to be adjusted is determined by the section position; wherein the key parameters include: metabolic rate, set point temperature, heat capacity or one or more; wherein, The thermal regulation model is: ； In the formula, is the serial number of the segment part for segmenting the target individual, ; The hierarchical sequence number for each section. ; For the The first section Heat capacity of the layer; For the The first section Temperature of the layer; The time during which the target individual is exposed to the cold environment; For the The first section Heat generation of the layer; For the The first section The blood in the layer exchanges heat; Respectively represent The first section Layer, Layer and Conductive heat exchange between other layers of the sections; For the The first section The breathing heat transfer of the layer; The target individuals are The radiation heat, convection heat and evaporation heat exchange between each section and the external environment respectively; step S103, based on the deep learning algorithm, cyclically adjust the key parameters of the target section according to the formatted data; Step S104, based on the thermal regulation model after the key parameters are adjusted, the skin temperature of the target individual is secondary predicted, and the cold stress risk of the target individual is determined according to the secondary prediction result, and the cold stress risk is fed back to the target individual; wherein, if the wind chill temperature calculated according to the environmental parameters of the target individual's location is less than the preset wind chill temperature, a wind chill temperature warning is issued to the target individual; in response to the skin temperature of the target individual predicted by the thermal regulation model being less than or equal to the preset frostbite risk index, a frostbite risk warning is issued to the target individual.

2. The personalized cold risk prediction and assessment method based on deep learning according to claim 1 is characterized in that: In step S101, The data cleaning includes performing outlier detection, missing value processing and noise filtering on the characteristic parameters and the environmental parameters in sequence; The formatting includes sequentially performing standardization processing, feature engineering and data reconstruction on the feature parameters and the environmental parameters after the data cleaning.

3. The personalized cold risk prediction and assessment method based on deep learning according to claim 1 is characterized in that: In step S103, Based on the genetic algorithm, the key parameters that need to be adjusted in the target section are adjusted in a cyclic and personalized manner according to the preset priority until the error between the prediction result of the thermal regulation model and the formatted characteristic parameter is less than the preset temperature threshold.

4. The personalized cold risk prediction and assessment method based on deep learning according to claim 1 is characterized in that: In step S104, according to the formatted environmental parameters of the location of the target individual, the formula: ； Calculate the wind chill temperature ; in, is the air temperature, is the wind speed at the location of the target individual, .

5. A personalized cold risk prediction and assessment method based on deep learning, characterized in that: include: Based on the configured wearable device, characteristic parameters and environmental parameters of the target individual are obtained, sent to the cloud, and the cold stress risk fed back by the cloud is received; wherein, the cloud performs personalized adjustment of key parameters of a pre-built thermal regulation model according to the characteristic parameters and environmental parameters, and determines the cold stress risk of the target individual based on the thermal regulation model after the key parameters are adjusted.

6. A personalized cold risk prediction and assessment system based on deep learning, characterized in that: include: A preprocessing unit is configured to perform data cleaning and formatting on the cloud in turn on the received characteristic parameters of the target individual and the environmental parameters of the location where the target individual is located; The target segment determination unit is configured to make a preliminary prediction of the skin temperature of the target individual based on a pre-built thermal regulation model, and according to an error evaluation between the preliminary prediction result and the formatted characteristic parameters, set the average error greater than The target section of the thermal regulation model where key parameters need to be adjusted is determined; wherein the key parameters include: metabolic rate, set point temperature, heat capacity or one or more thereof; wherein, The thermal regulation model is: ； In the formula, is the serial number of the segment part for segmenting the target individual, ; The hierarchical sequence number for each section. ; For the The first section Heat capacity of the layer; For the The first section Temperature of the layer; The time during which the target individual is exposed to the cold environment; For the The first section Heat generation of the layer; For the The first section The blood in the layer exchanges heat; Respectively represent The first section Layer, Layer and Conductive heat exchange between other layers of the sections; For the The first section The breathing heat transfer of the layer; The target individuals are The radiation heat, convection heat and evaporation heat exchange between each section and the external environment; a parameter adjustment unit configured to cyclically adjust the key parameters of the target segment according to the formatted data based on a deep learning algorithm; The prediction feedback unit is configured to perform a secondary prediction on the skin temperature of the target individual based on the thermal regulation model after the key parameters are adjusted, determine the cold stress risk of the target individual according to the secondary prediction result, and feed back to the target individual; wherein, if the wind chill temperature calculated according to the environmental parameters of the target individual's location is less than the preset wind chill temperature, a wind chill temperature warning is issued to the target individual; in response to the skin temperature of the target individual predicted by the thermal regulation model being less than or equal to the preset frostbite risk index, a frostbite risk warning is issued to the target individual.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program is a personalized cold risk prediction and assessment method based on deep learning as described in any one of claims 1-4.

8. An electronic device, characterized in that: include: A memory, a processor, and a program stored in the memory and executable on the processor, wherein when the processor executes the program, the personalized cold risk prediction and assessment method based on deep learning as described in any one of claims 1 to 4 is implemented.

Images