CN114504334A

CN114504334A - State prediction method, state prediction device, computer equipment and storage medium

Info

Publication number: CN114504334A
Application number: CN202210116350.9A
Authority: CN
Inventors: 不公告发明人
Original assignee: Suzhou Xiaowei Changxing Robot Co ltd
Current assignee: Suzhou Xiaowei Changxing Robot Co ltd
Priority date: 2022-02-07
Filing date: 2022-02-07
Publication date: 2022-05-17
Anticipated expiration: 2042-02-07
Also published as: CN114504334B

Abstract

The application relates to a state prediction method, a state prediction device, a computer device and a storage medium. The method comprises the following steps: acquiring corresponding electromyographic signals aiming at target muscles of a target object; determining an electromyographic characteristic index corresponding to the target muscle at the current moment according to the electromyographic signal of the target muscle within a preset time length, wherein the electromyographic characteristic index is used for representing the change condition of the electromyographic signal corresponding to the target muscle; and predicting the electromyographic characteristic indexes corresponding to the target muscles through a state prediction network to obtain a state prediction result corresponding to the target object at the current moment, wherein the state prediction result is used for representing the fatigue degree of the target object. The method can improve the state prediction precision.

Description

State prediction method, state prediction device, computer equipment and storage medium

Technical Field

The present application relates to the field of data processing technologies, and in particular, to a state prediction method, apparatus, computer device, and storage medium.

Background

In the case where a person maintains a posture for a long time, the relevant muscles may be fatigued. In the case of oral implant surgery, the patient must maintain the mouth-open state for a long time during the surgery, and the related muscles may be fatigued. Muscle fatigue may not only cause accidental injury, but also may bring potential safety hazards to the implant operation.

Currently, in a medical scene, a doctor can judge the current fatigue state of a patient through the autonomous feedback of the patient or the operation duration. The mode of judging the fatigue state by a doctor or a patient according to subjective consciousness has a low judgment precision on the fatigue state of the patient because the judgment result is not objective enough.

Disclosure of Invention

In view of the above, it is necessary to provide a state prediction method, a state prediction apparatus, a computer device, and a storage medium capable of improving a state prediction result.

In a first aspect, the present application provides a method for predicting a state, the method comprising:

acquiring corresponding electromyographic signals aiming at target muscles of a target object;

determining an electromyographic characteristic index corresponding to the target muscle at the current moment according to the electromyographic signal of the target muscle within a preset time length, wherein the electromyographic characteristic index is used for representing the change condition of the electromyographic signal corresponding to the target muscle;

and predicting the electromyographic characteristic indexes corresponding to the target muscles through a state prediction network to obtain a state prediction result corresponding to the target object at the current moment, wherein the state prediction result is used for representing the fatigue degree of the target object.

In one embodiment, the predicting, by a state prediction network, a myoelectric characteristic indicator corresponding to the target muscle to obtain a state prediction result of the target object at the current time includes:

acquiring personal attribute information of the target object;

and predicting myoelectric characteristic indexes corresponding to the target muscles and personal attribute information of the target object through a state prediction network to obtain a state prediction result corresponding to the target object at the current moment.

In one embodiment, the determining, according to the electromyographic signal of the target muscle within a preset time length, an electromyographic characteristic index corresponding to the target muscle at the current time includes:

determining a reference myoelectric characteristic value corresponding to the target muscle according to the myoelectric signal acquired by the target muscle in a resting state;

dividing the electromyographic signals into a plurality of segments, and determining electromyographic characteristic values corresponding to the segments;

and determining an electromyographic characteristic index corresponding to the target muscle at the current moment according to the difference value between the electromyographic characteristic value corresponding to each segment and the reference electromyographic characteristic value.

In one embodiment, the state prediction network comprises a first network for identifying a normal state and a fatigue state and a second network for identifying a fatigue level corresponding to the fatigue state,

the predicting process is carried out on the myoelectric characteristic indexes corresponding to the target muscles through a state prediction network to obtain the state prediction result corresponding to the target object at the current moment, and the predicting process comprises the following steps:

inputting the electromyographic characteristic indexes corresponding to the target muscles into the first network for state prediction to obtain a first state prediction result;

when the first state prediction result represents that the current state is the fatigue state, inputting the electromyographic characteristic indexes corresponding to the target muscles into the second network for state prediction to obtain a second state prediction result, wherein the second state prediction result is used for representing the fatigue grade corresponding to the fatigue state;

taking the fatigue grade corresponding to the fatigue state represented by the second state prediction result as the state prediction result corresponding to the target object at the current moment; or,

and taking the normal state as a state prediction result corresponding to the target object at the current moment when the first state prediction result represents that the current state is the normal state.

In one embodiment, the method further comprises:

and determining the risk level of the target object at the current moment according to the state prediction result.

In one embodiment, determining the risk level of the target object at the current time according to the state prediction result includes:

determining the risk level of the target object at the current moment as a low risk level under the condition that the state prediction result represents that the current state is a normal state;

or determining the risk level of the target object at the current moment as a high risk level under the condition that the fatigue level corresponding to the state prediction result representing the fatigue state is a severe fatigue state;

or determining a time interval between a transition time and the current time under the condition that the fatigue grade corresponding to the current state represented by the state prediction result is a light fatigue state, wherein the transition time is the time when the target object is determined to be in the light fatigue state for the first time;

and performing risk prediction on the target object according to the time interval, the myoelectric characteristic index corresponding to the target object at the current moment and the personal attribute information of the target object to obtain the risk level of the target object at the current moment.

In one embodiment, the method further comprises:

acquiring a first sample myoelectric signal corresponding to the target muscle of a simulation object in a resting state;

acquiring a second sample electromyographic signal corresponding to the target muscle under the state that the simulation object maintains the target posture, and acquiring state information corresponding to the second sample electromyographic signal at different moments;

constructing a sample group according to the corresponding state information of the first sample electromyographic signal, the second sample electromyographic signal and the second sample electromyographic signal at different moments;

and constructing a training set according to each sample group, and training the state prediction network through the training set.

In one embodiment, the constructing a sample group according to the state information of the first sample electromyographic signal, the second sample electromyographic signal and the second sample electromyographic signal at different time includes:

dividing the second sample electromyographic signal corresponding to the target muscle into a plurality of sample segments;

for any sample segment, obtaining a sample electromyographic characteristic index corresponding to the target muscle in the sample segment according to a second sample electromyographic signal and the first sample electromyographic signal in the sample segment;

and constructing a sample group according to the corresponding sample electromyography characteristic index of the target muscle in the sample segment, the state information of the simulation object in the time period corresponding to the sample segment and the personal attribute information of the simulation object.

In one embodiment, the training of the state prediction network by the training set includes:

for a first sample group of which the state information in the training set is in a light fatigue state or a heavy fatigue state, marking the state information in the first sample group as a fatigue state to obtain a first training set;

training an initial network according to the first training set to obtain the first network;

performing state prediction processing on the sample electromyographic characteristic indexes in each sample group in the training set according to the first network to obtain a prediction state result corresponding to each sample group;

determining a second sample group of which the predicted state result is a fatigue state from each sample group;

and training the initial network according to each second sample group to obtain the second network.

In one embodiment, the method further comprises:

and displaying corresponding warning information according to the risk level of the target object at the current moment.

In a second aspect, the present application further provides a state prediction apparatus, comprising:

the first acquisition module is used for acquiring corresponding electromyographic signals aiming at target muscles of a target object;

the first determination module is used for determining an electromyographic characteristic index corresponding to the target muscle at the current moment according to an electromyographic signal of the target muscle within a preset time length, wherein the electromyographic characteristic index is used for representing the change condition of the electromyographic signal corresponding to the target muscle;

and the prediction module is used for performing prediction processing on the myoelectric characteristic indexes corresponding to the target muscles through a state prediction network to obtain a state prediction result corresponding to the target object at the current moment, and the state prediction result is used for representing the fatigue degree of the target object.

In one embodiment, the prediction module is further configured to:

acquiring personal attribute information of the target object;

In one embodiment, the first determining module is further configured to:

In one embodiment, the state prediction network includes a first network and a second network, the first network is configured to identify a normal state and a fatigue state, the second network is configured to identify a fatigue level corresponding to the fatigue state, and the prediction module is further configured to:

when the first state prediction result represents that the current state is the fatigue state, inputting the electromyographic characteristic indexes corresponding to the target muscles into the second network for state prediction to obtain a second state prediction result, wherein the second state prediction result is used for representing the fatigue grade corresponding to the state;

In one embodiment, the apparatus further comprises:

and the second determination module is used for determining the risk level of the target object at the current moment according to the state prediction result.

In one embodiment, the second determining module is further configured to:

or determining a time interval between a transition time and the current time when the fatigue grade corresponding to the fatigue state represented by the state prediction result is a light fatigue state, wherein the transition time is the time when the target object is determined to be in the light fatigue state for the first time;

In one embodiment, the apparatus further comprises:

the second acquisition module is used for acquiring a first sample myoelectric signal corresponding to the target muscle of the simulation object in a resting state;

the third acquisition module is used for acquiring a second sample electromyographic signal corresponding to the target muscle under the condition that the simulation object maintains the target posture and acquiring corresponding state information of the second sample electromyographic signal at different moments;

the construction module is used for constructing a sample group according to the first sample electromyographic signal, the second sample electromyographic signal and corresponding state information of the second sample electromyographic signal at different moments;

and the training module is used for constructing a training set according to each sample group and training the state prediction network through the training set.

In one embodiment, the building module is further configured to:

and constructing a sample group according to the corresponding sample myoelectric characteristic index of the target muscle in the sample segment, the state information of the simulation object in the time period corresponding to the sample segment and the personal attribute information of the simulation object.

In one embodiment, the state prediction network includes a first network and a second network, and the training module is further configured to:

In one embodiment, the apparatus further comprises:

and the display module is used for displaying corresponding warning information according to the risk level of the target object at the current moment.

In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor implementing the steps of any of the above state prediction methods when the processor executes the computer program.

In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium, on which a computer program is stored, which computer program, when being executed by a processor, carries out the steps of any of the above state prediction methods.

In a fifth aspect, the present application further provides a computer program product. The computer program product comprising a computer program that when executed by a processor implements the steps of any of the above state prediction methods.

According to the state prediction method, the state prediction device, the computer equipment and the storage medium, the corresponding electromyographic signals are collected aiming at the target muscles of the target object, and the electromyographic characteristic indexes corresponding to the target muscles at the current moment are determined according to the electromyographic signals of the target muscles within the preset time length, wherein the electromyographic characteristic indexes are used for representing the change conditions of the electromyographic signals corresponding to the target muscles. Furthermore, the electromyographic characteristic indexes corresponding to the target muscles can be subjected to prediction processing through a state prediction network, so that a state prediction result corresponding to the target object at the current moment is obtained, and the state prediction result is used for representing the fatigue degree of the target object. Based on the state prediction method, the state prediction device, the computer equipment and the storage medium, the electromyographic characteristic index of the target muscle can be determined through the acquired electromyographic signal of the target muscle, then the electromyographic characteristic index is subjected to prediction processing through the state prediction network, a state prediction result is obtained, the state prediction result is not affected by subjective consciousness of people, the state prediction method, the state prediction device, the computer equipment and the storage medium are more objective, and the accuracy of state prediction can be improved.

Drawings

FIG. 1 is a flow diagram illustrating a method for state prediction in one embodiment;

FIG. 2 is a schematic illustration of a surface electrode arrangement in one embodiment;

FIG. 3 is a diagram illustrating an electromyographic signal denoising process in an embodiment;

FIG. 4 is a diagram illustrating an electromyographic signal denoising process in an embodiment;

FIG. 5 is a flowchart illustrating the method step 106 of the state prediction in one embodiment;

FIG. 6 is a flowchart illustrating the method step 104 of the state prediction method in one embodiment;

FIG. 7 is a diagram illustrating an exemplary method for obtaining electromyographic feature values;

FIG. 8 is a diagram illustrating obtaining electromyographic feature values according to an embodiment;

FIG. 9 is a flowchart illustrating the method step 106 of the state prediction in one embodiment;

FIG. 10 is a diagram illustrating a state prediction method in accordance with one embodiment;

FIG. 11 is a flow diagram that illustrates a method for state prediction, according to one embodiment;

FIG. 12 is a diagram of a training state prediction network in one embodiment;

FIG. 13 is a flowchart illustrating the method step 1106 of the state prediction method in one embodiment;

FIG. 14 is a flowchart illustrating the method step 1108 of the state prediction method in one embodiment;

FIG. 15 is a schematic diagram of training a first network in one embodiment;

FIG. 16 is a diagram of training a second network in one embodiment;

FIG. 17 is a schematic flow chart illustrating the evaluation of risk levels in one embodiment;

FIG. 18 is a diagram illustrating an exemplary alert message;

FIG. 19 is a flow diagram illustrating an embodiment of displaying alert information;

FIG. 20 is a flow diagram illustrating alert information in one embodiment;

FIG. 21 is a flow diagram that illustrates a method for state prediction, according to one embodiment;

FIG. 22 is a block diagram showing the structure of a state prediction device according to an embodiment;

FIG. 23 is a diagram illustrating an internal structure of a computer device according to an embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

In an embodiment, as shown in fig. 1, a state prediction method is provided, and this embodiment is illustrated by applying the method to a terminal, and it is to be understood that the method may also be applied to a server, and may also be applied to a system including the terminal and the server, and is implemented by interaction between the terminal and the server. In this embodiment, the method includes the steps of:

step 102, collecting corresponding electromyographic signals aiming at target muscles of a target object.

In the embodiment of the present application, the target object may be an object to be currently subjected to state estimation, and the target muscle may be a muscle associated with a current operation. For example, in the case of oral implantation, the target object may be a patient, and since the patient is required to maintain the maximum opening state for a long time during the oral implantation, the target muscles selected for monitoring are muscles related to opening and chewing, which may include temporalis, masseter, and sternocleidomastoid muscles. In the following embodiments, the oral implant operation scenario is taken as an example, and target muscles include a temporal muscle, a masseter and a sternocleidomastoid muscle, which are described in the embodiments of the present application.

The surface electrode is arranged on the surface of the target muscle, and the electromyographic signals corresponding to the target muscle are collected through the surface electrode. Illustratively, in the case where the target muscles include a temporal muscle, a masseter muscle and a sternocleidomastoid muscle, as shown in fig. 2, surface electrodes may be provided on the surfaces of the temporal muscle, the masseter muscle and the sternocleidomastoid muscle, respectively, to acquire electromyographic signals of the temporal muscle, the masseter muscle and the sternocleidomastoid muscle, respectively. In the embodiment of the present application, the surface electrode may be connected to the terminal in a wired manner or in a wireless manner, and the specific connection manner is not specifically limited in the embodiment of the present application.

In a possible implementation manner, referring to fig. 3, after acquiring an electromyographic signal corresponding to a target muscle, denoising/denoising may be performed on the electromyographic signal, all the electromyographic signals related in the following embodiments are electromyographic signals subjected to denoising/denoising, and no special description is provided in the following embodiments of the present application.

The electromyographic signals are low-frequency signals generally, so that the electromyographic signals with lower frequency can be retained by filtering high-frequency signals, and then denoising/noise reduction of the electromyographic signals is realized. Illustratively, referring to fig. 4, the electromyographic signals may be denoised based on a wavelet transform. It should be noted that wavelet transform denoising is only used as an example for denoising the electromyographic signals in the embodiment of the present application, and actually, the embodiment of the present application does not specifically limit the electromyographic signal denoising method, and all the methods that denoising or denoising the electromyographic signals can be applied to the embodiment of the present application.

And 104, determining an electromyographic characteristic index corresponding to the target muscle at the current moment according to the electromyographic signal of the target muscle within a preset time length, wherein the electromyographic characteristic index is used for representing the change condition of the electromyographic signal corresponding to the target muscle.

In the embodiment of the present application, the state of the target object may be predicted by an electromyographic signal within a preset duration, where the preset duration may be a preset duration, and a specific value may be determined by a person skilled in the art according to a prediction requirement, for example: the preset time duration can be 500ms, 800ms, 1000ms and the like. Assuming that the preset time length is 500ms, after each 500ms of electromyographic signals are collected, an electromyographic characteristic index corresponding to the target muscle corresponding to the electromyographic signal at the current time may be determined based on the electromyographic signal within 500ms, for example: the electromyographic characteristic index corresponding to the target muscle in 500ms can be determined for the electromyographic signals within 0ms to 500ms, and the electromyographic characteristic index corresponding to the target muscle in 1000ms can be determined for the electromyographic signals within 500ms to 1000 ms.

The electromyographic characteristic index can be information which is determined according to the electromyographic signal and can represent the change condition of the electromyographic signal within a preset time length.

Exemplarily, in the process of fatigue, a Frequency spectrum curve of an electromyographic signal is shifted left, and both a Frequency domain characteristic index MF (Median Frequency) and an MPF (Mean Power Frequency) are in a descending trend; with the left shift of the spectrum curve, the spectrum energy will move to the low frequency direction, so that the distribution of the electromyographic signals in the frequency domain is more concentrated, the complexity of the electromyographic structure is reduced, and the BSE (Band Spectral Entropy) will also be in a descending trend. Meanwhile, in the process of fatigue, the low-frequency information of the electromyographic signals is increased, and due to the low-pass filtering effect of body tissues, more electromyographic signal energy can be transmitted through the body tissues and is shown as the rising of the index RMS (Root Mean Square) reflecting the amplitude of the electromyographic signals. Therefore, the myoelectric characteristic index of the target muscle can be determined by selecting the change of the group of myoelectric characteristics of RMS, MF, MPF and BSE, in this example, each myoelectric characteristic of RMS, MF, MPF and BSE can obtain a corresponding myoelectric characteristic index, that is, the myoelectric characteristic index obtained by the single target muscle based on RMS, MF, MPF and BSE may include: RMS index, MF index, MPF index, and BSE index. And then the fatigue state of the target object can be evaluated through the electromyographic characteristic indexes corresponding to the electromyographic signals.

It should be noted that the selection of the electromyographic features RMS, MF, MPF, and BSE to calculate the electromyographic feature signal is only an example of the electromyographic feature signal in the embodiment of the present application, and actually, the evaluation of the muscle fatigue state may be performed based on a time domain, a frequency domain, or a time-frequency combined index, that is, a group of features may be arbitrarily selected from the time domain and the frequency domain to evaluate the muscle fatigue state.

The number of target muscles is not specifically limited in the embodiments of the present application, for example: under the condition that the target muscles comprise temporal muscles, masseter muscles and sternocleidomastoid muscles, a group of electromyographic characteristic indexes can be determined for each target muscle respectively, and then the fatigue degree of the target object can be predicted according to the electromyographic characteristic indexes of all the target muscles.

And 106, performing prediction processing on the myoelectric characteristic indexes corresponding to the target muscles through a state prediction network to obtain a state prediction result corresponding to the target object at the current moment, wherein the state prediction result is used for representing the fatigue degree of the target object.

In the embodiment of the application, after the electromyographic characteristic index corresponding to the target muscle is obtained, the fatigue degree of the target object at the current moment can be determined according to the electromyographic characteristic index corresponding to the target muscle. For example, the electromyographic characteristic index corresponding to the target muscle may be input to a state prediction network to perform state prediction processing, so as to obtain a state prediction result corresponding to the target object at the current time.

The state prediction network is a pre-trained neural network for predicting the fatigue degree of the target object, the network structure of the state prediction network is not specifically limited in the embodiment of the application, and all the neural networks capable of predicting the fatigue degree of the target object based on the electromyographic characteristic indexes are suitable for the embodiment of the application.

In one example, the state prediction result may be used to characterize the fatigue degree of the target object, for example: the state prediction result may include a result for characterizing a normal state in which the user does not feel tired, a result for characterizing that the user is in a light tired state, and a result for characterizing that the user is in a heavy tired state.

It should be noted that the normal state, the light fatigue state, and the heavy fatigue state may be taken as an example of dividing the fatigue degree of the target object in the embodiment of the present application, and actually, the fatigue degree of the target object may be divided more coarsely or more finely based on the predicted demand, for example: the normal state, the light fatigue state, and the heavy fatigue state are further classified into different grades, and the specific expression form of the fatigue degree of the target object is not specifically limited in the embodiment of the present application.

According to the state prediction method, the corresponding electromyographic signals are collected aiming at the target muscles of the target object, and after the electromyographic signals with preset time length are obtained through collection, the electromyographic characteristic indexes corresponding to the target muscles at the current moment are determined according to the electromyographic signals of the target muscles within the preset time length, and the electromyographic characteristic indexes are used for representing the change conditions of the electromyographic signals corresponding to the target muscles. Furthermore, the electromyographic characteristic indexes corresponding to the target muscles can be subjected to prediction processing through a state prediction network, so that a state prediction result corresponding to the target object at the current moment is obtained, and the state prediction result is used for representing the fatigue degree of the target object. Based on the state prediction method provided by the embodiment of the application, the electromyographic characteristic index of the target muscle can be determined through the acquired electromyographic signal of the target muscle, and then the electromyographic characteristic index is subjected to prediction processing through the state prediction network to obtain the state prediction result, so that the state prediction result is not influenced by subjective consciousness of people, is more objective, and can improve the accuracy of state prediction.

In an embodiment, as shown in fig. 5, in the step 106, performing prediction processing on the myoelectric characteristic indicator corresponding to the target muscle through a state prediction network to obtain a state prediction result corresponding to the target object at the current time includes:

step 502, acquiring personal attribute information of a target object;

and step 504, performing prediction processing on the electromyographic characteristic indexes corresponding to the target muscles and the personal attribute information of the target object through a state prediction network to obtain a state prediction result corresponding to the target object at the current moment.

In the embodiment of the application, as the age, the sex and the related medical history of the patient all have certain influence on the physical state, taking the age as an example, the elderly are more likely to feel tired compared with teenagers; for example, women are more fatigued than men, by gender; taking the related medical history as an example, people with the related medical history are more likely to feel tired. Therefore, the state prediction result corresponding to the target object at the current moment can be further predicted by combining the personal attribute information of the target object. The personal attribute information may include: the gender, age, related medical history and the like can generate information with certain influence on the physical state of the target object, and the embodiment of the application does not specifically limit personal attribute information.

The terminal display interface can comprise an information acquisition interface, and a doctor can input personal attribute information of the target object through the information acquisition interface; or, the personal attribute information of the target object may be acquired from the electronic medical record by associating the electronic medical record, and the acquisition manner of the personal attribute information is not particularly limited in this embodiment of the application.

After obtaining the personal attribute information of the target object, the personal attribute information may also be used as an input of a state prediction network (that is, in the training process of the state prediction network, the state prediction network is trained by combining the myoelectric characteristic index and the personal attribute information of the target object), and then the state prediction network may predict a state prediction result of the target object by combining the myoelectric characteristic index of the target muscle and the personal attribute information of the target object.

According to the state prediction method provided by the embodiment of the application, the fatigue state result of the target object can be predicted by combining the personal attribute information of the target object, and the accuracy of state prediction can be further improved.

In an embodiment, as shown in fig. 6, in the step 104, determining the corresponding electromyographic characteristic indicator of the target muscle at the current time according to the electromyographic signal of the target muscle within the preset time duration includes:

step 602, determining a reference electromyogram characteristic value corresponding to a target muscle according to an electromyogram signal acquired by the target muscle in a resting state;

step 604, dividing the electromyographic signal into a plurality of segments, and determining an electromyographic characteristic value corresponding to each segment;

step 606, determining the electromyographic characteristic index of the target muscle corresponding to the current moment according to the difference value between the electromyographic characteristic value corresponding to each segment and the reference electromyographic characteristic value.

In the embodiment of the application, the electromyographic signals of the target muscles of the target object in the resting state can be collected in advance. For example, in the oral implantation surgery, if the patient needs to continuously maintain the maximum mouth-opening state, the electromyographic signals of the target muscles of the patient can be collected in advance when the patient is not mouth-opened. After acquiring the electromyographic signals of the target muscles, the reference electromyographic characteristic values corresponding to the target muscles can be determined based on the electromyographic signals of the target muscles.

The electromyographic signal of the target muscle collected within the preset time length can be divided into a plurality of segments, for each segment, the electromyographic characteristic value corresponding to each segment can be respectively determined, the difference value between the electromyographic characteristic value of each segment and the reference electromyographic characteristic value is respectively determined, straight line fitting is carried out on each difference value according to time, and the change situation of the straight line slope obtained by fitting along with the time is respectively used as the electromyographic characteristic index corresponding to the target muscle at the current moment.

It should be noted that the electromyographic signals may be segmented by adopting an equal-length partitioning or a random partitioning method, and the electromyographic signals are segmented by the method in this embodiment of the application without specific limitation.

Illustratively, the electromyographic features are again taken to be RMS, MF, MPF, and BSE. Referring to fig. 7, RMS, MF, MPF and BSE may be respectively determined as reference electromyogram characteristic values corresponding to the target muscle according to the electromyogram signals collected in the resting state, where the following formula (i) may be referred to for the determination process of RMS, the following formula (ii) may be referred to for the determination process of MF, the following formula (iii) may be referred to for the determination process of MPF, and the following formula (iv) may be referred to for the determination process of BSE.

Where T is the integration variable and T is the time interval over which the integration is calculated.

Wherein PS (f) represents the frequency spectrum of the electromyographic signal, f₁And f₂Two end-point values representing the frequency range of the electromyographic signals, f is an integration variable.

Selecting appropriate frequency bandwidth (for example, selecting 5Hz as frequency bandwidth), decomposing the collected electromyographic signal frequency band into n sub-frequency bands, and calculating the power spectrum energy of each sub-frequency band, and recording the power spectrum energy as E_iWhere i is 1, …, n, for identifying the ith sub-band. Summing the functional spectrum energy of each sub-band to obtain the total spectrum energy

Wherein E represents the total spectral energy. Normalizing the power spectrum energy of each sub-band according to the total spectrum energy to obtain the probability density distribution of the power spectrum energy of the sub-bands

Wherein, P_iFor characterizing the probability density distribution of the ith sub-band. By integrating the relative band spectrum energy, the corresponding band spectrum entropy can be further obtained according to the definition of the Shannon entropy, and the formula (IV) is referred to.

After obtaining the electromyographic signals of the target muscles in the resting state, the corresponding RMS, MF, MPF and BSE can be determined as the reference electromyographic characteristic values by referring to the formula (i), the formula (ii), the formula (iii) and the formula (iv), respectively.

For the electromyographic signals of the target muscles collected within the preset time length, the electromyographic signals can be divided into a plurality of segments, and for each segment, the RMS, MF, MPF and BSE corresponding to each segment are determined by respectively referring to the formula (I), the formula (II), the formula (III) and the formula (IV) and are used as the electromyographic characteristic values corresponding to each segment. And respectively calculating the difference values of the RMS, MF, MPF and BSE of each segment and the RMS, MF, MPF and BSE in the reference electromyogram characteristic value. Taking RMS as an example, the difference between the RMS in the electromyographic feature values of each segment and the RMS in the reference electromyographic feature value may be calculated, and each difference may be linearly fitted according to time, so as to obtain an electromyographic feature index corresponding to the RMS. By analogy, myoelectric characteristic indexes corresponding to other myoelectric characteristics (MF, MPF, and BSE) can be obtained, and as shown in fig. 8, a group of myoelectric characteristic indexes corresponding to the target muscle at the current time can be obtained.

The myoelectric characteristic indexes (in the case that the target object includes a plurality of target muscles, here, the myoelectric characteristic indexes corresponding to the target muscles) are subjected to prediction processing through a state prediction network, so that a state prediction result corresponding to the target object at the current time can be obtained.

Based on the state prediction method provided by the embodiment of the application, the electromyographic characteristic index of the target muscle can be determined through the acquired electromyographic signal of the target muscle, and then the electromyographic characteristic index is subjected to prediction processing through the state prediction network to obtain the state prediction result, so that the state prediction result is not influenced by subjective consciousness of people, is more objective, and can improve the accuracy of state prediction.

In one embodiment, referring to fig. 9, the state prediction network includes a first network and a second network, the first network is used to identify a normal state and a fatigue state, and the second network is used to identify a fatigue level corresponding to the fatigue state, in the step 106, the obtaining of the state prediction result corresponding to the target object at the current time by performing prediction processing on the myoelectric characteristic index corresponding to the target muscle through the state prediction network includes:

step 902, inputting the electromyographic characteristic indexes corresponding to the target muscles into a first network for state prediction to obtain a first state prediction result;

step 904, inputting the electromyographic characteristic indexes corresponding to the target muscles into a second network for state prediction to obtain a second state prediction result under the condition that the first state prediction result represents that the current state is the fatigue state, wherein the second state prediction result is used for representing the fatigue grade corresponding to the fatigue state;

step 906, taking the fatigue grade corresponding to the fatigue state represented by the second state prediction result as a state prediction result corresponding to the target object at the current moment;

and 908, taking the normal state as the state prediction result corresponding to the target object at the current moment when the first state prediction result represents that the current state is the normal state.

In this embodiment of the present application, the first network and the second network may be networks for the second classification, and the network structure of the first network and the second network is not specifically limited in this embodiment of the present application. In an example, the state prediction network may be a two-stage soft-interval Support Vector Machine (SVM) based method, since the fatigue change of the muscle is a gradual process, absolute separability cannot be guaranteed between states, and a certain degree of transition is required for distinguishing different states, a first network in the first stage may be a first SVM (Support Vector Machine), and a second network in the second stage may be a second SVM. In the first phase, the first SVM needs to distinguish between a normal state and a fatigue state, and in the second phase, the second SVM needs to distinguish between fatigue levels corresponding to the fatigue state, and the fatigue levels can comprise a light fatigue state and a heavy fatigue state.

After obtaining the electromyographic characteristic index corresponding to the target muscle, the electromyographic characteristic index corresponding to the target muscle (in the case that the electromyographic characteristic index includes a plurality of electromyographic characteristic indexes, a group of electromyographic characteristic indexes is represented here) may be input into the first network to perform state prediction, so as to obtain a first state prediction result, where the first state prediction result may include a result representing that the current state is a fatigue state or a result representing that the current state is a normal state.

When the first state prediction result represents that the current state is a fatigue state, the electromyographic characteristic index needs to be further divided by the second network to identify the fatigue level of the fatigue state of the target object, for example: and identifying whether the target object is in a light fatigue state or a heavy fatigue state at present, outputting a corresponding second prediction result, and taking the fatigue grade corresponding to the fatigue state represented by the second prediction result as the state prediction result of the target object at the current moment.

In the embodiment of the present application, as shown in fig. 10, when the first state prediction result indicates that the current state is the normal state, the normal state may be directly used as the state prediction result corresponding to the target object at the current time, and the state prediction result is output, and when the first state prediction result indicates that the current state is the fatigue state, the myoelectric characteristic index corresponding to the target muscle is input to the second network for prediction, so as to obtain the state prediction result indicating the fatigue level corresponding to the fatigue state, for example: and obtaining a state prediction result corresponding to the light fatigue state or the heavy fatigue state corresponding to the characteristic fatigue state.

Based on the state prediction method provided by the embodiment of the application, the electromyographic characteristic indexes of the target object can be predicted by a two-stage soft interval vector machine method, the state prediction result is obtained, the prediction process is more in line with the gradual change characteristic of the fatigue state, the method is more objective, and the accuracy of state prediction can be further improved.

The state prediction network may be a network obtained by pre-training, and in the training process, a training set may be pre-constructed, so as to train the state prediction network through the training set.

In one embodiment, referring to fig. 11, the method may further include:

step 1102, collecting a first sample myoelectric signal corresponding to a target muscle of a simulation object in a resting state;

1104, collecting a second sample electromyographic signal corresponding to the target muscle under the state that the simulation object maintains the target posture, and acquiring state information corresponding to the second sample electromyographic signal at different moments;

step 1106, constructing a sample group according to the corresponding state information of the first sample electromyographic signal, the second sample electromyographic signal and the second sample electromyographic signal at different moments;

step 1108, a training set is established according to each sample group, and the state prediction network is trained through the training set.

In the embodiment of the application, the surface electrode can be placed on the surface of the target muscle of the simulation object to acquire the electromyographic signal. A section of electromyographic signals of the simulation object in a resting state can be collected firstly to serve as first sample electromyographic signals, then the simulation object is enabled to keep the target posture until the simulation object cannot insist, and the electromyographic signals in the process are collected to serve as second sample electromyographic signals. In the process of acquiring the electromyographic signals of the second sample, subjective feelings (no feeling, fatigue feeling or incapability of maintaining posture) of the simulation object can be recorded, and the corresponding state information of the electromyographic signals of the second sample at different moments can be determined according to the recorded subjective feelings (for example, no feeling corresponds to a normal state, fatigue feeling corresponds to a slight fatigue state, and incapability of maintaining posture corresponds to a severe fatigue state). Meanwhile, personal attribute information of the simulation object can be recorded.

Furthermore, the sample group may be formed according to the corresponding state information of the first sample electromyographic signal, the second sample electromyographic signal and the second sample electromyographic signal of the simulation object at different time (or may further include personal attribute information of the simulation object). The method comprises the steps of carrying out electromyographic signal acquisition on a plurality of simulation objects, constructing a plurality of sample groups, and further constructing a training set through the plurality of sample groups. Illustratively, referring to FIG. 12, a state prediction network may be trained by a training set.

In one example, in the process of training the state prediction network through the training set, in each training process, a sample electromyographic characteristic index can be determined through a first sample electromyographic signal and a second sample electromyographic signal, the sample electromyographic characteristic index is used as the input of the state prediction network, network parameters of the state prediction network are adjusted according to the difference of state information corresponding to the output corresponding state prediction result and the second sample electromyographic signal at different moments, so that the training of the current round is completed, and the training is stopped until the state prediction network reaches the training requirement after multiple rounds of training.

Based on the state prediction method provided by the embodiment of the application, the state prediction network for predicting the state of the object through the electromyographic signals can be trained, the prediction processing can be further carried out through the state prediction network and the acquired electromyographic signals, the state prediction result is obtained, the state prediction result is not influenced by subjective consciousness of people, the method is more objective, and the accuracy of state prediction can be improved.

In one embodiment, referring to fig. 13, in step 1106, constructing a sample group according to the state information of the first sample electromyographic signal, the second sample electromyographic signal and the second sample electromyographic signal corresponding to different time instants, where the constructing includes:

step 1302, dividing a second sample electromyographic signal corresponding to the target muscle into a plurality of sample segments;

step 1304, aiming at any sample segment, obtaining a sample electromyographic characteristic index corresponding to a target muscle in the sample segment according to a second sample electromyographic signal and a first sample electromyographic signal in the sample segment;

step 1306, constructing a sample group according to the corresponding sample electromyogram characteristic index of the target muscle in the sample segment, the state information of the simulation object in the time period corresponding to the sample segment, and the personal attribute information of the simulation object.

In the embodiment of the application, in the process of constructing the sample group, the second sample electromyographic signal corresponding to the target muscle can be divided into a plurality of sample segments, the division rule for the sample segments is not specifically limited, and the segment lengths may be the same or different. After the sample segments are divided, the electromyographic characteristic indexes of the samples of the sample segments can be determined according to the first sample electromyographic signals and the sample segments. The process of determining the electromyographic characteristic index of the sample corresponding to each sample segment may refer to the process of determining the electromyographic characteristic index, which is not described herein in detail in the embodiments of the present application.

After the sample electromyography characteristic index corresponding to each sample segment is determined, for any sample segment, a sample group may be constructed based on the sample electromyography characteristic index of the sample segment, the state information of the simulation object corresponding to the sample segment, and the personal attribute information of the simulation object. That is, a plurality of sample groups may be determined by a plurality of sample segments into which the second sample electromyogram signal is divided.

A training set can be constructed through a plurality of sample groups constructed through the electromyographic signals of the second samples, and then a state prediction network for predicting the fatigue state of the target object through the electromyographic signal indexes can be trained through the training set.

In one embodiment, referring to fig. 14, the state prediction network comprises a first network and a second network, and training 1108 the state prediction network with a training set may comprise:

step 1402, labeling the state information in the first sample group as a fatigue state to obtain a first training set, wherein the state information in the training set is a light fatigue state or a heavy fatigue state;

step 1404, training the initial network according to the first training set to obtain a first network;

1406, performing state prediction processing on the sample electromyographic characteristic indexes in each sample group in the training set according to the first network to obtain a prediction state result corresponding to each sample group;

step 1408, determining a second sample group of which the predicted state result is the fatigue state from the sample groups;

step 1410, training the initial network according to each second sample group to obtain a second network.

In the embodiment of the present application, the state prediction network may include a first network and a second network, and the first network and the second network may be, for example, support vector machines. The first network is used for distinguishing a normal state and a fatigue state, and the second network is used for distinguishing a light fatigue state and a heavy fatigue state. In the training process of the state prediction network, the first network and the second network may be trained synchronously, or the first network may be trained in advance and then the second network may be trained, which is not specifically limited in this embodiment of the application.

Illustratively, the first network is trained prior to the second network. Because the first network is used for distinguishing a normal state from a fatigue state, a training set can be copied, a first sample group with the state information of the simulation object being a light fatigue state and a heavy fatigue state is determined from the copied training set, the state information of the simulation object in the first sample group is relabeled to the fatigue state, a first training set is obtained, and the sample group in the first training set corresponds to the two states: normal state and fatigue state.

In the process of training the first network, as shown in fig. 15, the initial network may be trained through the myoelectric characteristic index of the target muscle in the sample group and the state information of the simulation object in the sample group (or may also include the personal attribute information of the simulation object, which is marked in fig. 15, but it should be understood that the personal attribute information of the simulation object is not necessary information, that is, the state prediction network in the embodiment of the present application is a loosely coupled structure), so as to obtain a corresponding parameter set, and obtain the trained first network according to the parameter set. For example, the electromyographic feature information of each sample (or the personal attribute information of each simulation object may also be included) may be subjected to prediction processing by using an initial network to obtain a state prediction result of each simulation object, and after network parameters of the initial network are adjusted according to a difference between the state prediction result of each simulation object and the state information of each simulation object, the previous process is continuously repeated to perform the next round of training until a first network is obtained after multiple rounds of iterative training.

Referring to fig. 16, after the first network is obtained by training, the prediction processing may be performed on the sample groups in the training set by the first network, so as to obtain the prediction state results corresponding to each sample group. And determining a second sample group of which the state prediction result is characterized as the fatigue state from the training set, and training the initial network through the second sample group to obtain a second network. For the training process of the second network, reference may be made to the training process of the first network, which is not described in detail herein.

After the training of the second network is completed, the training of the state prediction network is completed, and then the myoelectric characteristic index of the target object (or the myoelectric characteristic index may also include the personal attribute information of the target object) may be subjected to prediction processing by the trained state prediction network, so as to obtain a state prediction result of the target object.

Based on the state prediction method provided by the embodiment of the application, a two-stage state prediction network based on a soft interval vector machine can be trained to perform prediction processing on the electromyographic characteristic indexes of the target object to obtain a state prediction result, the prediction process is more in line with the gradual change characteristic of the fatigue state, the method is more objective, and the accuracy of state prediction can be further improved.

In one embodiment, the method may further include:

In the embodiment of the application, after the state prediction result corresponding to the target object at the current moment is obtained through prediction, the risk level of the target object at the current moment can be further determined according to the state prediction result. For example, the risk levels may be classified into a low risk level, a medium risk level and a high risk level, wherein the low risk level is used for representing that the target object can continuously maintain the current posture at present, the medium risk level is used for representing that the target object can maintain the current posture but needs to observe the state of the target object at any moment, and the high risk level is used for representing that the target object cannot continuously maintain the current posture.

It should be noted that the low risk level, the medium risk level, and the high risk level are only used as an example for dividing the risk levels in the embodiment of the present application, and actually, the risk levels may be divided according to a larger or smaller granularity according to a requirement, which is not specifically limited in the embodiment of the present application.

In an embodiment, the determining the risk level of the target object at the current time according to the state prediction result includes:

or determining the risk level of the target object at the current moment as a high risk level under the condition that the fatigue level corresponding to the fatigue state represented by the state prediction result is a severe fatigue state;

or determining a time interval between a transition time and the current time under the condition that the fatigue grade corresponding to the fatigue state represented by the state prediction result is a light fatigue state, wherein the transition time is the time when the target object is determined to be in the light fatigue state for the first time;

and performing risk prediction on the target object according to the time interval, the myoelectric characteristic index corresponding to the target object at the current moment and the personal attribute information of the target object to obtain the risk grade of the target object at the current moment.

In the embodiment of the application, under the condition that the state prediction result of the target object indicates that the current state is the normal state, the risk level of the target object at the current moment can be determined to be the low risk level; or, under the condition that the state prediction result of the target object represents that the current state is a fatigue state and the fatigue level corresponding to the fatigue state is a severe fatigue state, determining that the risk level of the target object at the current moment is a high risk level; alternatively, when the state prediction result of the target object indicates that the current state is the fatigue level and the fatigue level corresponding to the fatigue state is the light fatigue state, the risk level of the target object at the current time may be further determined.

For example, referring to fig. 17, in the embodiment of the present application, a state prediction result of a target object may be evaluated through a pre-constructed risk prediction model, so as to obtain a risk level of the target object at a current time. Under the condition that the target object is in a mild fatigue state, whether a high risk prediction condition is met or not can be judged, under the condition that the high risk prediction condition is met, the target object can be determined to be in a high risk level currently, and otherwise, the target object can be determined to be in a medium risk level currently.

The second electromyographic signals of the simulation objects can be predicted according to the state prediction network to obtain the time t0 when the simulation objects are in a light fatigue state, the time t1 when the simulation objects are in a heavy fatigue state, and the corresponding electromyographic characteristic indexes when the simulation objects are in the heavy fatigue state. Determining the time interval delta T between T0 and T1, and establishing the time interval delta T and the electromyographic signal index X according to the personal attribute information of each simulation object_ja(where a is used to identify the electromyographic signature, in this example the electromyographic signature indicators include the RMS indicator, MF indicator, MPF indicator, and BSE indicator, then a ═ 1, 2, 3, 4]J is used to identify the target muscle, which in this example includes the temporalis muscle, the masseter muscle, and the sternocleidomastoid muscle, and j is then [1, 2, 3 ═ d]) And the association relationship between the personal attribute information. Taking the example that the personal attribute information includes age, Δ T and electromyographic signal index X can be established_jaThe corresponding relation between them.

For example, when it is determined that the state prediction result of the target object indicates that the current state is a fatigue state and the fatigue level corresponding to the fatigue state is a light fatigue state, the risk prediction model may obtain the time interval Δ T and the myoelectric signal index X corresponding to the personal attribute information of the target object from the corresponding relationship_ja。

In a risk prediction mode, the time when the target object is determined to be in the mild fatigue state for the first time can be obtained, and the risk level of the target object at the current time can be determined according to the proportional relation between the time interval delta T and the time interval delta T between the time and the current time. For example: in the case that the proportional relation between the time interval Δ T and the time interval Δ T is greater than or equal to a first proportional threshold (e.g., 95%), it may be determined that the target object satisfies the high risk prediction condition, and it may be determined that the target object is at a high risk level at the current time; otherwise, determining that the target object does not meet the high risk prediction condition, and determining that the target object is at a medium risk level at the current moment.

In another risk prediction mode, myoelectric information of the current time of the target object can be acquiredNumber index x_jaAnd according to the electromyographic signal index x_jaAnd X_jaAnd determining the risk level of the target object at the current moment according to the proportional relation between the target object and the target object. For example: if at least one target muscle has at least two electromyographic signal indexes x in a group of electromyographic signals corresponding to each target muscle_jaAnd X_jaThe proportional relation between the target object and the target object is greater than or equal to a second proportional threshold, so that the target object can be determined to meet a high risk prediction condition, and the target object can be determined to be in a high risk level at the current moment; otherwise, it is determined that the target object does not satisfy the high risk prediction condition, and it may be determined that the target object is at a medium risk level at the present time.

In the two risk prediction modes, the target object can be determined to be at the high risk level at the current moment when any high risk level judgment condition is met. The first ratio threshold and the second ratio threshold may be preset values, specific values may be set by a person skilled in the art according to prediction accuracy, which is not specifically limited in this embodiment of the present application, and the first ratio threshold and the second ratio threshold may be the same values or different values.

It should be noted that the above risk prediction model is only an example of the risk prediction model in the embodiment of the present application, and actually, the embodiment of the present application does not specifically limit the form of the risk prediction model, and all models that can predict the risk level of the target object based on the state prediction result of the target object and information such as the myoelectric signal, the myoelectric characteristic index, and the personal attribute information are applicable to the embodiment of the present application, for example: the risk prediction model can be a model constructed based on a deep neural network.

Based on the state prediction method provided by the embodiment of the application, the risk level of the target object at the current moment can be further predicted according to the state prediction result corresponding to the predicted target object at the current moment, the target object can be timely adjusted based on the predicted risk level, and the safety of the target object can be ensured.

In one embodiment, the method further comprises:

In the embodiment of the present application, the warning information may include and is not limited to: speech, signal light, vibration, characters, images, etc. Referring to fig. 18, the embodiment of the present application may be applied to different scenarios, and the object receiving the warning information may be an object such as a doctor, an implantation navigation system, or an auxiliary mechanical arm system. The display mode of the specific warning information is different according to different application scenes.

In one example, referring to fig. 19, in a scenario where the object receiving the warning information is a doctor, the doctor may be prompted to inform the doctor of the current risk level of the target object through any one of visual, sound, vibration, and the like. Under the condition of low risk level, corresponding warning information can not be displayed so as to reduce the interference to doctors; in the case of intermediate risk levels, the alert information may be used to prompt the physician to focus on the patient's status; under the condition of high risk level, the warning information can be used for prompting the current risk of the doctor, and the doctor is advised to pause the operation so as to enable the patient to have a rest.

In one example, referring to fig. 20, in the case that the object receiving the warning information is a planting navigation system or an auxiliary mechanical arm system, the warning information may be fed back to the planting navigation system or the auxiliary mechanical arm system, and the planting navigation system or the auxiliary mechanical arm system presents the corresponding warning information to the user. Under the condition of low risk level, the planting navigation system or the auxiliary mechanical arm system can not display the displayed information so as to avoid information interference; under the condition of intermediate risk level, the planting navigation system or the auxiliary mechanical arm system can display warning information in the interactive interface to prompt a doctor to pay attention to the state of the patient; under the condition of high risk level, the planting navigation system or the auxiliary mechanical arm system can display warning information in the interactive interface, a doctor is suggested to pause operation through the warning information, and if the doctor does not take over the processing operation aiming at the warning information within the threshold duration, the planting navigation system or the auxiliary mechanical arm system can close the surgical instrument (such as a planting mobile phone) after prompting the doctor.

Based on the state prediction method provided by the embodiment of the application, the corresponding warning information can be displayed according to the risk level of the target object, so that the risk can be found in time according to the warning information, the safety of the target object is ensured, and the comfort level of the target object in the operation process is improved.

In one embodiment, referring to fig. 21, both the state prediction network and the risk prediction model may be continuously updated iteratively according to actual data during the application process. In the operation, the electromyographic signals of the patient are collected to monitor the fatigue state, and the occurrence time of the mild fatigue state and the severe fatigue state judged by the state prediction network, the actual experience of the patient and the observation of a doctor can have deviation. The state prediction network can record the actual occurrence time of the mild fatigue state and the severe fatigue state of the patient, update the original training set based on the recorded data when the recorded data amount reaches the preset proportion of the data in the original training set (the preset proportion is a preset value, such as 50 percent, 60 percent and the like), and iteratively update the state prediction network and the risk prediction model according to the updated training set to obtain the updated state prediction network and the updated risk prediction model.

The state prediction method provided by the embodiment of the application can be used for establishing a state prediction network and a risk prediction model by comprehensively considering factors such as age, gender and medical history based on the electromyographic signals collected in the simulation scene and the processing and analyzing results thereof, further monitoring the fatigue degree of a patient in an operation through the electromyographic signals of the patient, timely finding the fatigue symptom of the patient and predicting the possible risk, so that the patient can be prevented from being in the bud, the damage caused by the possible damage caused by muscle fatigue or the damage caused by the unexpected action caused by the muscle fatigue is avoided, and the flow of the whole operation is safer. Furthermore, the risk estimation result is fused with a conventional operation process, navigation or a mechanical arm auxiliary system, and the physical condition of the patient is fed back to the doctor in real time in the operation, so that the doctor can timely respond, and the comfort level of the patient in the process of the implant operation is improved. Meanwhile, the system can take data collected in actual operation as training data, and the applicability of the system can be improved.

It should be understood that, although the steps in the flowcharts related to the embodiments as described above are sequentially displayed as indicated by arrows, the steps are not necessarily performed sequentially as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a part of the steps in the flowcharts related to the embodiments described above may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the execution order of the steps or stages is not necessarily sequential, but may be rotated or alternated with other steps or at least a part of the steps or stages in other steps.

Based on the same inventive concept, the embodiment of the present application further provides a state prediction apparatus for implementing the above-mentioned state prediction method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so specific limitations in one or more embodiments of the state prediction device provided below can be referred to the limitations of the state prediction method in the foregoing, and details are not described here.

In one embodiment, as shown in fig. 22, there is provided a state prediction apparatus including: a first acquisition module 2202, a first determination module 2204, and a prediction module 2206, wherein:

a first collecting module 2202, configured to collect a corresponding myoelectric signal for a target muscle of a target object;

a first determining module 2204, configured to determine an electromyographic characteristic index corresponding to the target muscle at the current time according to an electromyographic signal of the target muscle within a preset time period, where the electromyographic characteristic index is used to represent a change condition of the electromyographic signal corresponding to the target muscle;

a predicting module 2206, configured to perform prediction processing on the myoelectric characteristic indicator corresponding to the target muscle through a state prediction network, so as to obtain a state prediction result corresponding to the target object at the current time, where the state prediction result is used to represent the fatigue degree of the target object.

The state prediction device acquires corresponding electromyographic signals aiming at target muscles of a target object, and determines corresponding electromyographic characteristic indexes of the target muscles at the current moment according to the electromyographic signals of the target muscles within a preset time length, wherein the electromyographic characteristic indexes are used for representing the change conditions of the electromyographic signals corresponding to the target muscles. Furthermore, the electromyographic characteristic indexes corresponding to the target muscles can be subjected to prediction processing through a state prediction network, so that a state prediction result corresponding to the target object at the current moment is obtained, and the state prediction result is used for representing the fatigue degree of the target object. Based on the state prediction device provided by the embodiment of the application, the myoelectric characteristic index of the target muscle can be determined through the collected myoelectric signal of the target muscle, and then the myoelectric characteristic index is subjected to prediction processing through the state prediction network to obtain the state prediction result, the state prediction result is not influenced by subjective consciousness of people, and the state prediction device is more objective and can improve the accuracy of state prediction.

In one embodiment, the prediction module 2206 is further configured to:

acquiring personal attribute information of the target object;

In one embodiment, the first determining module 2204 is further configured to:

In one embodiment, the state prediction network includes a first network and a second network, the first network is configured to identify a normal state and a fatigue state, the second network is configured to identify a fatigue level corresponding to the fatigue state, and the prediction module 2206 is further configured to:

taking the fatigue grade corresponding to the fatigue state represented by the second state prediction result as the state prediction result corresponding to the target object at the current moment;

or,

In one embodiment, the apparatus further comprises:

In one embodiment, the second determining module is further configured to:

In one embodiment, the apparatus further comprises:

In one embodiment, the building module is further configured to:

In one embodiment, the apparatus further comprises:

The respective modules in the state prediction apparatus described above may be implemented in whole or in part by software, hardware, and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.

In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in fig. 23. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement a state prediction method. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.

It will be appreciated by those skilled in the art that the configuration shown in fig. 23 is a block diagram of only a portion of the configuration associated with the present application, and is not intended to limit the computing device to which the present application may be applied, and that a particular computing device may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.

In one embodiment, a computer device is further provided, which includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the above method embodiments when executing the computer program.

In an embodiment, a computer-readable storage medium is provided, on which a computer program is stored which, when being executed by a processor, carries out the steps of the above-mentioned method embodiments.

In an embodiment, a computer program product is provided, comprising a computer program which, when being executed by a processor, carries out the steps of the above-mentioned method embodiments.

It should be noted that, the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, presented data, etc.) referred to in the present application are information and data authorized by the user or sufficiently authorized by each party.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, database, or other medium used in the embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high-density embedded nonvolatile Memory, resistive Random Access Memory (ReRAM), Magnetic Random Access Memory (MRAM), Ferroelectric Random Access Memory (FRAM), Phase Change Memory (PCM), graphene Memory, and the like. Volatile Memory can include Random Access Memory (RAM), external cache Memory, and the like. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others. The databases referred to in various embodiments provided herein may include at least one of relational and non-relational databases. The non-relational database may include, but is not limited to, a block chain based distributed database, and the like. The processors referred to in the embodiments provided herein may be general purpose processors, central processing units, graphics processors, digital signal processors, programmable logic devices, quantum computing based data processing logic devices, etc., without limitation.

The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.

The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present application. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A method of predicting a state, the method comprising:

2. The method according to claim 1, wherein the predicting myoelectric characteristic indicators corresponding to the target muscles through a state prediction network to obtain a state prediction result corresponding to the target object at the current time includes:

acquiring personal attribute information of the target object;

3. The method according to claim 1, wherein the determining the electromyographic characteristic index of the target muscle corresponding to the current time according to the electromyographic signal of the target muscle within a preset time duration comprises:

4. The method according to any one of claims 1 to 3, wherein the state prediction network comprises a first network for identifying a normal state and a fatigue state and a second network for identifying a fatigue level corresponding to the fatigue state,

or, when the first state prediction result represents that the current state is the normal state, taking the normal state as a state prediction result corresponding to the target object at the current moment.

5. The method according to any one of claims 1 to 4, further comprising:

6. The method of claim 5, wherein determining the risk level of the target object at the current time based on the state prediction comprises:

7. The method according to any one of claims 1 to 3, further comprising:

collecting a first sample myoelectric signal corresponding to the target muscle of a simulation object in a resting state;

8. The method according to claim 7, wherein the constructing a sample group according to the corresponding status information of the first sample electromyographic signal, the second sample electromyographic signal and the second sample electromyographic signal at different time instants comprises:

9. The method of claim 7 or 8, wherein the state prediction network comprises a first network and a second network, and wherein training the state prediction network through the training set comprises:

10. The method of claim 5 or 6, further comprising:

11. A state prediction apparatus, characterized in that the apparatus comprises:

the device comprises a first determination module, a second determination module and a third determination module, wherein the first determination module is used for determining an electromyographic characteristic index corresponding to a target muscle at the current moment according to the electromyographic signal of the target muscle within a preset time after acquiring the electromyographic signal of the preset time, and the electromyographic characteristic index is used for representing the change condition of the electromyographic signal corresponding to the target muscle;

12. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor realizes the steps of the method of any one of claims 1 to 10 when executing the computer program.

13. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 10.