CN-122004899-A - Task matching visual touch feedback exercise training device and method

CN122004899ACN 122004899 ACN122004899 ACN 122004899ACN-122004899-A

Abstract

The invention provides a task matching visual touch feedback exercise training device and a task matching visual touch feedback exercise training method, which are applied to the technical field of biological signal processing. The device comprises acquisition equipment, electric stimulation equipment, a processor, an updating coupling matrix, a vector machine and an electric stimulation equipment, wherein the electric stimulation equipment is used for applying first electric stimulation to a target object, applying second electric stimulation to the target object according to a second electric stimulation instruction, the processor is used for constructing and obtaining a training coupling matrix according to training brain electricity feature vectors in training brain electricity signals and training myoelectricity feature vectors in training myoelectricity signals, updating the training coupling matrix by using updating weights and the updating coupling matrix corresponding to historical training to obtain the updating coupling matrix, performing feature extraction fusion processing on the training brain electricity feature vectors and the training myoelectricity feature vectors based on a particle swarm fitness function and the updating coupling matrix to obtain target brain myoelectricity fusion features, performing feature classification recognition on the target brain electricity fusion features by using the vector machine, and applying the second electric stimulation instruction to the electric stimulation equipment according to training results.

Inventors

ZHOU YIJIE
Chu Zhuoyue
ZHANG MINGMING
CHENG SHENGCUI
WANG ZHONGPENG
MING DONG

Assignees

天津大学

Dates

Publication Date: 20260512
Application Date: 20260414

Claims (10)

1. A task matching visual touch feedback exercise training device, comprising: the acquisition equipment is arranged at the head and the hand of the target object and is used for acquiring the brain electrical signals and the muscle electrical signals of the target object; the electric stimulation equipment is attached to the hand of the target object and is used for applying first electric stimulation to the target object before the exercise training so as to prompt the exercise training; applying a second electrical stimulation to the target object according to a second electrical stimulation instruction to prompt a training result of the exercise training of the target object, wherein the intensities of the first electrical stimulation and the second electrical stimulation are different; and the processor is electrically connected with the acquisition device and the electric stimulation device and is used for: Performing exercise training in response to the target object, taking the electroencephalogram signal received from the acquisition equipment as a training electroencephalogram signal, taking the electromyogram signal received from the acquisition equipment as a training electromyogram signal, and constructing and obtaining a training coupling matrix according to a training electroencephalogram characteristic vector in the training electroencephalogram signal and a training electromyogram characteristic vector in the training electromyogram signal; Updating the training coupling matrix by using the updating weight corresponding to the historical training and the updating coupling matrix corresponding to the historical training to obtain an updating coupling matrix corresponding to the sports training; based on a particle swarm fitness function and an updated coupling matrix corresponding to motion training, performing feature extraction fusion processing on the training electroencephalogram feature vector and the training myoelectric feature vector to obtain target electroencephalogram fusion features; And carrying out feature classification and identification on the target brain myoelectricity fusion features by using a vector machine to obtain a training result, and applying the second electric stimulation instruction to the electric stimulation equipment according to the training result.
2. The apparatus of claim 1, wherein the processor is further configured to: Performing motion training in response to the target object, and performing synchronous energy feature extraction on the acquired training electroencephalogram signal based on event-related spectrum disturbance to obtain a training electroencephalogram feature vector; Based on a preset myoelectric channel, calculating an integral myoelectric value of the acquired training myoelectric signal to obtain a training myoelectric characteristic vector; And constructing the training coupling matrix according to the training electroencephalogram feature vector and the training myoelectricity feature vector.
3. The apparatus of claim 1, wherein the processor is further configured to: under the condition that the target object performs the nth exercise training in the N times of exercise training, according to the training result of the nth-1 exercise training serving as the historical training and the second electric stimulation instruction, determining an nth-1 updated weight corresponding to the nth-1 exercise training, wherein the updated weight is one of a first weight corresponding to the successful representation training of the training result and a second weight corresponding to the failed representation training of the training result, and N is more than 1, and N is more than or equal to 2; And updating the training coupling matrix of the nth exercise training by using the (n-1) th updating weight and the (n-1) th updating coupling matrix corresponding to the (n-1) th exercise training to obtain the (n) th updating coupling matrix corresponding to the nth exercise training.
4. The apparatus of claim 3, wherein the processor is further configured to: Updating the initial particle swarm velocity function according to the nth updated coupling matrix to obtain a bottom particle swarm velocity function corresponding to the nth motion training; Based on the bottom layer subgroup fitness function, respectively carrying out iterative updating on the electroencephalogram weight and the myoelectric weight corresponding to the nth exercise training by using a bottom layer particle group velocity function corresponding to the nth exercise training to obtain a target electroencephalogram weight and a target myoelectric weight; Obtaining a target electroencephalogram fusion characteristic corresponding to the nth exercise training according to the target electroencephalogram weight and a training electroencephalogram characteristic vector corresponding to the nth exercise training, and obtaining a target electromyogram fusion characteristic corresponding to the nth exercise training according to the target electromyogram weight and a training electromyogram characteristic vector corresponding to the nth exercise training; And based on a top-level subgroup fitness function, carrying out fusion processing on the target electroencephalogram fusion characteristic and the target myoelectric fusion characteristic by using a modal fusion coefficient to obtain the target electroencephalogram fusion characteristic.
5. The apparatus of claim 4, wherein the processor is further configured to: in the case of M rounds of iterative updates of the brain electrical weights, Based on a bottom particle group velocity function corresponding to the nth motion training, determining the mth update speed updated by the mth iteration according to the mth-1 update speed updated by the mth-1 iteration, the training brain electrical characteristic vector and the speed coefficient, wherein m is more than 1, and M is more than or equal to 2; processing the electroencephalogram weight updated with the m-1 th iteration by utilizing the m-th updating speed to obtain an m-th electroencephalogram weight; based on the bottom layer subgroup fitness function, obtaining an mth fitness electroencephalogram resolution value according to the mth electroencephalogram weight and the training electroencephalogram feature vector; And under the condition that the m-th adaptability degree electroencephalogram resolution value is larger than an electroencephalogram threshold value, confirming that the electroencephalogram fusion characteristic obtained according to the m-th electroencephalogram weight and the training electroencephalogram characteristic vector is the target electroencephalogram fusion characteristic, and under the condition that the m-th adaptability degree electroencephalogram resolution value is smaller than or equal to the electroencephalogram threshold value, repeatedly executing the iterative updating operation on the electroencephalogram weight until the target electroencephalogram fusion characteristic is obtained.
6. The apparatus of claim 4, wherein the processor is further configured to: In the case of P-round fusion of the target electroencephalogram fusion feature and the target myoelectric fusion feature, Obtaining a P-th electroencephalogram fusion characteristic according to the target electroencephalogram fusion characteristic, the target myoelectric fusion characteristic and a P-th modal fusion coefficient, wherein P is more than or equal to 1; based on the top-level subgroup fitness function, obtaining a p-th fitness fusion resolution value according to the p-th brain myoelectricity fusion characteristic and the fusion coefficient; And under the condition that the p-th fitness fusion resolution value is larger than a fusion threshold value, confirming that the p-th brain myoelectricity fusion feature is the target brain myoelectricity fusion feature, and under the condition that the p-th fitness fusion resolution value is smaller than or equal to the fusion threshold value, repeating the operation of fusing the fusion feature until the target brain myoelectricity fusion feature is obtained.
7. The apparatus of claim 3, wherein the processor is further configured to: Before the target object performs the first motion training, based on event-related spectrum disturbance, respectively performing synchronous energy feature extraction on a plurality of acquired adaptive training electroencephalograms to obtain a plurality of adaptive electroencephalogram feature vectors, wherein the plurality of adaptive training electroencephalograms are acquired under the condition that the target object performs multiple groups of motion adaptive training before the motion training; Based on a preset myoelectric channel, respectively calculating integral myoelectric values of a plurality of acquired adaptive training myoelectric signals to obtain a plurality of adaptive myoelectric feature vectors, wherein the adaptive training myoelectric signals are acquired under the condition that the target object performs multi-group motion adaptive training before motion training; determining a target adaptive electroencephalogram feature vector and a target adaptive myoelectricity feature vector from the plurality of adaptive electroencephalogram feature vectors and the plurality of adaptive myoelectricity feature vectors according to a preset feature screening rule; And constructing an initial coupling matrix according to the target adaptation electroencephalogram characteristic vector and the target adaptation myoelectric characteristic vector, and taking the initial coupling matrix as a first updated coupling matrix corresponding to the first motion training.
8. The apparatus of claim 1, wherein the processor is further configured to: Based on the similarity function, obtaining a brain myoelectricity similarity value according to the target brain myoelectricity fusion characteristic and the ideal brain myoelectricity fusion characteristic corresponding to the exercise training; Based on a preset recognition rule, optimizing a fusion coefficient according to the brain myoelectricity similarity value and the particle swarm to obtain a motion training decision score; And determining the motion completion degree corresponding to the motion training according to the motion training decision score and a preset decision threshold, and generating the training result.
9. The apparatus of claim 1, wherein the apparatus further comprises: and the display screen is electrically connected with the processor and is used for displaying the motion task and the motion completion degree corresponding to the motion training to the target object.
10. A task matching visual touch feedback exercise training method applied to the task matching visual touch feedback exercise rehabilitation training device as claimed in any one of claims 1 to 9, wherein the method comprises the following steps: Performing exercise training in response to the target object, taking the electroencephalogram signal received from the acquisition equipment as a training electroencephalogram signal, taking the electromyogram signal received from the acquisition equipment as a training electromyogram signal, and constructing and obtaining a training coupling matrix according to a training electroencephalogram characteristic vector in the training electroencephalogram signal and a training electromyogram characteristic vector in the training electromyogram signal; Updating the training coupling matrix by using the updating weight corresponding to the historical training and the updating coupling matrix corresponding to the historical training to obtain an updating coupling matrix corresponding to the sports training; based on a particle swarm fitness function and an updated coupling matrix corresponding to motion training, performing feature extraction fusion processing on the training electroencephalogram feature vector and the training myoelectric feature vector to obtain target electroencephalogram fusion features; And carrying out feature classification and identification on the target brain myoelectricity fusion features by using a vector machine to obtain a training result, and applying the second electric stimulation instruction to the electric stimulation equipment according to the training result.

Description

Task matching visual touch feedback exercise training device and method Technical Field The invention relates to the technical field of biological signal processing, in particular to a task matching visual touch feedback exercise training device and method. Background Due to the related auxiliary training technology based on brain-computer interfaces (Brain Computer Interface, BCI), the close interactive connection between the brain movement intention information and external movement auxiliary training equipment can be built in an auxiliary manner by fully playing the subjective motility of a training object, so that the synchronous activation effect of cortex and muscle is realized, and one of the development directions focused on currently is realized. In the process of realizing the inventive concept, researches show that the prior art only depends on the traditional motor imagination paradigm and a single perception feedback path, so that the initiative of a training object is lower, the fusion perception degree is poorer, the coupling coordination between myoelectricity and electroencephalogram under self-regulation is difficult to realize, and meanwhile, the technical problem that the training effect is more common is caused by the lack of multi-sense feedback and multi-mode interaction training paradigm based on real natural task guidance in the training process. Disclosure of Invention In view of the above problems, the invention provides a task matching visual touch feedback exercise training device and a task matching visual touch feedback exercise training method. According to the first aspect of the invention, a task matching visual touch feedback motion training device is provided, which comprises acquisition equipment, electrical stimulation equipment, a processor and a fusion device, wherein the acquisition equipment is arranged at the head and the hand of a target object and is used for acquiring brain electrical signals and electromyographic signals of the target object, the electrical stimulation equipment is attached to the hand of the target object and is used for applying first electrical stimulation to prompt motion training to the target object before motion training, the second electrical stimulation is applied to the target object according to a second electrical stimulation instruction so as to prompt a training result of the motion training of the target object, the processor is electrically connected with the acquisition equipment and the electrical stimulation equipment, is used for responding to the target object to perform motion training, the brain electrical signals received from the acquisition equipment are used as training brain electrical signals, the electromyographic signals are used as training electromyographic signals, a training electromyographic feature vector in the training brain electrical signals is used for constructing and a training electromyographic feature vector in the training to obtain a training coupling matrix, the training coupling matrix is updated by using an update weight corresponding to the history training and an update coupling matrix corresponding to the history training, the training coupling matrix is updated to obtain a training result corresponding to the motion training object, the electrical feature vector is fused with the training feature vector, the electrical feature vector is obtained by the fusion device, and the electrical stimulation device is fused with the training feature electrical feature vector is used to obtain the training feature fusion device. The second aspect of the invention provides a task matching visual touch feedback exercise training method, which comprises the steps of performing exercise training in response to a target object, taking an electroencephalogram signal received from acquisition equipment as a training electroencephalogram signal, taking an electromyogram signal received from the acquisition equipment as a training electromyogram signal, constructing a training coupling matrix according to a training electromyogram characteristic vector in the training electroencephalogram signal and a training electromyogram characteristic vector in the training electromyogram signal, updating the training coupling matrix by using an updating weight corresponding to historical training and an updating coupling matrix corresponding to the historical training to obtain an updating coupling matrix corresponding to the exercise training, performing characteristic extraction fusion processing on the training electromyogram characteristic vector and the training electromyogram characteristic vector based on a particle swarm fitness function and the updating coupling matrix corresponding to the exercise training to obtain a target brain electromyogram fusion characteristic, performing characteristic classification recognition on the target brain electromyogram fusion characteristic by using a vector machine to obtain