Exoskeleton action recognition method and device, equipment, storage medium and program product

Abstract

The embodiment of the invention provides an exoskeleton action recognition method and device, equipment, a storage medium and a program product, relates to the technical field of human exoskeletons, can realize accurate action recognition of a wearer of exoskeleton auxiliary equipment, reduces discomfort generated in the wearing process of the exoskeleton auxiliary equipment, and improves the wearing experience of the wearer. The auxiliary effect of the exoskeleton auxiliary equipment on the operating personnel is improved, and the movement speed and the operating efficiency of the wearing personnel can be further improved. The method comprises the following steps: acquiring an electromyographic signal acquisition point location of a target recognition person; on the basis of the electromyographic signal acquisition point location, acquiring an electromyographic action signal of the target recognition person; acquiring an auxiliary action point location of a target exoskeleton of the target recognition person; acquiring an auxiliary action signal of the target exoskeleton according to the auxiliary action point location; and determining action information of the target recognition person based on the myoelectricity action signal and the auxiliary action signal.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of human exoskeleton, and in particular to an exoskeleton motion recognition method, apparatus, device, storage medium and program product. Background Art

[0002] Activities involving heavy lifting and repeated lifting often cause certain impacts and damage to the body structure of the personnel involved. Especially when carrying heavy objects and working for long periods of time, the human skeleton, muscles, joints and other parts will be under greater pressure and fatigue, affecting the health of the personnel involved and increasing the risk of occupational diseases and other diseases. Therefore, for activities with high intensity, practitioners usually need to be equipped with exoskeleton devices to assist in movement in order to reduce the pressure on the human skeleton, muscles and joints.

[0003] Exoskeleton assistive equipment usually needs to assist movements based on the wearer's intended movements, but traditional exoskeleton assistive equipment usually finds it difficult to accurately identify the wearer's different movements, which can easily cause discomfort to the wearer or poor assistive effect, further affecting the wearer's movement speed and work efficiency. Summary of the invention

[0004] The exoskeleton motion recognition method, device, equipment, storage medium and program product provided in the embodiments of the present application can realize accurate motion recognition of the wearer of the exoskeleton auxiliary equipment, reduce the discomfort caused by the wearing of the exoskeleton auxiliary equipment, improve the auxiliary effect of the exoskeleton auxiliary equipment on the operating personnel, and further improve the movement speed and working efficiency of the wearer.

[0005] According to a first aspect of an embodiment of the present application, there is provided an exoskeleton action recognition method, comprising:

[0006] Obtain the electromyographic signal collection points of the target identification personnel;

[0007] Based on the electromyographic signal collection points, obtaining the electromyographic motion signals of the target identification personnel;

[0008] Acquire auxiliary action points of the target exoskeleton of the target recognition person;

[0009] Acquiring an auxiliary action signal of the target exoskeleton according to the auxiliary action point;

[0010] Based on the myoelectric action signal and the auxiliary action signal, the action information of the target identification person is determined.

[0011] In some embodiments, the electromyographic action signal includes at least one of an anterior deltoid signal, a middle deltoid signal, a lower trapezius signal, and a triceps signal.

[0012] In some implementations, the auxiliary action signal includes an angle adjustment signal, a force adjustment signal, and corresponding adjustment point information of the target exoskeleton.

[0013] In some embodiments, before the step of determining the motion information of the target identification person based on the electromyographic motion signal and the auxiliary motion signal, the method further includes:

[0014] The myoelectric compensation parameters corresponding to the myoelectric action signal are determined according to the myoelectric signal collection points and the auxiliary action points.

[0015] In some embodiments, determining the motion information of the target identification person based on the electromyographic motion signal and the auxiliary motion signal includes:

[0016] Determining a myoelectric correction signal of the target identification person based on the myoelectric compensation parameter and the auxiliary action signal;

[0017] The motion information of the target identification person is determined according to the myoelectric correction signal and the myoelectric motion signal.

[0018] In some embodiments, the exoskeleton motion recognition method further includes:

[0019] A target auxiliary signal of the target exoskeleton is determined according to the motion information and the myoelectric motion signal.

[0020] According to a second aspect of an embodiment of the present application, there is provided an exoskeleton action recognition device, comprising:

[0021] The first acquisition module is used to acquire the electromyographic signal collection points of the target identification personnel;

[0022] A second acquisition module is used to acquire the myoelectric motion signal of the target identification person based on the myoelectric signal collection point;

[0023] A third acquisition module is used to acquire auxiliary action points of the target exoskeleton of the target recognition person;

[0024] A fourth acquisition module, used for acquiring an auxiliary action signal of the target exoskeleton according to the auxiliary action point;

[0025] A determination module is used to determine the motion information of the target identification person based on the electromyographic motion signal and the auxiliary motion signal.

[0026] According to a third aspect of an embodiment of the present application, a computer device is provided, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the exoskeleton action recognition method described in any one of the first aspects are implemented.

[0027] According to a fourth aspect of the embodiments of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the exoskeleton action recognition method described in any one of the first aspects are implemented.

[0028] According to a fifth aspect of the embodiments of the present application, a computer program product is provided, comprising a computer program, wherein when the computer program is executed by a processor, the steps of the exoskeleton action recognition method described in any one of the first aspects are implemented.

[0029] The embodiments of the present application provide an exoskeleton motion recognition method, device, equipment, storage medium and program product, which obtain the electromyographic signal collection points of the target identification personnel, and based on the electromyographic signal collection points, obtain the electromyographic motion signals of the target identification personnel, so as to determine the muscle force state of different parts of the current target identification personnel, and further obtain the auxiliary action points of the target exoskeleton of the target identification personnel, and, according to the auxiliary action points, obtain the auxiliary action signals of the target exoskeleton, so as to determine the auxiliary force state of the target exoskeleton worn by the target identification personnel, so as to determine the motion state of different positions on the body of the target identification personnel based on the electromyographic motion signals and the auxiliary action signals, and further determine the motion information of the target identification personnel. It is possible to realize accurate motion recognition of the wearer of the exoskeleton auxiliary equipment, reduce the discomfort generated during the wearing of the exoskeleton auxiliary equipment, improve the auxiliary effect of the exoskeleton auxiliary equipment on the operator, and further improve the movement speed and work efficiency of the wearer. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0031] FIG1 is a schematic flow chart of an exoskeleton action recognition method provided in an embodiment of the present application;

[0032] FIG2 is a schematic structural diagram of an exoskeleton action recognition method and device provided in an embodiment of the present application;

[0033] FIG3 is a schematic structural diagram of a computer device provided in an embodiment of the present application;

[0034] FIG4 is a schematic structural diagram of a computer-readable storage medium provided in an embodiment of the present application;

[0035] FIG5 is a schematic structural diagram of a computer program product provided in an embodiment of the present application. DETAILED DESCRIPTION

[0036] In order to facilitate understanding of the present application, the present application will be described more fully below with reference to the relevant drawings. Embodiments of the present application are provided in the drawings. However, the present application can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present application more thorough and comprehensive.

[0037] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application.

[0038] When used herein, the singular forms "a", "an" and " / the" may also include the plural forms, unless the context clearly indicates otherwise. It should also be understood that the terms "include / comprise" or "have" and the like specify the presence of stated features, wholes, steps, operations, components, parts or combinations thereof, but do not exclude the possibility of the presence or addition of one or more other features, wholes, steps, operations, components, parts or combinations thereof. At the same time, in this specification, the term "and / or" includes any and all combinations of the relevant listed items.

[0039] It should be noted that activities involving heavy lifting and repeated lifting often cause certain impacts and damage to the body structure of the personnel involved. Especially when carrying heavy objects and working for long periods of time, the human skeleton, muscles, joints and other parts will be under greater pressure and fatigue, affecting the health of the personnel involved and increasing the risk of occupational diseases and other diseases. Therefore, for activities with high intensity, practitioners usually need to be equipped with exoskeleton devices to assist in movement in order to reduce the pressure on the human skeleton, muscles and joints.

[0040] Exoskeleton assistive equipment usually needs to assist movements based on the wearer's intended movements, but traditional exoskeleton assistive equipment usually finds it difficult to accurately identify the wearer's different movements, which can easily cause discomfort to the wearer or poor assistive effect, further affecting the wearer's movement speed and work efficiency.

[0041] The DC converter station is a complex workplace. We need to provide employees with a good working environment and working conditions. One of the key aspects is to continuously track the dynamics and postures of employees at work and provide timely feedback to help them optimize their work movements and reduce physical fatigue and potential work-related injury risks. However, due to the complexity of the site environment and the diversity of human movements, previous motion recognition methods often fail to accurately identify complex work movements, and the recognition effect of fast and continuous work movements is also poor.

[0042] In view of the complexity and diversity of human movements, commonly used motion recognition methods mainly include vision-based motion recognition and sensor-based motion recognition. Visual motion recognition mainly relies on the camera to capture the image information of individual movements, and then uses image processing and pattern recognition technology to perform motion recognition. Sensor motion recognition is to install sensors on the individual to capture motion-related physiological signals such as EEG signals and EMG signals, and then use signal processing and pattern recognition technology to perform motion recognition.

[0043] Although these two motion recognition methods have their own advantages, the visual recognition method is sensitive to problems such as lighting and occlusion, and it is difficult to obtain internal physiological state information; although the sensor method can obtain the internal physiological state, it has high requirements on the installation position and angle of the sensor, and in a complex working environment, the sensor may not be able to work stably, which brings certain troubles to motion recognition.

[0044] Therefore, proposing a method that can accurately identify the movements of people wearing exoskeleton-assisted equipment is an urgent problem to be solved.

[0045] As shown in Figure 1, Figure 1 is a schematic flow chart of an exoskeleton action recognition method provided in an embodiment of the present application. In a first aspect of an embodiment of the present application, an exoskeleton action recognition method is provided, comprising:

[0046] Step S110, obtaining the electromyographic signal collection points of the target identification person.

[0047] For example, the electromyographic signal collection points can be determined according to the wearing position of the target exoskeleton.

[0048] It should be noted that the myoelectric signal collection points are evenly and symmetrically distributed on the left and right sides of the target identification person's body. The myoelectric signal collection points include at least one group.

[0049] It should be noted that the attention mechanism, as an emerging data processing technology, has the unique ability to focus attention on the input sequence at a specified time step. This technology was first applied in the study of the human visual system. In actual image recognition tasks, humans usually focus their visual attention on a specific area of ​​the image over a continuous time span, and as time goes by, the focus will gradually change. Drawing on this principle, the artificial intelligence system can focus on the key areas of the image, while presetting other inconspicuous areas as invisible, which has a positive impact on the final recognition task. It has been proven to be effective in the field of natural language recognition. The data involved in CSI-based action recognition has characteristics such as high latitude and temporal continuity, which are extremely similar to the data characteristics in natural language and visual system research. Therefore, this study attempts to introduce the attention mechanism into the action recognition topic in order to further improve the accuracy of the recognition task.

[0050] Different actions show different characteristics in different wireless links. This solution proposes a feature map fusion method based on the attention mechanism (Feature Map Fusion Attention, FMFA), which globally retains the denoised data and extracts features from the three antennas separately to retain the features of different actions to the greatest extent. The data of each antenna is convolved by three convolution kernels, and the differences in different sizes can increase the flexibility and generalization ability of the model. After feature extraction, three feature maps are generated for each channel. For these three feature maps, this solution uses the attention mechanism for fusion instead of simple splicing or addition. As shown in the figure below, the purpose of FMFA is to allow the model to selectively focus on different parts of the feature map and assign them different weights to improve the performance of the model. In the process of feature map fusion, some feature maps may contribute more to the task, while some may contribute less. If the feature maps are simply spliced ​​together, the relationship between different feature maps cannot be fully utilized, which is prone to information redundancy and overfitting problems.

[0051] FMFA uses the relationship between feature maps of different scales to generate feature maps with different weights. This mechanism maps the input three channel feature maps into a new space, calculates their similarity through convolution, and uses the Relu function to strengthen the nonlinear relationship. The similarity is aggregated to obtain the attention weight of each channel, and convolution is used to map it to a scalar space, and then the Sigmoid function is used to limit the output to the range of [0,1] to obtain the attention weight of each channel. Finally, the original feature map is multiplied by the obtained attention weight to obtain the weighted feature map. The specific expression is as follows:

[0052]

[0053] Indicates channels, represents the convolution kernel parameter, represents the CSI data after data preprocessing, By The convolution kernel generates The feature map of each channel, It means that by introducing nonlinear mechanism Highlight important features. Indicates that the feature map is passed through Limit the output range to [0.1] It means that the three feature maps of each channel are converged so that their weight sum is 1. Represents the final fused feature map of each channel.

[0054] In the field of computer vision, channel attention (CA) plays a significant role. CA is used to strengthen the attention of deep neural networks to different channels, thereby improving the performance of the model. The core idea of ​​CA is to learn a weight on each channel, adjust the contribution of channel features, and make the model focus on important features. Under MIMO antenna diversity, multiple antennas have different representations of the same action. Therefore, it can be abstractly understood that each antenna corresponds to each feature channel, and processing antennas can be considered as processing feature channels. When processing time series data, previous studies often focus on the relationship between features between different time steps, while ignoring the different representation forms of action features by different channels. As mentioned above, this scheme extracts features by setting three channels of three antennas of the acquisition device to avoid the influence of different feature representations, improve the model's attention to important information in the input features, and reduce the attention to useless features. This scheme proposes a feature channel attention mechanism (FCA), which aims to capture the interdependence between different feature channels in convolutional neural networks. First, the input feature map is average pooled and max pooled, and the weights are aggregated in the channel dimension while retaining the spatial dimension. The resulting tensor is then processed through a series of convolutional layers and activation layers to generate a set of channel attention weights for each input tensor. As shown in the formula:

[0055]

[0056] Therefore, the EMG signal collection points can be determined based on the above-mentioned feature channel attention mechanism.

[0057] Step S120: based on the electromyographic signal collection points, obtain the electromyographic motion signals of the target identification person.

[0058] It should be noted that the electromyographic action signal is a bioelectric current signal caused by muscle contraction, and the voltage amplitude of the electromyographic action signal can reflect the intensity of muscle activity to a certain extent.

[0059] For example, the ME60-8 electromyography system can be used to capture human motion data. The sampling frequency can be 1200Hz, and the bandpass filter frequency ranges from 20Hz to 500Hz. The common mode rejection ratio is 110dB, the input impedance is 10GΩ, the gain is 1000, the noise level is <1μV, the sensitivity is 0.2μV, and the signal A / D conversion is 12bit.

[0060] Exemplarily, when acquiring electromyographic action signals, the skin area can be cleaned with alcohol wipes to reduce interference and noise, and silver-silver chloride surface electrodes are connected to the muscle belly in pairs with a center distance of 2.5 cm. In each experiment, all raw electromyographic signals were rectified and low-pass filtered using a fourth-order Butterworth filter (bidirectional) with a cutoff frequency of 3 Hz to generate a linear envelope. The electromyographic signal data collected in the first 3 seconds of each trial was not processed later to reduce transition period errors. Then in each trial, the RMS (Root Mean Square) value of the electromyographic signal was calculated, and the median of the muscle activity was determined.

[0061] Step S130: Acquire auxiliary action points of the target exoskeleton of the target recognition person.

[0062] For example, the auxiliary action point can be determined according to the movable range of the target exoskeleton, wherein the auxiliary action point and the electromyographic signal collection point can include the same position or different positions.

[0063] Step S140: Acquire the auxiliary action signal of the target exoskeleton according to the auxiliary action point.

[0064] Exemplarily, the control signal of the target exoskeleton may be obtained, and then the auxiliary action signal of the target exoskeleton may be obtained.

[0065] Step S150: Determine the motion information of the target person based on the electromyographic motion signal and the auxiliary motion signal.

[0066] For example, the action status of different positions on the target person's body can be determined based on the electromyographic action signal and the auxiliary action signal. The action information includes the action height of the target person, the limb extension state of the target person, the limb force state of the target person, etc., and can also include the weight information of the target person.

[0067] The embodiment of the present application provides an exoskeleton motion recognition method, which obtains the electromyographic signal collection points of the target identification personnel, and based on the electromyographic signal collection points, obtains the electromyographic motion signals of the target identification personnel, so as to determine the muscle force state of different parts of the current target identification personnel, and further obtains the auxiliary action points of the target exoskeleton of the target identification personnel, and, according to the auxiliary action points, obtains the auxiliary action signals of the target exoskeleton, so as to determine the auxiliary force state of the target exoskeleton worn by the target identification personnel, so as to determine the motion state of different positions on the body of the target identification personnel based on the electromyographic motion signals and the auxiliary action signals, and further determine the motion information of the target identification personnel. It is possible to realize accurate motion recognition of the wearer of the exoskeleton auxiliary equipment, reduce the discomfort generated during the wearing of the exoskeleton auxiliary equipment, improve the auxiliary effect of the exoskeleton auxiliary equipment on the operator, and further improve the movement speed and work efficiency of the wearer.

[0068] In some feasible implementations, the electromyographic action signal includes at least one of anterior deltoid muscle signal, middle deltoid muscle signal, lower trapezius muscle signal, and triceps brachii muscle signal.

[0069] It should be noted that the anterior deltoid muscle is mainly responsible for the flexion, internal rotation and horizontal adduction of the shoulder joint. By monitoring the anterior deltoid muscle signal of the target identification personnel. The main function of the middle deltoid muscle is to be responsible for the abduction of the shoulder joint, that is, to help the arm lift from the side of the body to a horizontal position, and can also assist in the external rotation of the shoulder joint, which is the action of rotating the arm outward so that the palm faces up. Therefore, the anterior deltoid muscle signal and the middle deltoid muscle signal can be used to determine the shoulder movement state of the target identification personnel. The lower trapezius muscle is mainly responsible for lowering the scapula, forming a scapular force couple, and tilting the scapula backward. The triceps brachii is mainly responsible for straightening the elbow joint, as well as extending and adducting the shoulder joint.

[0070] The embodiment of the present application provides an exoskeleton motion recognition method. By acquiring the anterior deltoid signal, the middle deltoid signal, the lower trapezius signal and the triceps signal, the comprehensiveness and diversity of the electromyographic motion signal can be increased, and the movement state of the shoulder of the target identification person can be further accurately determined, thereby facilitating the improvement of the accuracy of motion recognition, and further realizing accurate motion recognition of the wearer of the exoskeleton auxiliary equipment, reducing the discomfort caused by the wearing of the exoskeleton auxiliary equipment, improving the auxiliary effect of the exoskeleton auxiliary equipment on the operator, and further improving the movement speed and work efficiency of the wearer.

[0071] In some feasible implementations, the auxiliary action signal includes an angle adjustment signal, a force adjustment signal, and corresponding adjustment point information of the target exoskeleton.

[0072] Exemplarily, the angle adjustment signal of the target exoskeleton can be obtained through the angle sensor, the force adjustment signal of the target exoskeleton can be obtained through the pressure sensor, and the adjustment point information corresponding to the angle adjustment signal and the force adjustment signal can be determined through the setting positions of the angle sensor and the pressure sensor.

[0073] It should be noted that the more sensors a system has, the richer the information it obtains, and the larger the dimension of the classification feature, but a higher feature dimension does not mean a better classification effect. On the one hand, if the feature dimension is too high, the amount of calculation of the classification model will increase, and the calculation processing time will be longer. On the other hand, if there are too many features, the features are prone to interfere with each other, which directly affects the classification accuracy. Therefore, it is necessary to select features before classification. Common feature selection methods include filtering, encapsulation, and embedded methods. Among them, Relie F is a typical embedded method. The Relie F algorithm was proposed by scientists in 1992 and is mainly used to solve binary classification problems.

[0074] The core idea of ​​the Relief F algorithm is to assign different weights based on the different ability of the feature to distinguish the k nearest neighbor samples of the same type and the k nearest neighbor samples of different types. Specifically, first randomly select a sample R from the feature training set M, and then select k nearest neighbor samples from the same type of samples in R, denoted as Distance is Then select k nearest neighbor samples from R different types of samples, denoted as , the distance is When the distance between samples of the same type is greater than the distance between samples of different types, the weight of the feature is reduced, otherwise, it is increased. Repeat the above process n times, continuously update the weight, and finally get the weight of each feature. The weight calculation formula is as follows:

[0075]

[0076] In the formula Represents sample R and similar samples The distance on feature A, C is the sample category label, P(C) is the ratio of samples with sample category C to the total number of samples, Expressed as The ratio of samples of the same type to the total number of samples.

[0077] Principal component analysis (PCA) is a data dimensionality reduction method based on vector basis transformation. It linearly maps G-dimensional feature vectors to D-dimensional space (G>D), expecting the original feature vector dimension to be reduced, and the D-dimensional space to retain the principal component information as much as possible, using the least number of dimensions to represent the original features, and achieving the effect of simplifying the complex. The main idea of ​​PCA is to reduce the G-dimensional feature vector to D-dimensional, and its goal is to orderly find D unit orthogonal bases in the G-dimensional space, so that after the original feature vector is transformed to this set of bases, the covariance between the feature vectors is 0 and the variance is maximized.

[0078] Assume that the original feature vector is expressed as shown in the formula, where M is the number of samples and N is the feature dimension.

[0079]

[0080] The first step in solving PCA is to standardize the features to eliminate the influence of the magnitude of samples. Subtract the mean of each feature from the dimension and divide it by the standard deviation to obtain a new feature vector Y.

[0081]

[0082] In the formula is the j-th dimension mean of the feature vector, is the j-th dimension mean of the eigenvector, and .

[0083] The second step is to calculate the covariance matrix C, find the direction with the largest variance of the projection of all features in a certain direction to the center distance, and transform the dimensionality reduction problem into a maximum value problem. Since features of different dimensions are not independent of each other, covariance is introduced. The sample covariance calculation formula is as follows:

[0084]

[0085] The third step is to calculate the eigenvalues ​​of the covariance matrix C and the eigenvector :

[0086]

[0087] The eigenvalue λ is the variance projected on a principal component, and each λ corresponds to an eigenvector μ. The eigenvalues ​​λ are sorted from large to small, and the first k eigenvectors are selected to form the matrix P. Z=PX is the data after the original eigenvector is reduced to k dimensions. The last step is to determine the number of principal components, and use the formula to calculate the cumulative variance contribution rate of the first k eigenvectors. The PCA cumulative variance contribution rate represents the proportion of the variance contained in the first k principal components to the total variance, and is used to measure the amount of information retained in the data after PCA dimensionality reduction. Generally speaking, the first k principal components with a cumulative contribution rate of more than 95% are used as new eigenvectors, which retains more original signals while achieving the effect of removing redundancy.

[0088] The embodiment of the present application provides an exoskeleton motion recognition method. By acquiring the angle adjustment signal, force adjustment signal and corresponding adjustment point information of the target exoskeleton, the angle adjustment signal and the force adjustment signal can be corrected based on the adjustment point information when the auxiliary action point is inconsistent with the electromyographic signal acquisition point, thereby further improving the practicality, accuracy and objectivity of the auxiliary action signal, thereby facilitating improving the accuracy of motion recognition, and further realizing accurate motion recognition of the wearer of the exoskeleton auxiliary equipment, reducing the discomfort caused by the wearing of the exoskeleton auxiliary equipment, improving the auxiliary effect of the exoskeleton auxiliary equipment on the operator, and further improving the movement speed and work efficiency of the wearer.

[0089] In some feasible implementations, before the step of determining the motion information of the target identification person based on the electromyographic motion signal and the auxiliary motion signal, it also includes: determining the electromyographic compensation parameters corresponding to the electromyographic motion signal according to the electromyographic signal collection points and the auxiliary motion points.

[0090] It should be noted that when the target identification personnel performs an action, the muscles involved in the action can serve as active muscles. When external force is applied to the target identification personnel through the target exoskeleton, since the target exoskeleton has multiple interconnected driving devices, it may also drive some non-active muscles to move, thereby generating electromyographic signals, or exerting pressure on the active muscles, so that the active muscles need to exert a certain supporting force on the target exoskeleton, resulting in that not all of the electromyographic signals of the active muscles are used to participate in effective actions.

[0091] Exemplarily, when the auxiliary action point is inconsistent with the electromyographic signal collection point, the amplitude of the change in the electromyographic signal caused by the traction relationship can be determined based on the auxiliary action signal according to the traction relationship between the muscle group near the auxiliary action point and the muscle group near the electromyographic signal collection point, thereby determining the electromyographic compensation parameters corresponding to the electromyographic action signal.

[0092] Exemplarily, when the auxiliary action point is consistent with the electromyographic signal collection point, the auxiliary effect of the target exoskeleton on the target identification person's body activity at the auxiliary action point can be judged as positive or negative according to the activity direction of the target exoskeleton corresponding to the auxiliary action signal and the body activity direction of the target identification person corresponding to the electromyographic action signal. Among them, when the auxiliary effect is positive, it can be determined that the actual movement amplitude of the target identification person is greater than the movement amplitude corresponding to the electromyographic action signal. Similarly, when the auxiliary effect is negative, it can be determined that the actual movement amplitude of the target identification person is less than the movement amplitude corresponding to the electromyographic action signal.

[0093] The embodiment of the present application provides an exoskeleton motion recognition method. By determining the electromyographic compensation parameters corresponding to the electromyographic motion signals based on the electromyographic signal collection points and the auxiliary action points, it is possible to avoid interference of the auxiliary effect of the target exoskeleton on the motion recognition results, and further correct the motion recognition results of the target identification personnel, which can facilitate the improvement of the accuracy of motion recognition, and further can realize accurate motion recognition of the wearer of the exoskeleton auxiliary equipment, reduce the discomfort caused by the wearing of the exoskeleton auxiliary equipment, improve the auxiliary effect of the exoskeleton auxiliary equipment on the operator, and further can improve the movement speed and work efficiency of the wearer.

[0094] In some feasible implementations, the motion information of the target person is determined based on the electromyographic motion signal and the auxiliary motion signal, including: determining the electromyographic correction signal of the target person based on the electromyographic compensation parameter and the auxiliary motion signal; determining the motion information of the target person based on the electromyographic correction signal and the electromyographic motion signal.

[0095] The embodiment of the present application provides an exoskeleton motion recognition method, which determines the myoelectric correction signal of the target identification person through myoelectric compensation parameters and auxiliary motion signals, and further determines the motion information of the target identification person based on the myoelectric correction signal and the myoelectric motion signal, so as to avoid the auxiliary effect of the target exoskeleton from interfering with the motion recognition result, and further correct the motion recognition result of the target identification person, which can facilitate improving the accuracy of motion recognition, and further can realize accurate motion recognition of the wearer of the exoskeleton auxiliary equipment, reduce the discomfort caused by the wearing of the exoskeleton auxiliary equipment, improve the auxiliary effect of the exoskeleton auxiliary equipment on the operator, and further can improve the movement speed and work efficiency of the wearer.

[0096] In some feasible implementations, the exoskeleton motion recognition method further includes: determining a target auxiliary signal of the target exoskeleton based on the motion information and the myoelectric motion signal.

[0097] Exemplarily, the target auxiliary signal is used to drive the target exoskeleton.

[0098] The embodiment of the present application provides an exoskeleton motion recognition method, which determines the target auxiliary signal of the target exoskeleton through motion information and electromyographic motion signals, and can correct the auxiliary motion of the target exoskeleton, thereby improving the auxiliary effect of the target exoskeleton, further reducing the fatigue and discomfort of the wearer, improving the auxiliary effect of the exoskeleton auxiliary equipment on the operator, and further improving the movement speed and work efficiency of the wearer.

[0099] Based on the same inventive concept, the embodiment of the present application also provides an exoskeleton action recognition device for implementing the exoskeleton action recognition method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more exoskeleton action recognition device embodiments provided below can refer to the limitations on the exoskeleton action recognition method above, and will not be repeated here.

[0100] Fig. 2 is a schematic structural diagram of an exoskeleton motion recognition device provided by an embodiment of the present application. As shown in Fig. 2, the second aspect of the embodiment of the present application provides an exoskeleton motion recognition device, including: a first acquisition module 100, a second acquisition module 200, a third acquisition module 300, a fourth acquisition module 400 and a determination module 500. Among them, the first acquisition module 100 is used to obtain the electromyographic signal acquisition points of the target identification person; the second acquisition module 200 is used to obtain the electromyographic motion signals of the target identification person based on the electromyographic signal acquisition points; the third acquisition module 300 is used to obtain the auxiliary action points of the target exoskeleton of the target identification person; the fourth acquisition module 400 is used to obtain the auxiliary action signals of the target exoskeleton according to the auxiliary action points; the determination module 500 is used to determine the motion information of the target identification person based on the electromyographic motion signals and the auxiliary action signals.

[0101] The embodiment of the present application provides an exoskeleton motion recognition device, which obtains the electromyographic signal collection points of the target identification person, and based on the electromyographic signal collection points, obtains the electromyographic motion signals of the target identification person, so as to determine the muscle force state of different parts of the current target identification person, and further obtains the auxiliary action points of the target exoskeleton of the target identification person, and, according to the auxiliary action points, obtains the auxiliary action signals of the target exoskeleton, so as to determine the auxiliary force state of the target exoskeleton worn by the target identification person, so as to determine the motion state of different positions on the body of the target identification person based on the electromyographic motion signals and the auxiliary action signals, and further determine the motion information of the target identification person. It is possible to realize accurate motion recognition of the wearer of the exoskeleton auxiliary equipment, reduce the discomfort generated during the wearing of the exoskeleton auxiliary equipment, improve the auxiliary effect of the exoskeleton auxiliary equipment on the operator, and further improve the movement speed and work efficiency of the wearer.

[0102] Each module in the above exoskeleton action recognition device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each module.

[0103] FIG3 is a schematic structural diagram of a computer device provided in an embodiment of the present application. As shown in FIG3, in a third aspect of an embodiment of the present application, a computer device is provided, which may be a server. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile 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 the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, a near field communication (Near Field Communication, NFC) or other technologies. When the computer program is executed by the processor, an exoskeleton action recognition method is implemented. The display unit of the computer device is used to form a visually visible picture, and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball, or touchpad provided on the computer device housing, or an external keyboard, touchpad, or mouse, etc.

[0104] Those skilled in the art will understand that the structure shown in FIG. 3 is merely a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0105] Fig. 4 is a schematic structural diagram of a computer-readable storage medium provided in an embodiment of the present application. As shown in Fig. 4, in a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of any one of the exoskeleton action recognition methods in the first aspect are implemented.

[0106] Fig. 5 is a schematic structural diagram of a computer program product provided in an embodiment of the present application. As shown in Fig. 5, in a fifth aspect of an embodiment of the present application, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the steps of any one of the exoskeleton action recognition methods in the first aspect.

[0107] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0108] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0109] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. An exoskeleton action recognition method, characterized in that: include: Obtain the electromyographic signal collection points of the target identification personnel; Based on the electromyographic signal collection points, obtaining the electromyographic motion signals of the target identification personnel; Acquire auxiliary action points of the target exoskeleton of the target recognition person; Acquiring an auxiliary action signal of the target exoskeleton according to the auxiliary action point; Based on the myoelectric action signal and the auxiliary action signal, the action information of the target identification person is determined.

2. The exoskeleton action recognition method according to claim 1, characterized in that: The myoelectric action signal includes at least one of an anterior deltoid muscle signal, a middle deltoid muscle signal, a lower trapezius muscle signal, and a triceps brachii muscle signal.

3. The exoskeleton action recognition method according to claim 1, characterized in that: The auxiliary action signal includes an angle adjustment signal, a force adjustment signal and corresponding adjustment point information of the target exoskeleton.

4. The exoskeleton action recognition method according to claim 1, characterized in that: Before the step of determining the motion information of the target identification person based on the myoelectric motion signal and the auxiliary motion signal, the method further includes: The myoelectric compensation parameters corresponding to the myoelectric action signal are determined according to the myoelectric signal collection points and the auxiliary action points.

5. The exoskeleton action recognition method according to claim 4, characterized in that: The step of determining the motion information of the target identification person based on the electromyographic motion signal and the auxiliary motion signal includes: Determining a myoelectric correction signal of the target identification person based on the myoelectric compensation parameter and the auxiliary action signal; The motion information of the target identification person is determined according to the myoelectric correction signal and the myoelectric motion signal.

6. The exoskeleton action recognition method according to claim 1, characterized in that: Also includes: A target auxiliary signal of the target exoskeleton is determined according to the motion information and the myoelectric motion signal.

7. An exoskeleton motion recognition device, characterized in that: include: The first acquisition module is used to acquire the electromyographic signal collection points of the target identification personnel; A second acquisition module is used to acquire the myoelectric motion signal of the target identification person based on the myoelectric signal collection point; A third acquisition module is used to acquire auxiliary action points of the target exoskeleton of the target recognition person; A fourth acquisition module, used for acquiring an auxiliary action signal of the target exoskeleton according to the auxiliary action point; A determination module is used to determine the motion information of the target identification person based on the electromyographic motion signal and the auxiliary motion signal.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the exoskeleton action recognition method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the exoskeleton action recognition method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the exoskeleton action recognition method according to any one of claims 1 to 6 are implemented.

Images