阅读说明：本技术 具备用户自适应能力的肌电传感方法及系统 (Myoelectricity sensing method and system with user self-adaption capability ) 是由 易港 薛光辉 于 2021-01-20 设计创作，主要内容包括：本发明提供了一种具备用户自适应能力的肌电传感方法及系统,包括：步骤S1：采集前臂多块不同肌肉的表面肌电信号,将多通道的肌肉电信号汇总求多通道的肌肉电信号绝对值平均值,获取多通道的肌肉电信号绝对值平均值信息；步骤S2：根据多通道的肌肉电信号绝对值平均值信息,取多个连续时间序列对应的信号强度求平均；步骤S3：对各通道行动段求区间的平均值MAV,将区间的平均值MAV作为特征向量对信号进行特征提取,获取特征提取结果信息；步骤S4：根据特征提取结果信息,将肌肉电信号分为4类。本发明把开源算法中8通道肌电算法,转化成2通道肌电算法,使得识别率大大提高。(The invention provides a myoelectricity sensing method and a myoelectricity sensing system with user self-adaptive capacity, which comprise the following steps: step S1: collecting surface electromyographic signals of a plurality of different muscles on the forearm, summarizing the multichannel muscle electric signals to obtain the multichannel muscle electric signal absolute value average value, and obtaining the multichannel muscle electric signal absolute value average value information; step S2: averaging the signal intensities corresponding to a plurality of continuous time sequences according to the average value information of the absolute values of the muscle electric signals of the plurality of channels; step S3: calculating an average value MAV of the interval for each channel action section, and taking the average value MAV of the interval as a feature vector to perform feature extraction on the signal to obtain feature extraction result information; step S4: and (4) according to the feature extraction result information, dividing the muscle electric signals into 4 types. The invention converts the 8-channel electromyography algorithm in the open source algorithm into the 2-channel electromyography algorithm, so that the recognition rate is greatly improved.)

1. A myoelectric sensing method with user self-adaptive capability is characterized by comprising the following steps:

step S1: collecting surface electromyographic signals of a plurality of different muscles on the forearm, summarizing the multichannel muscle electric signals to obtain the multichannel muscle electric signal absolute value average value, and obtaining the multichannel muscle electric signal absolute value average value information;

step S2: averaging the signal intensities corresponding to a plurality of continuous time sequences according to the average value information of the absolute values of the muscle electric signals of the plurality of channels;

step S3: calculating an average value MAV of the interval for each channel action section, and taking the average value MAV of the interval as a feature vector to perform feature extraction on the signal to obtain feature extraction result information;

step S4: and (4) according to the feature extraction result information, dividing the muscle electric signals into 4 types.

2. The myoelectric sensing method with user adaptive capability according to claim 1, wherein the step S2 comprises:

step S2.1: if the average values of a plurality of signal points after the signals corresponding to the plurality of continuous time sequences all exceed a certain threshold value, the action is considered to be started, otherwise, if the local average values of a plurality of signal points after the signals corresponding to the plurality of continuous time sequences are all smaller than the threshold value, the action is considered to be ended.

3. The myoelectric sensing method with user adaptive capability according to claim 1, wherein the step S4 comprises:

step S4.1: and (4) according to the feature extraction result information, training an XGboost model, and dividing the muscle electric signals into 4 types.

4. The myoelectric sensing method with user adaptive capability according to claim 1, wherein the step S1 comprises:

step S1.1: collecting surface electromyographic signals of four different muscles of the forearm, summarizing the muscle electric signals of the four channels, solving the average value of the absolute values of the muscle electric signals of the four channels, and obtaining the average value information of the absolute values of the muscle electric signals of the four channels.

5. The myoelectric sensing method with user adaptive capability according to claim 1, wherein the step S2 comprises:

step S2.1: and averaging the signal intensities corresponding to the four continuous time sequences according to the absolute value average value information of the muscle electric signals of the four channels.

6. A myoelectric sensing system with user adaptive capability, comprising:

module M1: collecting surface electromyographic signals of a plurality of different muscles on the forearm, summarizing the multichannel muscle electric signals to obtain the multichannel muscle electric signal absolute value average value, and obtaining the multichannel muscle electric signal absolute value average value information;

module M2: averaging the signal intensities corresponding to a plurality of continuous time sequences according to the average value information of the absolute values of the muscle electric signals of the plurality of channels;

module M3: calculating an average value MAV of the interval for each channel action section, and taking the average value MAV of the interval as a feature vector to perform feature extraction on the signal to obtain feature extraction result information;

module M4: and (4) according to the feature extraction result information, dividing the muscle electric signals into 4 types.

7. The myoelectric sensing system with user adaptive capability according to claim 6, wherein said module M2 comprises:

module M2.1: if the average values of a plurality of signal points after the signals corresponding to the plurality of continuous time sequences all exceed a certain threshold value, the action is considered to be started, otherwise, if the local average values of a plurality of signal points after the signals corresponding to the plurality of continuous time sequences are all smaller than the threshold value, the action is considered to be ended.

8. The myoelectric sensing system with user adaptive capability according to claim 6, wherein said module M4 comprises:

module M4.1: and (4) according to the feature extraction result information, training an XGboost model, and dividing the muscle electric signals into 4 types.

9. The myoelectric sensing system with user adaptive capability according to claim 6, wherein said module M1 comprises:

module M1.1: collecting surface electromyographic signals of four different muscles of the forearm, summarizing the muscle electric signals of the four channels, solving the average value of the absolute values of the muscle electric signals of the four channels, and obtaining the average value information of the absolute values of the muscle electric signals of the four channels.

10. The myoelectric sensing system with user adaptive capability according to claim 6, wherein said module M2 comprises:

module M2.1: and averaging the signal intensities corresponding to the four continuous time sequences according to the absolute value average value information of the muscle electric signals of the four channels.

Technical Field

The invention relates to the technical field of interrupt guarantee, in particular to a myoelectric sensing method and system with user self-adaptive capacity.

Background

At present, there are many papers about myoelectricity sensing and electromechanical sensing algorithm frameworks of foreign open sources, but open source algorithms are more general and do not have user self-adaptive capability.

Patent document CN211723165U discloses an electromyographic sensor and a prosthetic device, the electromyographic sensor including: the fixing piece is used for being installed on the wearing piece, and a driving mechanism is arranged on the fixing piece; the sensor body is arranged on one side of the fixing piece and comprises a myoelectricity induction piece which is used for being in contact with a human body; the sensor body is connected with the driving mechanism, the driving mechanism is used for driving the sensor body to move towards the direction close to or far away from the fixing piece, and the sensor body is provided with an avoiding position close to the fixing piece and a sensing position far away from the fixing piece and used for being in contact with a human body. The utility model discloses among the technical scheme, realize the relative movement between mounting and the sensor body through actuating mechanism to traditional flesh electric sensor rigidity, the poor problem of flexibility have been solved.

Disclosure of Invention

Aiming at the defects in the prior art, the invention aims to provide a myoelectric sensing method and system with user self-adaptive capacity.

The myoelectric sensing method with the user self-adaptive capability provided by the invention comprises the following steps: step S1: collecting surface electromyographic signals of a plurality of different muscles on the forearm, summarizing the multichannel muscle electric signals to obtain the multichannel muscle electric signal absolute value average value, and obtaining the multichannel muscle electric signal absolute value average value information; step S2: averaging the signal intensities corresponding to a plurality of continuous time sequences according to the average value information of the absolute values of the muscle electric signals of the plurality of channels; step S3: calculating an average value MAV of an interval for each channel action segment, wherein the MAV values are obviously distinguished for different actions, and the average value MAV of the interval is used as a feature vector to perform feature extraction on the signal to obtain feature extraction result information; step S4: and (4) training an XGboost model by acquiring signals in the first three steps according to the feature extraction result information, and dividing the muscle electrical signals into 4 types. Through the field test of volunteers, the average identification accuracy of the method is up to 96.1%.

Preferably, the step S2 includes:

step S2.1: if the average values of a plurality of signal points after the signals corresponding to the plurality of continuous time sequences all exceed a certain threshold value, the action is considered to be started, otherwise, if the local average values of a plurality of signal points after the signals corresponding to the plurality of continuous time sequences are all smaller than the threshold value, the action is considered to be ended.

Preferably, the step S4 includes:

step S4.1: and (4) training an XGboost model by acquiring signals in the first three steps according to the feature extraction result information, and dividing the muscle electrical signals into 4 types. Through the field test of volunteers, the average identification accuracy of the method is up to 96.1%.

Preferably, the step S1 includes:

step S1.1: collecting surface electromyographic signals of four different muscles of the forearm, summarizing the muscle electric signals of the four channels, solving the average value of the absolute values of the muscle electric signals of the four channels, and obtaining the average value information of the absolute values of the muscle electric signals of the four channels.

Preferably, the step S2 includes:

step S2.1: and averaging the signal intensities corresponding to the four continuous time sequences according to the absolute value average value information of the muscle electric signals of the four channels.

The invention provides a myoelectric sensing system with user self-adaptive capacity, which comprises: module M1: collecting surface electromyographic signals of a plurality of different muscles on the forearm, summarizing the multichannel muscle electric signals to obtain the multichannel muscle electric signal absolute value average value, and obtaining the multichannel muscle electric signal absolute value average value information; module M2: averaging the signal intensities corresponding to a plurality of continuous time sequences according to the average value information of the absolute values of the muscle electric signals of the plurality of channels; module M3: calculating an average value MAV of an interval for each channel action segment, wherein the MAV values are obviously distinguished for different actions, and the average value MAV of the interval is used as a feature vector to perform feature extraction on the signal to obtain feature extraction result information; module M4: and (4) training an XGboost model by acquiring signals in the first three steps according to the feature extraction result information, and dividing the muscle electrical signals into 4 types. Through the field test of volunteers, the average identification accuracy of the system is up to 96.1%.

Preferably, said module M2 comprises:

module M2.1: if the average values of a plurality of signal points after the signals corresponding to the plurality of continuous time sequences all exceed a certain threshold value, the action is considered to be started, otherwise, if the local average values of a plurality of signal points after the signals corresponding to the plurality of continuous time sequences are all smaller than the threshold value, the action is considered to be ended.

Preferably, said module M4 comprises:

module M4.1: and (4) training an XGboost model by acquiring signals in the first three steps according to the feature extraction result information, and dividing the muscle electrical signals into 4 types. Through the field test of volunteers, the average identification accuracy of the system is up to 96.1%.

Preferably, said module M1 comprises:

module M1.1: collecting surface electromyographic signals of four different muscles of the forearm, summarizing the muscle electric signals of the four channels, solving the average value of the absolute values of the muscle electric signals of the four channels, and obtaining the average value information of the absolute values of the muscle electric signals of the four channels.

Preferably, said module M2 comprises:

module M2.1: and averaging the signal intensities corresponding to the four continuous time sequences according to the absolute value average value information of the muscle electric signals of the four channels.

Compared with the prior art, the invention has the following beneficial effects:

1. the invention converts the 8-channel electromyography algorithm in the open source algorithm into the 2-channel electromyography algorithm, so that the recognition rate is greatly improved.

2. The XGboost model can be trained, 10 ten thousand database models are collected by the model at present, the model is trained, a user self-adaptive system is originally created through the model, and the traditional algorithm needs to be adjusted and calibrated individually according to the actual muscle condition of each user. Therefore, the system greatly reduces the product debugging cost, improves the product production efficiency and enables the product to be produced in a standardized way;

3. the invention has reasonable flow structure and convenient use and can overcome the defects of the prior art.

Drawings

Other features, objects and advantages of the invention will become more apparent upon reading of the detailed description of non-limiting embodiments with reference to the following drawings:

FIG. 1 is a schematic overall flow chart of the present invention.

Fig. 2 is a diagram showing a sequence of action potentials of a motion unit in an embodiment of the present invention.

Fig. 3 is a schematic diagram of a general processing flow of sEMG-based motion recognition in an embodiment of the present invention.

Detailed Description

The present invention will be described in detail with reference to specific examples. The following examples will assist those skilled in the art in further understanding the invention, but are not intended to limit the invention in any way. It should be noted that variations and modifications can be made by persons skilled in the art without departing from the spirit of the invention. All falling within the scope of the present invention.

The myoelectric sensing method with the user self-adaptive capability provided by the invention comprises the following steps: step S1: collecting surface electromyographic signals of a plurality of different muscles on the forearm, summarizing the multichannel muscle electric signals to obtain the multichannel muscle electric signal absolute value average value, and obtaining the multichannel muscle electric signal absolute value average value information; step S2: averaging the signal intensities corresponding to a plurality of continuous time sequences according to the average value information of the absolute values of the muscle electric signals of the plurality of channels; step S3: calculating an average value MAV of an interval for each channel action segment, wherein the MAV values are obviously distinguished for different actions, and the average value MAV of the interval is used as a feature vector to perform feature extraction on the signal to obtain feature extraction result information; step S4: and (4) training an XGboost model by acquiring signals in the first three steps according to the feature extraction result information, and dividing the muscle electrical signals into 4 types. Through the field test of volunteers, the average identification accuracy of the method is up to 96.1%.

The step S2 includes:

step S2.1: if the average values of a plurality of signal points after the signals corresponding to the plurality of continuous time sequences all exceed a certain threshold value, the action is considered to be started, otherwise, if the local average values of a plurality of signal points after the signals corresponding to the plurality of continuous time sequences are all smaller than the threshold value, the action is considered to be ended.

The step S4 includes:

step S4.1: and (4) training an XGboost model by acquiring signals in the first three steps according to the feature extraction result information, and dividing the muscle electrical signals into 4 types. Through the field test of volunteers, the average identification accuracy of the method is up to 96.1%.

The step S1 includes:

step S1.1: collecting surface electromyographic signals of four different muscles of the forearm, summarizing the muscle electric signals of the four channels, solving the average value of the absolute values of the muscle electric signals of the four channels, and obtaining the average value information of the absolute values of the muscle electric signals of the four channels.

The step S2 includes:

step S2.1: and averaging the signal intensities corresponding to the four continuous time sequences according to the absolute value average value information of the muscle electric signals of the four channels.

The invention provides a myoelectric sensing system with user self-adaptive capacity, which comprises: module M1: collecting surface electromyographic signals of a plurality of different muscles on the forearm, summarizing the multichannel muscle electric signals to obtain the multichannel muscle electric signal absolute value average value, and obtaining the multichannel muscle electric signal absolute value average value information; module M2: averaging the signal intensities corresponding to a plurality of continuous time sequences according to the average value information of the absolute values of the muscle electric signals of the plurality of channels; module M3: calculating an average value MAV of an interval for each channel action segment, wherein the MAV values are obviously distinguished for different actions, and the average value MAV of the interval is used as a feature vector to perform feature extraction on the signal to obtain feature extraction result information; module M4: and (4) training an XGboost model by acquiring signals in the first three steps according to the feature extraction result information, and dividing the muscle electrical signals into 4 types. Through the field test of volunteers, the average identification accuracy of the system is up to 96.1%.

The module M2 includes:

module M2.1: if the average values of a plurality of signal points after the signals corresponding to the plurality of continuous time sequences all exceed a certain threshold value, the action is considered to be started, otherwise, if the local average values of a plurality of signal points after the signals corresponding to the plurality of continuous time sequences are all smaller than the threshold value, the action is considered to be ended.

The module M4 includes:

module M4.1: and (4) training an XGboost model by acquiring signals in the first three steps according to the feature extraction result information, and dividing the muscle electrical signals into 4 types. Through the field test of volunteers, the average identification accuracy of the system is up to 96.1%.

The module M1 includes:

module M1.1: collecting surface electromyographic signals of four different muscles of the forearm, summarizing the muscle electric signals of the four channels, solving the average value of the absolute values of the muscle electric signals of the four channels, and obtaining the average value information of the absolute values of the muscle electric signals of the four channels.

The module M2 includes:

module M2.1: and averaging the signal intensities corresponding to the four continuous time sequences according to the absolute value average value information of the muscle electric signals of the four channels.

The CNN convolutional neural network is adopted, the information of the electromyographic signals is used as input, the features are extracted and abstracted through convolution, and the recognition results are directly output to judge the arm actions of the user.

The invention converts the 8-channel electromyography algorithm in the open source algorithm into the 2-channel electromyography algorithm, so that the recognition rate is greatly improved. The XGboost model can be trained, 10 ten thousand database models are collected by the model at present, the model is trained, a user self-adaptive system is originally created through the model, and the traditional algorithm needs to be adjusted and calibrated individually according to the actual muscle condition of each user. Therefore, the system greatly reduces the product debugging cost, improves the product production efficiency and enables the product to be produced in a standardized way; the invention has reasonable flow structure and convenient use and can overcome the defects of the prior art.

Specifically, in one embodiment, as shown in FIGS. 2-3, the surface electromyography signal is the result of a combined superposition of the sequence of action potentials issued by multiple locomotor units, both temporally and spatially, presented at the skin surface.

In the description of the present application, it is to be understood that the terms "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", and the like indicate orientations or positional relationships based on those shown in the drawings, and are only for convenience in describing the present application and simplifying the description, but do not indicate or imply that the referred device or element must have a specific orientation, be constructed in a specific orientation, and be operated, and thus, should not be construed as limiting the present application.

The foregoing description of specific embodiments of the present invention has been presented. It is to be understood that the present invention is not limited to the specific embodiments described above, and that various changes or modifications within the scope of the appended claims can be made by those skilled in the art without affecting the essence of the present invention. The embodiments and features of the embodiments of the present application can be arbitrarily combined with each other without conflict.