Operation method and device of control system based on myoelectricity

Abstract

The invention discloses an operation method of a control system based on myoelectricity. The operation method comprises the following steps: acquiring a myoelectricity signal generated by a gesture action of a user; a certain window and recognition frequency are set on the time dimension, the electromyographic signals are recognized according to the set window and recognition frequency, and a recognition result is obtained; and performing weighted voting on the identification result to obtain an operation instruction. The invention provides a visual and natural interaction mode, so that a user can operate a virtual interface more easily; through various gesture operations, richer interaction functions are provided, and the user experience and efficiency are improved. The second aspect of the invention provides a device applying the operation method of the myoelectricity-based control system.

Description

Technical Field

[0001] The present invention belongs to the field of human-computer interaction control realized through bioelectricity, and in particular, relates to an operation method and device of a control system based on myoelectricity. Background Art

[0002] At present, the main control scheme for AR / VR is to control the system through eye tracking combined with gesture recognition. Eye tracking technology is used to detect the user's line of sight and gaze point. By using a dedicated eye tracking device or camera, the system can accurately track the movement and focus position of the user's eyes. By identifying the user's line of sight, the system can determine the user's point of interest in the virtual scene and adjust the display of virtual content accordingly. The gesture recognition technology in the system is mainly based on optical sensors such as cameras, depth sensors, and ToF to capture user hand movements and convert them into instructions. Users can use both hands to perform various gestures, such as pinching, swiping, rotating, etc., to operate and control AR / VR content. Existing technical solutions often rely on multiple cameras and high-precision cameras, so the cost is high and requires a lot of computing resources.

[0003] In view of this, the present invention is proposed. Summary of the invention

[0004] The technical problem to be solved by the present invention is to overcome the shortcomings of the prior art and provide an operating method for a control system based on electromyography, which collects electromyographic signals generated by user gestures and then converts them into corresponding operating instructions to achieve control of objects with low cost and simple device.

[0005] In order to solve the above technical problems, the basic concept of the technical solution adopted by the present invention is:

[0006] A first aspect of the present invention provides an operation method of a control system based on myoelectricity, comprising:

[0007] S1, obtaining the electromyographic signal generated by the user's gesture action;

[0008] S2. Set a certain window and recognition frequency in the time dimension, recognize the electromyographic signal according to the set window and recognition frequency, and obtain the recognition result;

[0009] S3. Perform weighted voting on the recognition results to obtain operation instructions.

[0010] Optionally, S2 also includes:

[0011] A certain window and recognition frequency are set in the time dimension, and the signals in at least three windows are extracted and recognized to obtain the recognition results.

[0012] Optionally, S2 also includes:

[0013] S21, setting a first window and a recognition frequency in a time dimension;

[0014] S22, setting a second window and a third window of equal length to the first window in the time dimension, and the first window, the second window and the third window partially overlap in the time dimension;

[0015] S23, extracting and identifying the electromyographic signals of the first window, the second window and the third window respectively, and obtaining identification results.

[0016] Optionally, the operation instructions include at least: double-clicking, pinching, dragging and rotating, and each operation instruction includes different sub-actions.

[0017] Optionally, the corresponding operation instruction is inferred based on at least three operation instructions identified in previous detections and the sub-actions of the operation instruction acquired this time.

[0018] Optionally, the operation instruction corresponding to the gesture action is further determined based on the detected operation instruction and the duration of the sub-action of the operation instruction.

[0019] Optional, before S2, including:

[0020] S11, matching and calculating the acquired electromyographic signal with the existing basic model to obtain parameters corresponding to the electromyographic signal;

[0021] S12. Match the obtained parameters with the user information to obtain the parameters unique to the user.

[0022] Optional, before S1, including:

[0023] S1', obtaining the electromyographic signal of the action performed by the user;

[0024] S2', matching and calculating the acquired electromyographic signal with the existing basic unlocking model to obtain characteristic information of the electromyographic signal;

[0025] S3', matching the obtained feature information with the existing user unlocking feature information, if the difference between the two is within a certain threshold, unlocking is performed.

[0026] Optionally, in S2', the basic unlocking model is a model trained by a large amount of user data and used to extract feature information of low-precision electromyographic signals.

[0027] A second aspect of the present invention provides a device for a myoelectric control system, characterized in that the device adopts an operation method of a myoelectric control system according to any one of claims 1 to 9, comprising:

[0028] A signal acquisition module is used to obtain the electromyographic signals generated when the user makes movements;

[0029] A signal processing module, used for converting the acquired electromyographic signals into operation instructions;

[0030] The data transmission module is used to transmit the operation instructions obtained after processing by the processing module to the host computer; wherein the host computer includes: an electromyographic bracelet and an AR / VR device.

[0031] After adopting the above technical scheme, the present invention has the following beneficial effects compared with the prior art.

[0032] 1. The present invention provides an intuitive and natural interaction method, allowing users to operate the virtual interface more easily; through a variety of gesture operations, it provides richer interactive functions and enhances the user experience.

[0033] 2. The present invention analyzes the user's gestures through electromyographic signals and converts them into corresponding action information, thereby reducing the use of equipment and simplifying the analysis process.

[0034] 3. The present invention divides the time when the electromyographic signal is generated into multiple windows, extracts the action information in each window, and then obtains the operation instruction corresponding to the electromyographic signal through weighted voting, thereby increasing the accuracy of recognition.

[0035] 4. The present invention further subdivides the divided windows into windows with shorter time intervals, extracts and identifies the subdivided windows, obtains a recognition result of a window, and then weights the recognition results of all windows with shorter time intervals to obtain the operation instructions corresponding to the windows with longer time intervals, and further subdivides the identification, thereby improving the recognition accuracy and ensuring that the recognized operation instructions are more closely matched with the user's gesture actions.

[0036] 5. The present invention matches the finally recognized action information with the user's unique identity information as the user's characteristic information, thereby increasing the accuracy of use by a single user and making the use process more convenient for the user.

[0037] The specific implementation modes of the present invention are further described in detail below in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The accompanying drawings are part of the present invention and are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but do not constitute an improper limitation of the present invention. Obviously, the drawings described below are only some embodiments. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the accompanying drawings:

[0039] Figure 1 It is a schematic flow chart of the myoelectric signal recognition process in the operation method of the myoelectric control system of the present invention;

[0040] Figure 2 The process flow diagram of the myoelectric signal recognition process in the operation method of the myoelectric control system of the present invention is as follows Figure 2 ;

[0041] Figure 3 The process flow diagram of the myoelectric signal recognition process in the operation method of the myoelectric control system of the present invention is as follows Figure 3 ;

[0042] Figure 4 It is a flow chart of user customized parameters in the operation method of the myoelectric control system of the present invention;

[0043] Figure 5 It is a schematic flow chart of the unlocking process in the operation method of the control system based on electromyography of the present invention;

[0044] Figure 6 It is a schematic diagram of the electromyographic signal recognition process of the present invention;

[0045] Figure 7 It is a schematic diagram of device signal transmission of the myoelectricity-based control system of the present invention.

[0046] It should be noted that these drawings and textual descriptions are not intended to limit the conceptual scope of the present invention in any way, but are intended to illustrate the concept of the present invention for those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The following embodiments are used to illustrate the present invention but are not used to limit the scope of the present invention.

[0048] In the description of the present invention, it should be noted that the directions or positional relationships indicated by terms such as “upper”, “lower”, “front”, “back”, “left”, “right”, “vertical”, “inside” and “outside” are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as a limitation on the present invention.

[0049] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0050] The present invention provides an operation method of a control system based on myoelectricity, such as Figures 1 to 3 As shown, it includes steps S1 to S3.

[0051] In step S1, the electromyographic signal generated by the user's gesture action is obtained.

[0052] In step S2, a certain window and recognition frequency are set in the time dimension, and the electromyographic signal is recognized according to the set window and recognition frequency to obtain a recognition result.

[0053] In step S3, weighted voting is performed on the identified results to obtain a final operation instruction.

[0054] Optionally, in step S2, it includes:

[0055] A certain window and recognition frequency are set in the time dimension, and the signals in at least three windows are extracted and recognized to obtain the recognition results.

[0056] Optionally, step S2 further includes:

[0057] In step S21, a first window and a recognition frequency are set in the time dimension;

[0058] In step S22, a second window and a third window are set to have the same length as the first window in terms of time dimension, and the first window, the second window and the third window partially overlap in terms of time dimension;

[0059] In step S23, the electromyographic signals of the first window, the second window and the third window are extracted and identified respectively to obtain identification results.

[0060] Specifically, Figure 6 As shown, the acquired electromyographic signal is segmented in the time dimension, a certain window is set, a single window is identified, and an identification result is obtained. Then, a weighted vote is performed on the identification result to obtain an operation instruction.

[0061] In this embodiment, a certain window and recognition frequency are set for the acquired electromyographic signal in a practical dimension, the signals in at least three windows are recognized, the recognition results are obtained, and then weighted voting is performed to obtain the operation instructions.

[0062] More specifically, the electromyographic signal is set to a first window, a second window, and a third window in the time dimension, and the time of the first window, the second window, and the third window is equal in length, and the first window, the second window, and the third window partially overlap in the time dimension. The signal in the first window is identified to obtain a recognition result, and then the signal in the second window is identified to obtain a recognition result, and then the signal in the third window is identified to obtain a recognition result, and the obtained recognition result is weighted voted to obtain an operation instruction. For example, using 300ms of data as the length of the identified electromyographic signal, the data within 300ms can be divided into multiple 100-200ms sub-data, and each 100-200ms of data is extracted and identified to obtain a recognition result, and then the obtained recognition result is weighted voted to obtain the operation instruction corresponding to the electromyographic signal.

[0063] Optionally, the operation instructions include at least: double-clicking, pinching, dragging and rotating, and each operation instruction includes different sub-actions.

[0064] Specifically, double-clicking, that is, the thumb touches the index finger twice. The user performs a double-click operation on the virtual element by touching the index finger twice to achieve the operation of zooming in or opening.

[0065] Pinch, that is, pinching the thumb and index finger together for a period of time, the user performs a highlighting operation with the virtual element by pinching the thumb and index finger together and holding it for a period of time. This operation can be used to highlight text or other elements.

[0066] Drag, that is, pinch and drag with the thumb and index finger. Users can scroll and move the window by pinching the thumb and index finger and dragging them. Users can scroll or move the window at different angles, and the system will make corresponding adjustments based on the gesture speed.

[0067] Rotation, that is, pinching the index finger and thumb of the right hand at the same time and rotating clockwise, the user manipulates the direction of the virtual object by pinching the index finger and thumb of the right hand and rotating clockwise. This operation can be used to rotate virtual elements.

[0068] The operation instructions shown may also include: click and zoom.

[0069] Specifically, a single click is when the thumb touches the index finger once. The user performs a click operation on the virtual element by touching the index finger once, thereby realizing the selection of the virtual element.

[0070] Zooming means pinching the index fingers and thumbs of both hands together and pulling them apart to the left and right. Users can zoom in on the window by simultaneously moving the index fingers and thumbs of both hands together, pinching them together and pulling them apart to the left and right. This operation can be used to expand the size of the virtual window.

[0071] For example, taking the index finger and thumb action as an example, the process and action of clicking includes three parts: pressing, touching, and lifting. Finger movements are divided into clicking and pinching. Clicking is the rapid contact between the index finger and the thumb, which produces a more obvious hand vibration. Its force feedback information will be reflected in the neural electrical signal, and its tiny vibration information will be reflected in the vibration signal. The click action can be identified by using the force feedback signal and vibration signal of pressing combined with deep learning methods.

[0072] When pinching, the index finger and thumb come into contact and exert a certain force. The contact time is longer than clicking, which results in no similar vibration information, but only force feedback information. The neural electrical signal of force feedback may also be different from the pressure during clicking. Therefore, deep learning of the neural electrical signal of the pinching action can also identify the information of the pinching process.

[0073] It should be noted that, in addition to the single click, double click, pinch, drag, zoom and rotate mentioned above, the operation instructions supported by this myoelectric system may also include operation instructions generated by other gesture actions.

[0074] Optionally, the corresponding operation instruction is inferred based on at least three operation instructions identified in previous detections and the sub-actions of the operation instruction acquired this time.

[0075] Specifically, the previous recognition results are used as the basis for judgment to judge the intention. For example, the action results obtained by calculating the previous 2 to 5 data are weighted. If the recognition rate of click or pinch action increases, or the "press" action is recognized, the probability of identifying the click (contact) increases; in addition, since electromyography represents the muscle force, which is an indicator of action acceleration, this part of the previous action data can also be directly input into the model as the model input (in this case, the previous action results need to be added in the training stage for multimodal training), so as to improve the recognition rate in the calculation stage. This processing method can speed up the signal processing process and can identify the operation instructions represented by the gesture action in a short time. Optionally, the operation instruction corresponding to the gesture action is further determined based on the duration of the detected operation instruction and the operation instruction sub-action.

[0076] Specifically, in order to further improve the certainty of the recognition results, the recognition results cannot be directly used as the final recognized action, but need to be combined and judged to obtain gesture judgment. For example, among the six gestures, click is "click" or "pinch" for a short time (such as <200ms); double-click is "click" twice within a certain time window; pinch is "pinch" for a long time (such as >200ms); drag is "pinch" + accelerometer movement detection; zoom is dragging of both hands; rotation is "pinch" + gyroscope rotation detection.

[0077] Before step S2, Figure 4 As shown, it includes steps S1 to S12.

[0078] In step S11, the acquired electromyographic signal is matched with the existing basic model to obtain parameters corresponding to the electromyographic signal.

[0079] In step S12, the obtained parameters are matched with the user information as parameters unique to the user.

[0080] Specifically, due to the particularity of electromyography, each person's muscle state, muscle control level and electromyography filtering by the fat cortex are likely to be quite different, so it is necessary to collect and calculate the user's electromyography characteristics in advance, such as MVC, baseline, mean and other parameters.

[0081] Specifically, the user is asked to click and pinch multiple times to collect a small amount of his electromyographic signal, which is then recognized using an existing recognition model. Different parameters are automatically calculated and matched to achieve the highest recognition accuracy, thereby obtaining customized parameters unique to the user.

[0082] In this embodiment, the existing recognition model is trained and stored before obtaining the electromyographic signal of the user's gesture action. By recognizing the electromyographic signal of the gesture action made by the user multiple times, and matching it with the existing recognition model, the matching is automatically calculated to switch the parameters, the recognition accuracy is increased, and the corresponding feature information is obtained.

[0083] Before step S1, Figure 5 As shown, it includes steps S1'~S3'.

[0084] Specifically, the control system based on electromyography belongs to a human-computer interaction system, so it is necessary to know whether the user has the intention to operate, so as to prevent the user from being identified as operating when there is no intention to operate, resulting in the occurrence of erroneous operation.

[0085] In step S1', the electromyographic signal of the action performed by the user is obtained

[0086] Specifically, the user makes corresponding actions and obtains the electromyographic signals generated by the actions made by the user. It should be noted that since each user's muscle state, muscle control level and fat cortex are different, when the user makes an action, a corresponding electromyographic signal will be generated. Moreover, different users make different electromyographic signals when making the same action. Therefore, the acquired electromyographic signals are specific to each user. It should be noted that the action made by the user can be a gesture action made by the system to remind the user, or it can be an unlocking action set by the user according to his own situation.

[0087] In step S2', the acquired electromyographic signal is matched with the existing basic unlocking model to obtain feature information of the electromyographic signal.

[0088] In this embodiment, the existing basic unlocking model is a model trained by a large amount of user data and is used to extract feature information of low-precision electromyographic signals.

[0089] In step S3', the obtained feature information is matched with the existing user unlocking feature information, and if the difference between the two is within a certain threshold, unlocking is performed.

[0090] Specifically, first, a basic unlocking model based on deep learning of low-precision signals is trained with a large amount of user data to extract the feature information of low-precision electromyographic signals; then, before using the device, the user makes a specific action once or multiple times. It should be noted that this specific action can be the user's own special action or a specific action specified by the system. The information of this action is bound as the user's unlocking information, and the electromyographic signal of this action is matched and calculated on the trained basic unlocking model to obtain the feature information of this data, and correspond it with the user's identity information and save it as the user's unlocking feature information.

[0091] Specifically, when in the locked state, information with a lower number of collected samples, or high-frequency electromyography with a smaller number of channels is used to identify the unlocking action made by the user. The system extracts the electromyography signal of the resulting action made by the user, and matches and calculates the extracted electromyography signal with the existing basic unlocking model to obtain the characteristic information of the electromyography signal. This characteristic information is then compared with the recorded characteristic information of the user. If the distance between the two is within a certain threshold, the unlocking is successful, otherwise it is unsuccessful.

[0092] It should be noted that, because everyone's muscles are different, the system cannot be unlocked even if the same action is not performed by the designated user.

[0093] In this embodiment, the calculation method can use parameters such as the maximum and minimum value intervals and the average value to adjust these parameters of the user's electromechanical signal to parameters that are more compatible with the existing recognition model. For example, the maximum value of the data with a better match to the original recognition model is max1, the minimum value is min1, and the average value is avg1; the maximum value of the user data is max2, the minimum value is min2, and the average value is avg2. Then the new user data x can be adjusted to (x-.avg2+avg1) / (max2-min2)*(max1-min1). In addition, parameters such as MVC, baseline, and RMS can also be used for complex adjustments.

[0094] In this embodiment, through the above-mentioned recognition and matching process, the user parameters can be updated, the user data can be recorded, and the model can be fine-tuned at regular intervals; at the same time, the existing model can be updated, that is, the model can be fitted in the direction of the user. Specifically, it is equivalent to treating the original recognition model as a pre-trained model, and constantly fine-tuning it in actual use, thereby improving the recognition accuracy of specific users without the need to retrain the model and with a relatively small amount of data.

[0095] A second aspect of the present invention provides a device for a myoelectric control system, using the above-mentioned method for operating a myoelectric control system, such as Figure 7 As shown, including:

[0096] A signal acquisition module is used to obtain the electromyographic signal generated when the user makes a gesture;

[0097] A signal processing module, used for converting the acquired electromyographic signals into operation instructions;

[0098] The data transmission module is used to transmit the operation instructions obtained after processing by the processing module to the host computer; wherein the host computer includes: an electromyographic bracelet and an AR / VR device.

[0099] In this embodiment, the signal acquisition module relies on a bracelet or armband signal acquisition device. The acquisition device is mainly composed of a neural signal acquisition part and an IMU gyroscope part. The neural signal acquisition part is mainly used to identify the myoelectric neural signals generated by hand movements such as wrist rotation, palm expansion and contraction, and finger changes; the IMU gyroscope part is mainly used to identify information on forearm rotation and forearm position changes; the data processing part is mainly used to process the collected data, including but not limited to simple data processing such as filtering and rectification, and complex calculation and recognition processes such as deploying artificial intelligence models; the data transmission part is mainly used to transmit the processed data to the host computer.

[0100] In this embodiment, six gestures are described above, and each action can be divided into finger movement and wrist movement. Finger movement is divided into clicking and pinching. Taking the index finger and thumb movements as an example, clicking is the rapid contact between the index finger and the thumb, which produces a more obvious hand vibration. Its force feedback information will be reflected in the neural electrical signal, and its tiny vibration information will be reflected in the IMU gyroscope signal. The click action can be identified by combining the force feedback signal and vibration signal of the press with the deep learning method. When pinching, the process in which the index finger contacts the thumb and generates a certain force has a longer contact time than clicking, which will result in no similar vibration information, but only force feedback information, and the neural electrical signal of the force feedback may also be different from the pressing during clicking. Therefore, the information of the kneading process can also be identified by deep learning of the neural electrical signal of the kneading action.

[0101] It is worth noting that there are many possibilities of misidentification when using neural electrical signals and gyroscope signals for gesture recognition. For example, the process and action of clicking includes three action parts: pressing, touching, and lifting. The "touching" part is what needs to be recognized, while the pressing and lifting signals do not need to be recognized. However, the signals of this part may be more obvious than the "touching" part, so errors are prone to occur during data segmentation and recognition. Therefore, in addition to using a large amount of negative sample data for training, capacitor and resistor sensors are also used to assist in accurately marking the "touching" part. That is, capacitor and resistor sensors are worn on the hands, and corresponding information will appear when touching. By synchronizing the information with the electromyography and gyroscope information, the "touching" part can be accurately marked.

[0102] In this embodiment, in addition to finger movements, wrist movements are also included. Wrist movements are mainly used to identify the dynamics of hand position and direction through the IMU gyroscope. It is worth noting that the IMU and gyroscope cannot completely and accurately identify the spatial position, but can only identify the dynamics, and the dragging, scaling and rotation we need do not require complete positioning, but only need to obtain fuzzy dynamics.

[0103] The specific usage process includes unlocking the device, setting user parameters, and identifying gestures. In the locked state, you can use information with a lower number of collected samples, such as IMU gyroscopes and low-frequency electromyography to identify specific actions that are different between each person, or use high-frequency electromyography with fewer channels to identify the differences between specific actions between different people, to unlock, thereby reducing the power consumption of the collection while taking into account the unlocking function based on biometric information. During the parameter setting process, the user can click and pinch multiple times to collect a small amount of electromyography and gyroscope information samples, and use the existing recognition model for recognition, automatically calculate and match and switch different parameters, so that the recognition accuracy is the highest, thereby obtaining customized parameters unique to the user. During the recognition process, the entire signal can be recognized, or inferences can be made based on the previous data and the sub-actions of this recognition.

[0104] The above description is only a preferred embodiment of the present invention, and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technician familiar with this patent can make some changes or modifications to equivalent embodiments of equivalent changes by using the technical content suggested above without departing from the scope of the technical solution of the present invention. The implementation scheme in the above embodiment can also be further combined or replaced. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the scope of the solution of the present invention.

Claims

1. An operating method of a control system based on electromyography, characterized in that: S1, obtaining the electromyographic signal generated by the user's gesture action; S2. Set a certain window and recognition frequency in the time dimension, recognize the electromyographic signal according to the set window and recognition frequency, and obtain the recognition result; S3. Perform weighted voting on the recognition results to obtain operation instructions.

2. The method for operating a control system based on myoelectricity according to claim 1, characterized in that: In S2, it includes: A certain window and recognition frequency are set in the time dimension, and the signals in at least three windows are extracted and recognized to obtain the recognition results.

3. The method for operating a control system based on myoelectricity according to claim 2, characterized in that: In S2, it also includes: S21, setting a first window and a recognition frequency in a time dimension; S22, setting a second window and a third window of equal length to the first window in the time dimension, and the first window, the second window and the third window partially overlap in the time dimension; S23, extracting and identifying the electromyographic signals of the first window, the second window and the third window respectively, and obtaining identification results.

4. The method for operating a control system based on myoelectricity according to claim 1, characterized in that: The operation instructions at least include: double-clicking, pinching, dragging and rotating, and each of the operation instructions includes different sub-actions.

5. The operation method based on myoelectric control system according to any one of claims 1 to 4, characterized in that: The corresponding operation instruction is inferred based on at least three operation instructions recognized in the previous detection and the sub-actions of the operation instruction obtained this time.

6. The method for operating a control system based on myoelectricity according to any one of claims 1 to 4, characterized in that: The operation instruction corresponding to the gesture action is further determined according to the detected operation instruction and the duration of the sub-action of the operation instruction.

7. The method for operating a control system based on myoelectricity according to claim 1, characterized in that: Prior to S2, this included: S11, matching and calculating the acquired electromyographic signal with the existing basic model to obtain parameters corresponding to the electromyographic signal; S12. Match the obtained parameters with the user information to obtain the parameters unique to the user.

8. The method for operating a control system based on myoelectricity according to claim 1, characterized in that: Prior to S1, this included: S1', obtaining the electromyographic signal of the action performed by the user; S2', matching and calculating the acquired electromyographic signal with the existing basic unlocking model to obtain characteristic information of the electromyographic signal; S3', matching the obtained feature information with the existing user unlocking feature information, if the difference between the two is within a certain threshold, unlocking is performed.

9. The method for operating a control system based on myoelectricity according to claim 8, characterized in that: In S2', the basic unlocking model is trained with a large amount of user data and is used to extract feature information of low-precision electromyographic signals.

10. A device for a control system based on electromyography, characterized in that: The method for operating the myoelectric control system according to any one of claims 1 to 9 comprises: A signal acquisition module is used to obtain the electromyographic signals generated when the user makes movements; A signal processing module, used for converting the acquired electromyographic signals into operation instructions; The data transmission module is used to transmit the operation instructions obtained after processing by the processing module to the host computer; wherein the host computer includes: an electromyographic bracelet and an AR / VR device.

Images