Personnel stress state detection method and device based on human intelligence, and medium

Abstract

The invention provides a human factor intelligence-based personnel stress state detection method, device and equipment and a medium, and the method comprises the steps: obtaining a PPG signal of a to-be-tested person, the PPG signal of the to-be-tested person comprising first resting state data and to-be-detected state data; data slicing is conducted on the first resting state data and the to-be-detected state data, corresponding first fragment data and second fragment data are obtained, and the length of the first fragment data is larger than that of the second fragment data; the first fragment data and the second fragment data are input into a trained stress state detection model for prediction, and a first EDA (Electrocortical Activity) prediction result is obtained; and judging the stress state of the to-be-tested person based on the first EDA prediction result. According to the method provided by the invention, the problem that the detection precision of the short-time stress state is not high when the PPG data is utilized to detect the stress state of the personnel can be solved.

Description

Technical Field

[0001] The present application relates to the field of stress state detection, and in particular to a method, device, equipment and medium for detecting a person's stress state based on human factor intelligence. Background Art

[0002] In related technologies, EEG (Electroencephalogram) data or PPG (Photoplethysmographic) data is used to detect a person's stress state. However, collecting EEG data requires wearing an EEG acquisition device, which is a cumbersome process and the collected data is easily disturbed by movement. When using PPG data to predict a person's stress state, it is usually necessary to use longer PPG data to extract HRV (Heart Rate Variability) features. The PPG data corresponding to a short-term stress state may be submerged in the long-term PPG data, affecting the detection accuracy. Summary of the invention

[0003] The present application provides a method, device, equipment and medium for detecting the stress state of personnel based on human factors intelligence, which helps to solve the problem of low detection accuracy of short-term stress state when using PPG data to detect the stress state of personnel.

[0004] In a first aspect, the present application provides a method for detecting a person's stress state based on human factor intelligence, comprising:

[0005] Acquire a PPG signal of the person to be tested, where the PPG signal of the person to be tested includes first resting state data and data in a state to be detected;

[0006] Slicing the first resting state data and the to-be-detected state data respectively to obtain corresponding first fragment data and second fragment data, wherein the length of the first fragment data is greater than the length of the second fragment data;

[0007] Inputting the first segment data and the second segment data into a trained personnel stress state detection model for prediction to obtain a first EDA prediction result;

[0008] The stress state of the person to be tested is determined based on the first EDA prediction result.

[0009] In one possible implementation, the first segment data and the second segment data are input into a trained personnel stress state detection model for prediction to obtain a first EDA prediction result, including:

[0010] Encoding is performed based on the first segment data to obtain a first encoding result;

[0011] Encoding based on the second segment data to obtain a second encoding result;

[0012] Decoding is performed based on the first encoding result and the second encoding result to obtain a first EDA prediction result.

[0013] In one possible implementation, encoding is performed based on the first segment data to obtain a first encoding result, including:

[0014] Using a first convolutional neural network to extract features from the first segment data to obtain first feature information;

[0015] A self-attention calculation is performed based on the first feature information to obtain a first encoding result.

[0016] In one possible implementation, encoding is performed based on the second segment data to obtain a second encoding result, including:

[0017] The second segment data is encoded using a second convolutional neural network to obtain a second encoding result.

[0018] In one possible implementation, decoding is performed based on the first encoding result and the second encoding result to obtain a first EDA prediction result, including:

[0019] Calculate self-attention based on the second encoding result, and then calculate cross-attention based on the calculation result of self-attention and the first encoding result to obtain state change information;

[0020] The third convolutional neural network is used to convert the state change information into the first EDA prediction result.

[0021] In one possible implementation, the training process of the personnel stress state detection model includes:

[0022] Acquire a PPG signal and an EDA signal of the trainee, wherein the PPG signal of the trainee includes the second resting state data and the first experimental data, and the EDA signal of the trainee includes the third resting state data and the second experimental data, and the PPG signal and the EDA signal of the trainee are synchronously acquired by the physiological signal acquisition device when the trainee performs the same stress task;

[0023] Performing data normalization based on the third resting state data and the second experimental data to obtain normalized EDA data;

[0024] Slicing the second resting state data, the first experimental data, and the normalized EDA data respectively to obtain corresponding third segment data, fourth segment data, and fifth segment data, wherein the length of the third segment data is equal to the length of the first segment data, and the lengths of the fourth segment data and the fifth segment data are both equal to the length of the second segment data;

[0025] Encoding based on the third segment data to obtain a third encoding result;

[0026] Encoding based on the fourth segment data to obtain a fourth encoding result;

[0027] Decoding is performed based on the third encoding result and the fourth encoding result to obtain a second EDA prediction result;

[0028] A loss value is calculated based on the second EDA prediction result and the fifth segment data, so as to adjust the model parameters using the loss value.

[0029] In one possible implementation, the stress state of the person to be tested includes a low stress state, a medium stress state, and a high stress state, and judging the stress state of the person to be tested based on the first EDA prediction result includes:

[0030] If the value of the first EDA prediction result is within the first preset range, the stress state of the person to be tested is a low stress state; or

[0031] If the value of the first EDA prediction result is within the second preset range, the stress state of the person to be tested is a medium stress state; or

[0032] If the value of the first EDA prediction result is within the third preset range, the stress state of the person to be tested is a high stress state.

[0033] In one possible implementation, after obtaining the PPG signal of the person to be tested, the method further includes:

[0034] Filter the PPG signal.

[0035] In a second aspect, the present application provides a device for detecting a stress state of a person based on human factor intelligence, comprising:

[0036] An acquisition module, used for acquiring a PPG signal of a person to be tested, wherein the PPG signal of the person to be tested includes first resting state data and data of a state to be detected;

[0037] A data slicing module, used to slice the first resting state data and the to-be-detected state data respectively to obtain corresponding first fragment data and second fragment data, wherein the length of the first fragment data is greater than the length of the second fragment data;

[0038] A prediction module, used for inputting the first segment data and the second segment data into a trained personnel stress state detection model for prediction, and obtaining a first EDA prediction result;

[0039] The state judgment module is used to judge the stress state of the person to be tested based on the first EDA prediction result.

[0040] In a third aspect, the present application provides an electronic device comprising: a processor and a memory, the memory being used to store a computer program; the processor being used to run the computer program to implement the method for detecting a person's stress state based on human factor intelligence as described in the first aspect.

[0041] In a fourth aspect, the present application provides a computer-readable storage medium, which stores a computer program. When the computer-readable storage medium is executed on a computer, the computer implements the method for detecting a person's stress state based on human factor intelligence as described in the first aspect.

[0042] Compared with the prior art, this application has at least the following technical effects:

[0043] The present application provides a method for detecting the stress state of personnel based on human factor intelligence, which uses the PPG signal that is easy to collect to detect the stress state of personnel. By converting the PPG signal into an EDA signal, the PPG signal required for detecting the stress state of personnel is shorter, which is conducive to improving the detection accuracy of short-term stress state. At the same time, after converting the PPG signal into an EDA signal, the stress state of the personnel can be judged according to the change of the EDA signal. The judgment process is independent of the specific task, and the decoupling of the stress state detection of the personnel and the specific task can be achieved. Furthermore, the baseline data of the personnel (i.e., resting state data) is considered in the detection of the stress state of the personnel, and the baseline data can be used to suppress individual differences. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A schematic diagram of a process for detecting a person's stress state based on human intelligence provided in an embodiment of the present application;

[0045] Figure 2 It is a flowchart of using the trained personnel stress state detection model for prediction;

[0046] Figure 3 It is a structural schematic diagram of a personnel stress state detection model provided in an embodiment of the present application;

[0047] Figure 4 It is a flowchart of the training process of the personnel stress state detection model provided in the embodiment of the present application;

[0048] Figure 5 A schematic diagram of the structure of a device for detecting a stress state of a person based on human factors intelligence provided in an embodiment of the present application;

[0049] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0050] In the embodiments of the present application, unless otherwise specified, the character " / " indicates that the objects before and after the association are in an or relationship. For example, A / B can represent A or B. "And / or" describes the association relationship of the associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone.

[0051] It should be pointed out that the words "first", "second", etc. involved in the embodiments of the present application are only used to distinguish the description purpose, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated, nor can they be understood as indicating or implying order.

[0052] In the embodiments of the present application, "at least one" refers to one or more, and "plurality" refers to two or more. In addition, "at least one of the following" or similar expressions refers to any combination of these items, which may include any combination of single items or plural items. For example, at least one of A, B, or C may represent: A, B, C, A and B, A and C, B and C, or A, B and C. Among them, each of A, B, and C may be an element itself, or a set containing one or more elements.

[0053] In the embodiments of the present application, "exemplary", "in some embodiments", "in another embodiment", etc. are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" in the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present concepts in a concrete way.

[0054] In the embodiments of the present application, "of", "corresponding", and "corresponding" can sometimes be used interchangeably. It should be noted that when the distinction between them is not emphasized, the meanings to be expressed are consistent. In the embodiments of the present application, communication and transmission can sometimes be used interchangeably. It should be noted that when the distinction between them is not emphasized, the meanings to be expressed are consistent. For example, transmission can include sending and / or receiving, which can be a noun or a verb.

[0055] The equal to involved in the embodiments of the present application can be used in conjunction with greater than, and is applicable to the technical solution adopted when greater than, and can also be used in conjunction with less than, and is applicable to the technical solution adopted when less than. It should be noted that when equal to is used in conjunction with greater than, it cannot be used in conjunction with less than; when equal to is used in conjunction with less than, it cannot be used in conjunction with greater than.

[0056] In the related art, the detection of personnel stress state has the following problems:

[0057] 1. Currently, the commonly used models for detecting stress status of personnel are often based on EEG or PPG modality data.

[0058] In actual application, not all tasks are suitable for wearing EEG acquisition devices during execution. In the case of wearable EEG acquisition devices, wearing EEG devices will also have a certain impact on the subject's state, and the collected data will also be affected by the task, such as motion interference.

[0059] Prediction of a person's stress state based on PPG signals usually requires extraction of HRV features. HRV feature extraction requires longer PPG data, such as 3 minutes or even longer. The PPG signal corresponding to a short-term stress state may be submerged in the long-term PPG signal, affecting the prediction accuracy.

[0060] 2. The existing models for detecting personnel stress states usually detect the stress states induced by specific tasks, and the cross-task detection capabilities of the models are insufficient.

[0061] 3. When using physiological signals to judge the stress state of the subjects, the individuality of the subjects will have a great influence on the judgment results. The existing personnel stress state detection algorithms do not adequately consider the issue of individual differences.

[0062] Based on the above problems, an embodiment of the present application proposes a method for detecting personnel stress status based on human factor intelligence.

[0063] Now combined Figure 1-Figure 4 The method for detecting a person's stress state based on human factor intelligence provided in an embodiment of the present application is described.

[0064] Figure 1 The flowchart of the method for detecting the stress state of a person based on human factor intelligence provided in the embodiment of the present application specifically includes the following steps:

[0065] Step S11, obtaining a PPG signal of the person to be tested, wherein the PPG signal of the person to be tested includes first resting state data and data in a state to be detected.

[0066] Specifically, a portable biometric information collection device or a telemetry device is used to collect the PPG signal of the person to be tested. The PPG signal of the person to be tested includes first resting state data (also called baseline data) and data in a state to be detected. The collection process needs to record the start and end time of the first resting state data so as to divide the PPG signal of the person to be tested into the first resting state data and data in a state to be detected according to the start and end time of the first resting state data.

[0067] Optionally, after obtaining the PPG signal of the person to be tested, the PPG signal can also be filtered. The effective frequency band range of the PPG signal is [0.5-8] Hz, and other frequency band signals can be considered as noise signals, which need to be filtered out. Therefore, a 0.5-8 Hz bandpass filter can be used for processing to obtain a filtered PPG signal. Filtering the PPG signal can remove noise and improve detection accuracy.

[0068] Step S12, slicing the first resting state data and the to-be-detected state data respectively to obtain corresponding first fragment data and second fragment data, wherein the length of the first fragment data is greater than the length of the second fragment data.

[0069] Specifically, the first resting state data and the state data to be detected are segmented, the first resting state data is segmented into first segment data, and the state data to be detected is segmented into second segment data, wherein the length of the first segment data is greater than the length of the second segment data. For example, 60s of the first segment data is intercepted from the first resting state data, and the state data to be detected is divided into 1s of the second segment data. Acquiring a longer segment of the first static data is helpful for the subsequent extraction of information about the person to be tested in a static state, and acquiring a shorter time segment of the data to be detected is helpful for the model to use a shorter PPG signal to detect the stress state of the person, and the real-time prediction of the model is better. Optionally, the length of the first segment data can be 60s, or 2min, 3min, etc., and the length of the second segment data can be 1s, 2s, 3s, etc. The number of the first segment data and the second segment data can be one or more, which is not limited in this application.

[0070] For example, the first resting state data can be divided into 60s segment PPG baseseg , divide the state data to be detected into 60 1s segments PPG token , then [PPG baseseg ,PPG token1 ,PPG token2,… PPG token60 ] as the input of the model for prediction.

[0071] Step S13: input the first segment data and the second segment data into a trained personnel stress state detection model for prediction to obtain a first EDA prediction result.

[0072] Specifically, Figure 2 As shown, Figure 2 This is a flowchart of using the trained personnel stress state detection model for prediction. Figure 3is a schematic diagram of the structure of a personnel stress state detection model provided in an embodiment of the present application. The personnel stress state detection model of the present application specifically refers to a Transformer model. The Transformer model includes a first convolutional neural network ( Figure 3 The Baseline Conv Block shown in the figure), the second convolutional neural network ( Figure 3 Exp Conv Block shown in ), the third convolutional neural network ( Figure 3 Output Conv Block shown in ), encoder ( Figure 3 The encoder and decoder shown in Figure 3 The Decoder shown in ). It includes the following steps:

[0073] Step S131: Encode based on the first fragment data to obtain a first encoding result.

[0074] Specifically, the first convolutional neural network is used to process the first segment data ( Figure 3 PPG shown in baseseg ) to extract features and obtain the first feature information. Then the first feature information is input into the encoder, and the multi-head attention module performs self-attention calculation based on the first feature information to obtain the first encoding result. The calculation will focus on the entire PPG baseseg The individual information of the person to be tested and the information of the person to be tested in a resting state are extracted.

[0075] Step S132: Encode based on the second fragment data to obtain a second encoding result.

[0076] Specifically, the second segment data ( Figure 3 PPG shown in token ) is encoded to obtain a second encoding result.

[0077] Step S133: Decoding is performed based on the first encoding result and the second encoding result to obtain the first EDA prediction result.

[0078] Specifically, the second encoding result is input into the masked multi-head attention module of the decoder, and the self-attention is calculated based on the second encoding result; the calculation result of the self-attention and the first encoding result are input into the cross multi-head attention module, and the cross-attention is calculated based on the calculation result of the self-attention and the first encoding result to obtain the state change information; finally, the state change information is converted into the first EDA prediction result ( Figure 3 The EDA shown in pred ).

[0079] The calculation process also takes into account the baseline PPG signal (i.e., PPG baseseg ) and the PPG signal of the state to be tested (ie PPG token ), which can extract the information of the state change of the test subject and suppress individual differences to a certain extent. Finally, the calculation results are input into the convolutional neural network to convert the state change information that suppresses individual differences into EDA signals.

[0080] Figure 3 The structure of the personnel stress state detection model shown in also includes: Feed Forward (feedforward neural network layer), Add&Norm (residual connection and layer normalization operation module), and softmax layer, wherein Feed Forward is used to perform nonlinear transformation on the calculation results output by the multi-head attention module. Feed Forward combined with the multi-head attention module can obtain more comprehensive information; each layer in the model (such as the multi-head attention layer and the feedforward neural network layer) is followed by a residual connection and layer normalization operation module, which is used to add the output of each layer to its input to form a residual connection, and then perform layer normalization; the softmax layer is used to convert the output of the decoder into the final prediction result.

[0081] Step S14: determining the stress state of the person to be tested based on the first EDA prediction result.

[0082] In one possible embodiment, the stress state of the person to be tested includes a low stress state, a medium stress state and a high stress state, and the stress state of the person to be tested is judged based on the first EDA prediction result, including: if the value of the first EDA prediction result is within a first preset range, the stress state of the person to be tested is a low stress state; or if the value of the first EDA prediction result is within a second preset range, the stress state of the person to be tested is a medium stress state; or if the value of the first EDA prediction result is within a third preset range, the stress state of the person to be tested is a high stress state.

[0083] For example, if the value of the first EDA prediction result is between 0.6 and 1, the person to be tested is considered to be in a high stress state; if the value of the first EDA prediction result is between 0.3 and 0.6, the person to be tested is considered to be in a medium stress state; if the value of the first EDA prediction result is between 0 and 0.3, the person to be tested is considered to be in a low stress state.

[0084] EDA based on model prediction results pred Determine the stress state of personnel and the prediction results of the model EDA pred By setting different thresholds, the classification of different stress levels can be obtained for the relative changes in the skin electricity of personnel. Furthermore, the threshold can be adjusted according to the difficulty of the experimental task. For example, when the task difficulty is low, a lower threshold can be set to increase sensitivity; and when the task difficulty is high, the threshold can be increased to increase specificity.

[0085] In the present application, the stress state of a person is detected by using a PPG signal that is easy to collect. By converting the PPG signal into an EDA signal, the PPG signal required for detecting the stress state of a person is shorter, which is beneficial to improving the detection accuracy of the short-term stress state. At the same time, after the PPG signal is converted into an EDA signal, the stress state of the person can be judged according to the change of the EDA signal. This judgment process is independent of the specific task. The model can detect the stress state of a person across different tasks or situations, thereby decoupling the stress state detection of a person from the specific task. This is because the EDA signal directly reflects the physiological stress response of the human body, rather than relying on the induction of a specific task. . Furthermore, the baseline data of the person (i.e., resting state data) is considered in the detection of the stress state of the person, and the result of the analysis is the relative change of the PPG, so the baseline data can be used to suppress individual differences.

[0086] In this application, the training process of the personnel stress state detection model is as follows: Figure 4 As shown, Figure 4 : is a flow chart of the training process of the personnel stress state detection model provided in the embodiment of the present application, which specifically includes the following steps:

[0087] Step S21, obtaining a PPG signal and an EDA signal of the trainee, wherein the PPG signal of the trainee includes second resting state data and first experimental data, and the EDA signal of the trainee includes third resting state data and second experimental data.

[0088] Specifically, a portable bio-information acquisition device (also referred to as a physiological signal acquisition device) or a telemetry device is used to acquire the PPG signal of the trainee, and a portable bio-information acquisition device is used to acquire the EDA signal of the trainee. Among them, the PPG signal and EDA signal of the trainee are synchronously acquired by the physiological signal acquisition device when the trainee performs the same stress task. That is, when the trainee performs the same stress task, the resting state start and end time of the EDA signal is the same as the resting state start and end time of the PPG signal, and the start and end time of the experimental data acquisition of the EDA signal is the same as the start and end time of the experimental data acquisition of the PPG signal.

[0089] The acquisition process needs to record the start and end time of the second resting state data and the third resting state data, so as to divide the PPG signal of the trainee into the second resting state data and the first experimental data according to the start and end time of the second resting state data, and divide the EDA signal of the trainee into the third resting state data and the second experimental data according to the start and end time of the third resting state data.

[0090] Optionally, the collected PPG signal and EDA signal are filtered. Specifically, the PPG signal can be processed by a 0.5-8 Hz bandpass filter to obtain a filtered PPG signal, and the effective frequency of the EDA signal is relatively low, and the EDA signal is processed by a 3 Hz low-pass filter to obtain a filtered EDA signal.

[0091] Step S22, performing data normalization based on the third resting state data and the second experimental data to obtain normalized EDA data.

[0092] Since the EDA data of trainees are different, in order to effectively suppress individual differences, it is necessary to normalize the EDA data. The specific calculation formula is as follows:

[0093]

[0094] Among them, EDA norm Refers to the normalized EDA data. exp Refers to the second experimental data, EDA base It refers to the third resting state data. The second experimental data is divided by the mean of the baseline data (i.e., the third resting state data), and then divided by 10 in order to normalize the data to the range of 0-1.

[0095] By normalizing the EDA data, the differences in baseline levels of different individuals can be eliminated and the data can be adjusted to a common scale to facilitate comparison of data from different individuals or analysis of data changes of the same individual under different conditions, making the data more comparable. Scaling the data to the range of 0-1 helps with subsequent model training, and normalized data can improve the convergence speed and performance of the model.

[0096] Step S23, slicing the second resting state data, the first experimental data and the normalized EDA data respectively to obtain corresponding third segment data, fourth segment data and fifth segment data.

[0097] Specifically, the second resting state data, the first experimental data and the normalized EDA data are segmented, the second resting state data is segmented into the third segment data, the first experimental data is segmented into the fourth segment data, and the normalized EDA data is segmented into the fifth segment data. The length of the third segment data is equal to the length of the first segment data, and the lengths of the fourth segment data and the fifth segment data are equal to the length of the second segment data, which meets the model's requirements for the size of the data input data. For example, the length of the third segment data is equal to the length of the first segment data, both of which are 60s, and the lengths of the fourth segment data and the fifth segment data are equal to the length of the second segment data, both of which are 1s. The number of the third segment data, the fourth segment data and the fifth segment data can be one or more, and this application does not limit it.

[0098] In an optional implementation, the second resting state data is divided into a third segment data PPG of 60 seconds. sbaseseg , the first experimental data is divided into 60 1s fourth segment data PPG stoken , the normalized EDA data is divided into 60 1s fifth segment data EDA token , then [PPG sbaseseg ,PPG stoken1 ,PPG stoken2,… PPG stoken60, EDA token1, EDA token2… EDA token60 ] is used as input data for model training.

[0099] Step S24: Encode based on the third fragment data to obtain a third encoding result.

[0100] The specific implementation principle refers to the above step S131 and will not be elaborated here.

[0101] Step S25: Encode based on the fourth fragment data to obtain a fourth encoding result.

[0102] The specific implementation principle refers to the above step S132 and will not be elaborated here.

[0103] Step S26, decoding is performed based on the third encoding result and the fourth encoding result to obtain a second EDA prediction result.

[0104] The specific implementation principle refers to the above step S133 and will not be elaborated here.

[0105] Step S27, calculating a loss value based on the second EDA prediction result and the fifth segment data, so as to adjust model parameters using the loss value.

[0106] Specifically, step S26 calculates and obtains a second EDA prediction result, and compares the second EDA prediction result with the fifth segment data EDA token Compare, calculate loss and backpropagate to adjust model parameters using loss.

[0107] In an optional embodiment, for a set of data [PPG sbaseseg ,PPG stoken1 ,PPG stoken2,… PPG stoken60, EDA token1, EDA token2… EDA token60 ], first execute step S24, for PPG sbaseseg Encode to obtain a third encoding result; then execute step S25 to encode PPG stoken1 Encode to obtain a fourth encoding result; then execute step S26 to decode the third encoding result and the fourth encoding result to obtain a second EDA prediction result, and finally compare the second EDA prediction result with the EDA token1 Compare and calculate the loss value to complete the first training. After completing the first training, update the input of step S25 to [PPG stoken1, PPG stoke2 ], repeat steps S24-S26, and compare the prediction result obtained in step S26 with EDA token2 Compare and calculate the loss value to complete the second training. Repeat the above process until 60 PPGs are obtained. stoken All completed training.

[0108] This application encodes PPG data based on the Transformer model and then decodes it into EDA. Use EDA data to detect the stress state of personnel. Effectively utilize baseline data in the decoding and encoding process of the model to suppress individual differences in personnel physiological signals to a certain extent. In addition, the PPG signal is converted into an EDA signal that is more strongly correlated with the stress state, and the stress state of the person is judged according to the changes in the EDA signal. After the PPG signal is converted into an EDA signal, the judgment process is independent of the specific task, and the decoupling of personnel stress state detection and specific tasks can be achieved; at the same time, the PPG signal required for personnel state detection is shorter, which is conducive to the identification of short-term stress states.

[0109] Based on the same idea, Figure 5 A schematic diagram of the structure of a device for detecting a stress state of a person based on human intelligence provided in an embodiment of the present application is shown in FIG. Figure 5 As shown, the personnel stress state detection device 50 based on human factor intelligence may include:

[0110] An acquisition module 51 is used to acquire a PPG signal of a person to be tested, wherein the PPG signal of the person to be tested includes first resting state data and data in a state to be detected;

[0111] A data slicing module 52, configured to slice the first resting state data and the to-be-detected state data respectively to obtain corresponding first segment data and second segment data, wherein the length of the first segment data is greater than the length of the second segment data;

[0112] A prediction module 53, used for inputting the first segment data and the second segment data into a trained personnel stress state detection model for prediction, to obtain a first EDA prediction result;

[0113] The state judgment module 54 is used to judge the stress state of the person to be tested based on the first EDA prediction result.

[0114] In one possible implementation, the prediction module 53 may also be used to:

[0115] Encoding is performed based on the first segment data to obtain a first encoding result;

[0116] Encoding based on the second segment data to obtain a second encoding result;

[0117] Decoding is performed based on the first encoding result and the second encoding result to obtain a first EDA prediction result.

[0118] In one possible implementation, encoding is performed based on the first segment data to obtain a first encoding result, including:

[0119] Using a first convolutional neural network to extract features from the first segment data to obtain first feature information;

[0120] A self-attention calculation is performed based on the first feature information to obtain a first encoding result.

[0121] In one possible implementation, encoding is performed based on the second segment data to obtain a second encoding result, including:

[0122] The second segment data is encoded using a second convolutional neural network to obtain a second encoding result.

[0123] In one possible implementation, decoding is performed based on the first encoding result and the second encoding result to obtain a first EDA prediction result, including:

[0124] Calculate self-attention based on the second encoding result, and then calculate cross-attention based on the calculation result of self-attention and the first encoding result to obtain state change information;

[0125] The third convolutional neural network is used to convert the state change information into the first EDA prediction result.

[0126] In one possible implementation, the training process of the personnel stress state detection model includes:

[0127] Acquire a PPG signal and an EDA signal of the trainee, wherein the PPG signal of the trainee includes the second resting state data and the first experimental data, and the EDA signal of the trainee includes the third resting state data and the second experimental data, and the PPG signal and the EDA signal of the trainee are synchronously acquired by the physiological signal acquisition device when the trainee performs the same stress task;

[0128] Performing data normalization based on the third resting state data and the second experimental data to obtain normalized EDA data;

[0129] Slicing the second resting state data, the first experimental data, and the normalized EDA data respectively to obtain corresponding third segment data, fourth segment data, and fifth segment data, wherein the length of the third segment data is equal to the length of the first segment data, and the lengths of the fourth segment data and the fifth segment data are both equal to the length of the second segment data;

[0130] Encoding based on the third segment data to obtain a third encoding result;

[0131] Encoding based on the fourth segment data to obtain a fourth encoding result;

[0132] Decoding is performed based on the third encoding result and the fourth encoding result to obtain a second EDA prediction result;

[0133] A loss value is calculated based on the second EDA prediction result and the fifth segment data, so as to adjust the model parameters using the loss value.

[0134] In one possible implementation, the stress state of the person to be tested includes a low stress state, a medium stress state and a high stress state, and the state determination module 54 can also be used for:

[0135] If the value of the first EDA prediction result is within the first preset range, the stress state of the person to be tested is a low stress state; or

[0136] If the value of the first EDA prediction result is within the second preset range, the stress state of the person to be tested is a medium stress state; or

[0137] If the value of the first EDA prediction result is within the third preset range, the stress state of the person to be tested is a high stress state.

[0138] In one possible implementation, the device further includes a filtering module, which is used to filter the PPG signal after acquiring the PPG signal of the person to be tested.

[0139] Figure 5 The human factor intelligence-based personnel stress state detection device 50 provided in the illustrated embodiment can be used to execute the technical solution of the method embodiment shown in the present application. Its implementation principle and technical effects can be further referred to the relevant description in the method embodiment.

[0140] It should be understood that the above Figure 5 The division of the modules of the human factor intelligence-based personnel stress state detection device 50 shown is only a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. And these modules can all be implemented in the form of software calling through processing elements; they can also be all implemented in the form of hardware; some modules can also be implemented in the form of software calling through processing elements, and some modules can be implemented in the form of hardware. For example, the acquisition module can be a separately established processing element, or it can be integrated in a chip of an electronic device. The implementation of other modules is similar. In addition, all or part of these modules can be integrated together, or they can be implemented independently. In the implementation process, each step of the above method or each of the above modules can be completed by an integrated logic circuit of hardware in the processor element or instructions in the form of software.

[0141] For example, the above modules may be one or more integrated circuits configured to implement the above methods, such as one or more application specific integrated circuits (ASIC), or one or more microprocessors (DSP), or one or more field programmable gate arrays (FPGA). For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0142] In the above embodiments, the processor involved may include, for example, a CPU, a DSP, a microcontroller or a digital signal processor, and may also include a GPU, an embedded neural network processor (Neural-network Process Units; hereinafter referred to as: NPU) and an image signal processor (Image Signal Processing; hereinafter referred to as: ISP). The processor may also include necessary hardware accelerators or logic processing hardware circuits, such as ASIC, or one or more integrated circuits for controlling the execution of the program of the technical solution of the present application. In addition, the processor may have the function of operating one or more software programs, and the software programs may be stored in a storage medium.

[0143] An embodiment of the present application also provides a computer-readable storage medium, in which a computer program is stored. When the computer-readable storage medium is run on a computer, the computer executes the method provided by the embodiment shown in the present application.

[0144] Combine the following Figure 6 The exemplary electronic device provided in the embodiments of the present application is further introduced. Figure 6 A schematic structural diagram of an electronic device 6000 is shown.

[0145] The above-mentioned electronic device 6000 may include: at least one processor; and at least one memory communicatively connected to the above-mentioned processor, wherein: the above-mentioned memory stores program instructions that can be executed by the above-mentioned processor, and the processor calls the above-mentioned program instructions to execute the personnel stress state detection method based on human factors intelligence provided in the embodiment shown in this application.

[0146] Figure 6 A block diagram of an exemplary electronic device 6000 suitable for implementing the embodiments of the present application is shown. Figure 6 The electronic device 6000 shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0147] like Figure 6 As shown, the electronic device 6000 is in the form of a general computing device. The components of the electronic device 6000 may include, but are not limited to: one or more processors 6010, a memory 6020, a communication bus 6040 connecting different system components (including the memory 6020 and the processor 6010), and a communication interface 6030.

[0148] The communication bus 6040 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor or a local bus using any of a variety of bus structures. For example, these architectures include but are not limited to Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MAC) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus and Peripheral Component Interconnection (PCI) bus.

[0149] The electronic device 6000 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the electronic device, including volatile and non-volatile media, removable and non-removable media.

[0150] The memory 6020 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) and / or cache memory. The electronic device may further include other removable / non-removable, volatile / non-volatile computer system storage media. Figure 6Not shown, a disk drive for reading and writing a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing a removable non-volatile optical disk (e.g., a compact disc read only memory (CD-ROM), a digital versatile disc read only memory (DVD-ROM), or other optical media) may be provided. In these cases, each drive may be connected to the communication bus 6040 via one or more data medium interfaces. The memory 6020 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present application.

[0151] A program / utility having a set (at least one) of program modules may be stored in the memory 6020, such program modules including but not limited to an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. The program modules generally perform the functions and / or methods of the embodiments described herein.

[0152] The electronic device 6000 may also communicate with one or more external devices (e.g., keyboard, pointing device, display, etc.), one or more devices that enable a user to interact with the electronic device, and / or any device that enables the electronic device to communicate with one or more other computing devices (e.g., network card, modem, etc.). Such communication may be performed through the communication interface 6030. In addition, the electronic device 6000 may also communicate with the network adapter ( Figure 6 The network adapter can communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through the communication bus 6040. It should be understood that although Figure 6 Not shown, other hardware and / or software modules may be used in conjunction with the electronic device 6000, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, disk arrays (Redundant Arrays of Independent Drives; hereinafter referred to as: RAID) systems, tape drives, and data backup storage systems.

[0153] The processor 6010 executes various functional applications and data processing by running the programs stored in the memory 6020, such as implementing the method provided in the embodiment of the present application.

[0154] It is understandable that the interface connection relationship between the modules illustrated in the embodiment of the present application is only a schematic illustration and does not constitute a structural limitation on the electronic device 6000. In other embodiments of the present application, the electronic device 6000 may also adopt different interface connection methods in the above embodiments, or a combination of multiple interface connection methods.

[0155] In the above embodiments, the processor involved may include, for example, a CPU, a DSP, a microcontroller or a digital signal processor, and may also include a GPU, an embedded neural network processor (Neural-network Process Units; hereinafter referred to as: NPU) and an image signal processor (Image Signal Processing; hereinafter referred to as: ISP). The processor may also include necessary hardware accelerators or logic processing hardware circuits, such as ASIC, or one or more integrated circuits for controlling the execution of the program of the technical solution of the present application. In addition, the processor may have the function of operating one or more software programs, and the software programs may be stored in a storage medium.

[0156] Those of ordinary skill in the art will appreciate that the various units and algorithm steps described in the embodiments disclosed herein can be implemented in a combination of electronic hardware, computer software, and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of 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 this application.

[0157] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0158] In several embodiments provided in the present application, if any function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory; hereinafter referred to as: ROM), random access memory (Random Access Memory; hereinafter referred to as: RAM), disk or optical disk, and other media that can store program codes.

[0159] The above is only a specific implementation of the present application. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. The protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for detecting personnel stress state based on human factor intelligence, characterized in that: The method comprises: Acquire a PPG signal of a person to be tested, wherein the PPG signal of the person to be tested includes first resting state data and state data to be detected; Slicing the first resting state data and the to-be-detected state data respectively to obtain corresponding first fragment data and second fragment data, wherein the length of the first fragment data is greater than the length of the second fragment data; Inputting the first segment data and the second segment data into a trained personnel stress state detection model for prediction to obtain a first EDA prediction result; The stress state of the person to be tested is determined based on the first EDA prediction result.

2. The method according to claim 1, characterized in that The step of inputting the first segment data and the second segment data into a trained personnel stress state detection model for prediction to obtain a first EDA prediction result includes: Encoding is performed based on the first fragment data to obtain a first encoding result; Perform encoding based on the second fragment data to obtain a second encoding result; Decoding is performed based on the first encoding result and the second encoding result to obtain the first EDA prediction result.

3. The method according to claim 2, characterized in that The encoding based on the first fragment data to obtain a first encoding result includes: Using a first convolutional neural network to extract features from the first segment data to obtain first feature information; A self-attention calculation is performed based on the first feature information to obtain a first encoding result.

4. The method according to claim 3, characterized in that The encoding based on the second fragment data to obtain a second encoding result includes: The second fragment data is encoded using a second convolutional neural network to obtain a second encoding result.

5. The method according to claim 4, characterized in that The decoding based on the first encoding result and the second encoding result to obtain the first EDA prediction result includes: Calculating self-attention based on the second encoding result, and then calculating cross-attention based on the calculation result of the self-attention and the first encoding result to obtain state change information; The state change information is converted into the first EDA prediction result by using a third convolutional neural network.

6. The method according to claim 1, characterized in that The training process of the personnel stress state detection model includes: Acquire a PPG signal and an EDA signal of a trainee, wherein the PPG signal of the trainee includes the second resting state data and the first experimental data, and the EDA signal of the trainee includes the third resting state data and the second experimental data, and the PPG signal and the EDA signal of the trainee are synchronously acquired by a physiological signal acquisition device when the trainee performs the same stress task; Performing data normalization based on the third resting state data and the second experimental data to obtain normalized EDA data; Slicing the second resting state data, the first experimental data, and the normalized EDA data respectively to obtain corresponding third segment data, fourth segment data, and fifth segment data, wherein the length of the third segment data is equal to the length of the first segment data, and the lengths of the fourth segment data and the fifth segment data are both equal to the length of the second segment data; Perform encoding based on the third segment data to obtain a third encoding result; Perform encoding based on the fourth segment data to obtain a fourth encoding result; Decoding is performed based on the third encoding result and the fourth encoding result to obtain a second EDA prediction result; A loss value is calculated based on the second EDA prediction result and the fifth segment data, so as to adjust the model parameters using the loss value.

7. The method according to claim 1, characterized in that The stress state of the person to be tested includes a low stress state, a medium stress state and a high stress state, and judging the stress state of the person to be tested based on the first EDA prediction result includes: If the value of the first EDA prediction result is within a first preset range, the stress state of the person to be tested is a low stress state; or If the value of the first EDA prediction result is within the second preset range, the stress state of the person to be tested is a medium stress state; or If the value of the first EDA prediction result is within a third preset range, the stress state of the person to be tested is a high stress state.

8. The method according to any one of claims 1 to 7, characterized in that: After obtaining the PPG signal of the person to be tested, the method further includes: The PPG signal is filtered.

9. A device for detecting stress state of personnel based on human factor intelligence, characterized in that: include: An acquisition module, used for acquiring a PPG signal of a person to be tested, wherein the PPG signal of the person to be tested includes first resting state data and state data to be detected; a data slicing module, configured to slice the first resting state data and the to-be-detected state data respectively to obtain corresponding first fragment data and second fragment data, wherein the length of the first fragment data is greater than the length of the second fragment data; A prediction module, used for inputting the first segment data and the second segment data into a trained personnel stress state detection model for prediction, to obtain a first EDA prediction result; A state judgment module is used to judge the stress state of the person to be tested based on the first EDA prediction result.

10. An electronic device, characterized in that: include: A processor and a memory, wherein the memory is used to store a computer program; the processor is used to run the computer program to implement the method for detecting a personnel stress state based on human factor intelligence as described in any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed on a computer, the method for detecting a stress state of a person based on human factor intelligence as described in any one of claims 1 to 8 is implemented.

Images