CN113940680A

CN113940680A - Motion control method and system based on electromyographic signals

Info

Publication number: CN113940680A
Application number: CN202111276798.9A
Authority: CN
Inventors: 楚喆
Original assignee: Individual
Current assignee: Individual
Priority date: 2021-10-29
Filing date: 2021-10-29
Publication date: 2022-01-18

Abstract

The invention discloses a motion control method and a system thereof based on electromyographic signals, comprising the following steps: collecting myoelectric signals of a human body; conditioning and filtering the electromyographic signals to generate filtering signals; obtaining a pre-trained action matching model, and inputting the filtering signal into the action matching model to generate an action prediction result; and outputting a motion control instruction to a motion execution unit based on the motion prediction result. According to the invention, a plurality of groups of movement electromyographic signals are collected in a mode of surrounding the same target muscle group, and three-dimensional electromyographic distribution characteristics with three parameters of collection time, electromyographic signal values and pasting electrode numbers can be generated, so that the action type of a user can be accurately represented, the accuracy of a movement control instruction output by a control unit is effectively improved under the condition of not generating extra computational load, and the problems of low movement control reliability and low sensitivity of the traditional movement control method based on the electromyographic signals are solved.

Description

Motion control method and system based on electromyographic signals

Technical Field

The invention relates to the technical field of electromyographic signal detection, in particular to a motion control method and a motion control system based on electromyographic signals.

Background

The Surface Electromyography (SEMG) is a result of the comprehensive superposition of time and space obtained by detecting potential sequences emitted by a plurality of motor units on the Surface layer of muscle on the Surface of skin through electrodes in the autonomous activity of a human, and is a bioelectricity signal accompanied with the activity of a neuromuscular system. Since it relates to the activity state and the functional state of muscles to a certain extent, electromyographic signals are generally regarded as the best tools for reflecting joint muscle activities, and are widely used in the fields of medical rehabilitation, exercise control and the like.

However, the conventional myoelectric signal-based motion control method often cannot avoid system errors caused by interference of various signals, such as: the real exertion of the user cannot be accurately predicted based on the collected electromyographic signals, and then an erroneous motion control instruction is output. The existing methods for solving the above defects are generally: the variety of the collected electromyographic signals is increased, a rear-end filtering algorithm (such as a wavelet algorithm) is added, the improvement effect is limited, the rear-end computational load is further improved, and the system cannot achieve ideal response speed easily. Also, conventional control methods typically require detecting arm movement, thereby requiring the user to move the arm significantly.

In summary, the existing motion control method based on the electromyographic signals has the problems of low reliability and low sensitivity of motion control.

Disclosure of Invention

In view of the above, the present invention provides a motion control method and system based on an electromyographic signal, which solves the problems of low reliability and low sensitivity of motion control in the conventional motion control method based on an electromyographic signal by improving an electromyographic signal acquisition method and a data processing method.

In order to solve the above problems, the present invention provides a method for controlling a motion based on an electromyographic signal, comprising: collecting myoelectric signals of a human body; conditioning and filtering the electromyographic signals to generate filtering signals; obtaining a pre-trained action matching model, and inputting the filtering signal into the action matching model to generate an action prediction result; and outputting a motion control instruction to a motion execution unit based on the motion prediction result.

Optionally, acquiring an electromyographic signal of a body includes: collecting a reference electromyographic signal at the joint of the shoulder and the upper arm of the human body within unit time t; and collecting a plurality of groups of movement electromyographic signals of the upper arm of the body in unit time t.

Optionally, conditioning and filtering the electromyographic signal to generate a filtered signal, includes: respectively subtracting the reference electromyographic signal values of the same acquisition time stamp from the multiple groups of the motion electromyographic signal values to generate multiple groups of first electromyographic signals; and after differential derivation and high-pass filtering are carried out on the multiple groups of first electromyographic signals, generating a filtering signal capable of representing electromyographic distribution characteristics at one week of the upper arm of the human body.

Optionally, training the motion matching model comprises: constructing an initialization network model, wherein the network model comprises a semantic matching model; acquiring a training data set and a test data set which are composed of electromyographic signal samples containing labeling action types and electromyographic distribution characteristics; training and testing the network model based on the training dataset and the testing dataset.

Optionally, the motion control method further comprises: collecting the electromyographic signals of the body, and caching the original electromyographic signals.

Accordingly, the present invention provides a motion control system based on electromyographic signals, comprising: the electromyographic signal collector is used for collecting the electromyographic signal of the human body; the control unit is used for conditioning and filtering the electromyographic signals to generate filtering signals, acquiring a pre-trained action matching model, inputting the filtering signals into the action matching model to generate action prediction results, and outputting motion control instructions to the action execution unit based on the action prediction results; and the action execution unit is used for executing the motion control instruction.

Optionally, the electromyographic signal collector comprises a plurality of application electrodes, wherein a first application electrode is arranged at the joint of the shoulder and the upper arm of the body and is used for collecting a reference electromyographic signal at the joint of the shoulder and the upper arm of the body within a unit time t; the plurality of the application electrodes except the first application electrode are equidistantly arranged on one circle of the upper arm of the body and are used for collecting a plurality of groups of movement electromyographic signals on one circle of the upper arm of the body in unit time t.

Optionally, the control unit is configured to: respectively subtracting the reference electromyographic signal values of the same acquisition time stamp from the multiple groups of the motion electromyographic signal values to generate multiple groups of first electromyographic signals; and after differential derivation and high-pass filtering are carried out on the multiple groups of first electromyographic signals, generating a filtering signal capable of representing electromyographic distribution characteristics at one week of the upper arm of the human body.

Optionally, the motion control system further includes a buffer module, configured to store the electromyographic signal.

Optionally, the action execution unit is an electronic control unit for controlling the mechanical action unit.

The invention has the primary improvement that the provided motion control method based on the electromyographic signals collects a plurality of groups of motion electromyographic signals according to the mode of surrounding the same target muscle group, and collecting reference electromyographic signals for removing electromyographic signal interference of the target external muscle group, so that after the reference electromyographic signal values of the same collection time stamp, differential derivation of each group of the motion electromyographic signal values and high-pass filtering are respectively subtracted from a plurality of groups of the motion electromyographic signal values, can generate three-dimensional myoelectric distribution characteristics with three parameters of acquisition time, myoelectric signal value and application electrode number, thereby accurately representing the action type of a user, under the condition of not generating extra computational load, the accuracy of the motion control instruction output by the control unit is effectively improved, and the problems of low motion control reliability and low sensitivity of the traditional motion control method based on the electromyographic signal are solved.

Drawings

FIG. 1 is a simplified flow diagram of a myoelectric signal-based motion control method of the present invention;

FIG. 2 is an exemplary diagram of an arrangement of an electromyographic signal collector according to the present invention;

FIG. 3 is a simplified unit connection diagram of the myoelectric signal-based motion control system of the present invention.

Detailed Description

In order to make the technical solutions of the present invention better understood by those skilled in the art, the present invention will be further described in detail with reference to the accompanying drawings and specific embodiments.

As shown in fig. 1, a method for controlling a motion based on an electromyographic signal includes: collecting myoelectric signals of a human body; conditioning and filtering the electromyographic signals to generate filtering signals; obtaining a pre-trained action matching model, and inputting the filtering signal into the action matching model to generate an action prediction result; and outputting a motion control instruction to a motion execution unit based on the motion prediction result. Wherein the motion control method further comprises: collecting the electromyographic signals of the body, and caching the original electromyographic signals.

Further, the method for acquiring the electromyographic signals of the human body comprises the following steps: collecting a reference electromyographic signal at the joint of the shoulder and the upper arm of the human body within unit time t; and collecting a plurality of groups of movement electromyographic signals of the upper arm of the body in unit time t.

Further, conditioning and filtering the electromyographic signal to generate a filtering signal includes: respectively subtracting the reference electromyographic signal values of the same acquisition time stamp from the multiple groups of the motion electromyographic signal values to generate multiple groups of first electromyographic signals; and after differential derivation and high-pass filtering are carried out on the multiple groups of first electromyographic signals, generating a filtering signal capable of representing electromyographic distribution characteristics at one week of the upper arm of the human body.

Further, training the motion matching model includes: constructing an initialization network model, wherein the network model comprises a semantic matching model; acquiring a training data set and a test data set which are composed of electromyographic signal samples containing labeling action types and electromyographic distribution characteristics; training and testing the network model based on the training dataset and the testing dataset. It should be noted that the semantic matching model used in the present application is conventional in the art, and does not involve further improvement of the model architecture, and therefore, the type and architecture of the semantic matching model are not specifically limited.

In order to facilitate understanding of the working principle of the present application, an embodiment in which an electronic control unit for acquiring four groups of motion electromyographic signals and controlling a mechanical exoskeleton is used as a motion execution unit, and a single-chip microcomputer MCU is used as a control unit is now described:

as shown in fig. 2, the electromyographic signal collector is provided with five application electrodes, the first application electrode 1 is arranged at the joint of the shoulder and the upper arm of the body, and the second application electrode 2, the third application electrode 3, the fourth application electrode 4 and the fifth application electrode 5 are arranged on the periphery of the upper arm of the body in an equidistant surrounding manner. The acquisition frequencies of the first application electrode 1, the second application electrode 2, the third application electrode 3, the fourth application electrode 4 and the fifth application electrode 5 are the same, the acquisition frequencies are not particularly limited, the acquisition frequencies meet the requirement that multiple acquisition can be carried out in a single acquisition cycle t, and the three-dimensional myoelectric distribution characteristics with the acquisition time, the myoelectric signal value and the application electrode number as three parameters can be generated.

After the control unit sends a collection instruction, in a single collection period t, the first application electrode 1 collects a reference electromyographic signal at the joint of the shoulder and the upper arm of the body in unit time t, the second application electrode 2 collects a first movement electromyographic signal at the upper arm of the body in unit time t, the third application electrode 3 collects a second movement electromyographic signal at the upper arm of the body in unit time t, the fourth application electrode 4 collects a third movement electromyographic signal at the upper arm of the body in unit time t, and the fifth application electrode collects a fourth movement electromyographic signal at the upper arm of the body in unit time t.

The control unit generates a first corrected electromyographic signal by subtracting the reference electromyographic signal value of the same acquisition time stamp from a plurality of first electromyographic signal values acquired within a unit time t, generates a second corrected electromyographic signal by subtracting the reference electromyographic signal value of the same acquisition time stamp from a plurality of second electromyographic signal values acquired within the unit time t, generates a third corrected electromyographic signal by subtracting the reference electromyographic signal value of the same acquisition time stamp from a plurality of third electromyographic signal values acquired within the unit time t, generates a fourth corrected electromyographic signal by subtracting the reference electromyographic signal value of the same acquisition time stamp from a plurality of fourth electromyographic signal values acquired within the unit time t, thereby removing the electromyographic signal interference of the target external muscle group. The first modified electromyographic signal, the second modified electromyographic signal, the third modified electromyographic signal and the fourth modified electromyographic signal all belong to the first electromyographic signal.

The control unit respectively carries out differential derivation on the first corrected movement electromyographic signal, the second corrected movement electromyographic signal, the third corrected movement electromyographic signal and the fourth corrected movement electromyographic signal to remove interference caused by peak-to-peak value floating and phase drifting of signal values, then carries out high-pass filtering on the four groups of corrected movement electromyographic signals respectively to remove system interference such as heartbeat of a user and the like, generates filtering signals formed by the four groups of conditioned and filtered corrected movement electromyographic signals, and the filtering signals form three-dimensional electromyographic distribution characteristics with three types of parameters including acquisition time, electromyographic signal values and numbers of the pasting electrodes, so that the reliability of front-end acquired data and the representation capability of semantic characteristics are greatly improved.

The control unit inputs the filtering signal into the action matching model to generate an action prediction result of the user, namely, a motion control instruction can be output to the action execution unit based on the action prediction result of the user. The type of the action prediction result of the present invention is specifically limited, and the type of the action prediction result may be various, for example: arm relaxation, forward arm, backward arm, arm deployment, arm retraction, and the like. Accordingly, the present invention does not specifically limit the specific mapping relationship between the motion prediction result type and the motion control command, and is related to the setting mode of the mechanical exoskeleton, for example: when the mechanical exoskeleton is arranged on the upper limb of the user as a motion rehabilitation device and the motion prediction result is that the left arm of the user moves forwards, the motion control instruction generated by the control unit can assist the left arm of the user to move forwards; when the mechanical exoskeleton is arranged on the lower limbs of the user as a motion enhancement device, and the motion prediction result is that the left arm of the user is forward, the motion control instruction generated by the control unit can assist the right leg of the user to move forward.

According to the invention, multiple groups of movement electromyographic signals are collected in a mode of surrounding the same target muscle group, and reference electromyographic signals for removing electromyographic signal interference of the target external muscle group are collected, so that after the reference electromyographic signal values with the same collection time stamp are respectively subtracted from the multiple groups of movement electromyographic signal values, differential derivation of each group of movement electromyographic signal values and high-pass filtering, three-dimensional electromyographic distribution characteristics with three types of parameters of collection time, electromyographic signal values and numbers of the applied electrodes can be generated, thus the user action type is accurately represented, and the accuracy of a movement control instruction output by a control unit is effectively improved under the condition of not generating extra calculation force load. And the invention does not need large-scale actual movement in space by detecting whether the muscle is applied with force or not. Therefore, the control unit can send out instructions more quickly, and can sense motion signals more sensitively, and the problems of low motion control reliability and low sensitivity of the traditional motion control method based on the electromyographic signals are solved.

Accordingly, as shown in fig. 3, the present invention provides a motion control system based on electromyographic signals, comprising: the electromyographic signal collector is used for collecting the electromyographic signal of the human body; the control unit is used for conditioning and filtering the electromyographic signals to generate filtering signals, acquiring a pre-trained action matching model, inputting the filtering signals into the action matching model to generate action prediction results, and outputting motion control instructions to the action execution unit based on the action prediction results; and the action execution unit is used for executing the motion control instruction. The motion control system further comprises a cache module used for storing the electromyographic signals; the action execution unit can be an electronic control unit for controlling a mechanical action unit; the mechanical action unit may be a mechanical exoskeleton.

Further, the electromyographic signal collector comprises a plurality of application electrodes, wherein a first application electrode is arranged at the joint of the shoulder and the upper arm of the body and is used for collecting a reference electromyographic signal at the joint of the shoulder and the upper arm of the body within unit time t; the plurality of the application electrodes except the first application electrode are equidistantly arranged on one circle of the upper arm of the body and are used for collecting a plurality of groups of movement electromyographic signals on one circle of the upper arm of the body in unit time t.

Further, the control unit is configured to: respectively subtracting the reference electromyographic signal values of the same acquisition time stamp from the multiple groups of the motion electromyographic signal values to generate multiple groups of first electromyographic signals; and after differential derivation and high-pass filtering are carried out on the multiple groups of first electromyographic signals, generating a filtering signal capable of representing electromyographic distribution characteristics at one week of the upper arm of the human body.

The electromyographic signal-based motion control method and the system thereof provided by the embodiment of the invention are described in detail above. The embodiments are described in a progressive manner in the specification, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description. It should be noted that, for those skilled in the art, it is possible to make various improvements and modifications to the present invention without departing from the principle of the present invention, and those improvements and modifications also fall within the scope of the claims of the present invention.

Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative components and steps have been described above generally in terms of their functionality in order to clearly illustrate this interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may reside in Random Access Memory (RAM), memory, Read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

Claims

1. A method for controlling a movement based on an electromyographic signal, comprising:

collecting myoelectric signals of a human body;

conditioning and filtering the electromyographic signals to generate filtering signals;

obtaining a pre-trained action matching model, and inputting the filtering signal into the action matching model to generate an action prediction result;

and outputting a motion control instruction to a motion execution unit based on the motion prediction result.

2. The exercise control method according to claim 1, wherein acquiring an electromyographic signal of a body includes:

collecting a reference electromyographic signal at the joint of the shoulder and the upper arm of the human body within unit time t;

and collecting a plurality of groups of movement electromyographic signals of the upper arm of the body in unit time t.

3. The method of claim 2, wherein conditioning the electromyographic signals to generate filtered signals comprises:

respectively subtracting the reference electromyographic signal values of the same acquisition time stamp from the multiple groups of the motion electromyographic signal values to generate multiple groups of first electromyographic signals;

and after differential derivation and high-pass filtering are carried out on the multiple groups of first electromyographic signals, generating a filtering signal capable of representing electromyographic distribution characteristics at one week of the upper arm of the human body.

4. The motion control method of claim 3, wherein training the motion matching model comprises:

constructing an initialization network model, wherein the network model comprises a semantic matching model;

acquiring a training data set and a test data set which are composed of electromyographic signal samples containing labeling action types and electromyographic distribution characteristics;

training and testing the network model based on the training dataset and the testing dataset.

5. The motion control method according to claim 1, characterized in that the motion control method further comprises:

collecting the electromyographic signals of the body, and caching the original electromyographic signals.

6. A myoelectric signal-based motion control system, comprising:

the electromyographic signal collector is used for collecting the electromyographic signal of the human body;

the control unit is used for conditioning and filtering the electromyographic signals to generate filtering signals, acquiring a pre-trained action matching model, inputting the filtering signals into the action matching model to generate action prediction results, and outputting motion control instructions to the action execution unit based on the action prediction results;

and the action execution unit is used for executing the motion control instruction.

7. The motion control system of claim 6, wherein the electromyographic signal collector comprises a plurality of application electrodes, wherein,

the first pasting electrode is arranged at the joint of the shoulder and the upper arm of the human body and is used for collecting a reference electromyographic signal at the joint of the shoulder and the upper arm of the human body within unit time t;

the plurality of the application electrodes except the first application electrode are equidistantly arranged on one circle of the upper arm of the body and are used for collecting a plurality of groups of movement electromyographic signals on one circle of the upper arm of the body in unit time t.

8. The motion control system of claim 7, wherein the control unit is configured to: respectively subtracting the reference electromyographic signal values of the same acquisition time stamp from the multiple groups of the motion electromyographic signal values to generate multiple groups of first electromyographic signals; and after differential derivation and high-pass filtering are carried out on the multiple groups of first electromyographic signals, generating a filtering signal capable of representing electromyographic distribution characteristics at one week of the upper arm of the human body.

9. The exercise control system of claim 6, further comprising a buffer module to store the electromyographic signals.

10. The motion control system of claim 6, wherein the motion-performing unit is an electronic control unit that manipulates a mechanical action unit.