363 results about "Motor imagery" patented technology

The invention provides a system and method for controlling a brain computer interface (BCI) based on multimode fusion. The system comprises a brain electrostimulation and feedback module, an electroencephalogram (EEG) signal acquisition module, an EEG signal processing module and an execution module, wherein the brain electrostimulation and feedback module is used for evoking an SSVEP (Steady State Visual Evoked Potential) and inducing an MI (Motor Imagery); the EEG signal acquisition module is used for acquiring EEG signals; the EEG signal processing module is used for extracting, identifying and classifying SSVEP characteristics and MI EEG characteristics in the EEG signals and feeding classified results which respectively correspond to the SSVEP characteristics and the MI EEG characteristics back to the brain electrostimulation and feedback module; and the execution module is used for executing the classified results. According to the system and the method, a multimode fusion BCI is constructed, so that the information transmission rate, reliability and flexibility of a control system are improved, the low information transmission rate of a BCI in a single MI mode is reduced, meanwhile, the visual burden under a single SSVEP task is reduced, and the adaptation crowds of the BIC-based control system are increased.

The invention belongs to the technical field of brain-computer interaction (BCI), and particularly relates to a method and device for classifying active brain-computer interaction system motor imagery tasks. The method and device for classifying active brain-computer interaction system motor imagery tasks resolve the technical problems that an existing classification mode for the active motor imagery tasks is not high in accuracy and long in execution time. The classification method for the active BCI system motor imagery tasks comprises the steps that (1) original brain electrical signals in the training period are collected; (2) processing and characteristic extraction are carried out on the brain electrical signals, and the detection threshold Vd is calculated; (3) collection and state detection are carried out on the brain electric signals in the active imagery period, and the idle state and the imagery motion state of the brain are detected according to the threshold; (4) the motor imagery brain electric characteristics are classified into left hand motor imagery tasks and right hand motor imagery tasks. The brain electric characteristic classification method which combines threshold detection and supporting of a support vector machine is stable and reliable, improves the accuracy of classifying the motor imagery brain electric characteristics, and effectively shortens the classification execution time.

The invention relates to an intelligent rehabilitation robot system for upper limbs in motor imagery and a training method thereof, and belongs to the field of motor imagery. The system comprises a motion visual stimulation module, an electroencephalogram acquisition module, an upper computer, and an upper limb rehabilitation movement bracket; the upper computer produces corresponding motion stimulation videos according to a rehabilitation training program, and displays the videos to a user through the motion visual stimulation module; the electroencephalogram acquisition module collects synchronous electroencephalogram signals of the user, carries out an amplification and filtering pretreatment, and transmits the electroencephalogram signals to the upper computer; after receiving the electroencephalogram signals, the upper computer generates rehabilitation training control signals through a feature extraction and pattern classification algorithm and outputs the rehabilitation training control signals to the upper limb rehabilitation movement bracket; the user fastens the limb through a fixing mechanism to the upper limb rehabilitation movement bracket, and the rehabilitation movement bracket receives the training control signals and conducts rehabilitation exercise training on the limb according to the corresponding control signals. According to the intelligent rehabilitation robot system for upper limbs in motor imagery and the training method thereof, the precision of rehabilitation training is improved, and the visual comfort of the user is improved.

The invention relates to a multi-class motor imagery brain electrical signal classification method based on phase synchronization. According to the method, firstly, phase synchronization features of a training sample and a test sample are calculated respectively through a phase locking value; secondly, correlation coefficients of the training sample and the test sample are calculated and arrayed from large to small after an average value is removed and an absolute value is obtained; thirdly, brain electrical signals are roughly classified according to the arrayed correlation coefficients, and then disaggregated classification is conducted according to the brain electrical signals which are roughly classified, wherein the process is involved in a shared airspace mode feature extraction method and a linear discriminant analysis and classification method. The method comprises the steps of brain electrical signal collection, data pre-processing, filtering, calculation of the correlation coefficients of the phase synchronization features, feature extraction and classification and classification accuracy calculation. Classification results show that by the adoption of the brain electrical signal classification method based on phase synchronization, the classification results are good, the rough class where the test sample belongs can be efficiently and accurately determined through brain electrical signal rough classification based on phase synchronization, and the calculated amount is reduced.

The invention discloses a brain computer interface mouse control-based Internet browsing method, which comprises the following steps of: initializing a system; controlling a mouse to move to a target area in a browser interface based on motor imagery and P300 electroencephalogram potentials; if a target is a text input target, switching the interface to a P300 text input interface, inputting characters by using P300 signal detection, returning to the browser interface after finishing inputting the characters and writing text input results into a selected text input box; and if the target is not the text input target, judging whether to click the target or not according to the motor imagery and P300 electroencephalogram potential information, clicking the target to open a new interface if determining to click the target, otherwise, controlling the mouse to continuously move in the interface. The brain computer interface mouse control-based Internet browsing method has the advantages of high control speed, high accuracy, good effect, working stability and capacity of realizing continuous control, and can be used under the condition of irremovability of a body or the computer mouse due to external or internal factors.

The invention discloses a two-dimensional cursor control method and a two-dimensional cursor control device for a brain-computer interface. The method comprises that: a user generates a scalp electroencephalogram signal according to a working interface command in a display device; an electrode cap acquires the scalp electroencephalogram signal, and the scalp electroencephalogram signal is converted by an analog-to-digital conversion module and amplified by a signal amplifier and is transmitted to a computer; a signal processing module in the computer respectively preprocesses the transmitted signal, extracts characteristics of the transmitted signal and sorts the transmitted signal; and a control module in the computer converts the classified information into a control command to control a cursor to move on the display device. The device comprises the electrode cap, the signal amplifier, the computer and the display device; and a preprocessing module, a characteristic extraction module and a sorting algorithm module are respectively arranged inside the computer according to P300 information and motor imagery ERD/ERS information. The two-dimensional cursor control method and the two-dimensional cursor control device have the advantages of high control accuracy, good effect, stable work and capacity of realizing continuous two-dimensional motion of the cursor, and can be applied to motion control of a computer mouse, a wheel chair, a mechanical arm and the like.

The invention relates to a method based on LSTM-. The electroencephalogram signal feature extraction and classification recognition method for the FC is characterized by comprising the following steps: S1, collecting and preprocessing electroencephalogram signals; S2, defining a basic model structure of the LSTM network; S3, enabling the feature matrix passing through the LSTM network to pass through a two-layer FC network to obtain an LSTM model fused with the FC network, namely, LSTM-; An FC model; S4, inputting the training set into a corresponding model for training, and updating the network by using error back propagation; S5, after the corresponding model is trained, inputting the test set into the model to obtain the final classification accuracy of the motor imagery task, and thenthe performance of the model is evaluated; S6: comparing performance of LSTM and LSTM-FC models to obtain the best model.

The invention relates to an apoplexy recovery degree index calculation method based on brain electrical signals. The method includes the following steps of firstly, enabling a preferred examinee to execute a left/right hand motor imagery task according to a prompt, utilizing multi-channel brain electrical signal collecting equipment to collect the brain electrical signals produced when the examinee executes the motor imagery task, then, utilizing a public average eliminating reference method to lower the level of public noise, utilizing an independent component analysis method to eliminate ocular artifacts, improving the signal to noise ratio of the brain electrical signals, utilizing a Butterworth filter to draw data of frequency ranges which are closely relevant to the motor imagery task for later analysis, finally, calculating a coefficient of polymerization and an overall efficiency property of a brain function network, calculating accuracy of the execution of the motor imagery task, utilizing the brain function network property and motor imagery task accuracy as recovery degree indexes, and storing calculation results in a special data base. The apoplexy recovery degree index calculation method based on the brain electrical signals has the advantages of being free of will interference and strong in objectivity. An objective assessing basis can be obtained through the method.

The invention provides a parallel convolutional neural network motor imagery electroencephalogram classification method based on spatial-temporal feature fusion. According to the invention, motion imagery electroencephalogram signals are used as research objects, and a novel deep network model-parallel convolutional neural network method is provided for extracting spatial-temporal features of themotion imagery electroencephalogram signals. Different from a traditional electroencephalogram classification algorithm which often discards electroencephalogram spatial feature information, Theta waves (4-8 Hz), alpha waves (8-12 Hz) and beta waves (12-36 Hz) are extracted through fast Fourier transform to generate a 2D electroencephalogram feature map; training of the electroencephalogram feature map is conducted based on a multi-convolutional neural network so as to extract spatial features; in addition, a time convolution neural network is used for parallel training so as to extract time-order features; and finally, the spatial features and the time-order features are fused and classified based on Softmax. Experimental results show that the parallel convolutional neural network has good recognition precision and is superior to other latest classification algorithms.

The invention provides an upper limb exoskeleton control method based on motor imagery. The method comprises following steps: 1), according to left arrow prompt or down arrow prompt in a display screen, a subject performs corresponding motor imagery, namely, left hand motor imagery and foot motor imagery, and produced original electroencephalogram signals are pre-processed; 2), the pre-processed original electroencephalogram signals are input into a CNN structure and subjected to feature extraction and classification, and an identification result is output; 3), a driver control signal is acquired according to the output identification result, upper limb exoskeleton is driven and front arms are driven to correspondingly move; if the classification result indicates the left hand motor imagery, the upper limb exoskeleton performs a stretching action, and if the classification result indicates the foot motor imagery, the upper limb exoskeleton performs a bending action. When the CNN structure is designed, a convolution kernel is set as a vector according to the characteristic of time and space feature combination of the electroencephalogram signals. The invention further provides an upper limb exoskeleton control system based on the motor imagery. The classification and identification rate is higher, and experience effect is good.

The invention discloses a motor imagery identification method and system fusing a CNN-BiLSTM model and probability cooperation, belonging to the field of man-machine interfaces. The invention aims tosolve the technical problem of how to effectively extract features and how to quickly realize classification, and the method adopts the technical scheme that (1) the method captures and extracts time-space depth features in an EEG signal by merging a CNN model of a BiLSTM network, and inputs captured and extracted time-space depth features into a ProCRC classifier for classification, and uses testset data pairs to construct the CNN-BiLSTM model for performance evaluation to achieve intent recognition. The method comprises the following specific steps: S1, acquiring an electroencephalogram signal; S2, constructing a deep neural network; S3, establishing a classifier; and S4, testing and evaluating the model. The system comprises an electroencephalogram signal acquisition unit, a deep neural network construction unit, a classifier construction unit and a model test and evaluation unit.

The invention relates to an EEMD (Ensemble Empirical Mode Decomposition) and wavelet threshold based motor imagery electroencephalogram signal denoising method. The method comprises the steps of firstly, performing EEMD on an original signal to obtain a series of IMF (Intrinsic Mode Function) components; secondly, improving a conventional wavelet threshold method with a new threshold function and a threshold selection method; thirdly, processing a high-frequency IMF component with the improved wavelet threshold method; and finally, reconstructing the processed IMF component and other IMF components to obtain a denoised motor imagery EEG (electroencephalogram) signal. The method has the advantages that effective information in the high-frequency component is reserved, the suppression of the wavelet threshold method to a weak-energy effective signal is reduced, most of useful detailed information is reserved while a large amount of noises are eliminated, and a good foundation is laid for motor imagery EEG signal feature extraction and mode identification in the next step.

The invention relates to a motor imagery electroencephalogram data classification method which comprises the following steps: acquiring motor imagery electroencephalogram signals generated when a subject executes a motor imagery task to obtain original electroencephalogram data; performing data preprocessing on the original electroencephalogram data to obtain preprocessed data; dividing the preprocessed data into a training set and a test set; establishing an inception deep neural network model, and inputting the training set into the inception deep neural network model for training to obtaina trained model; and testing the trained model by utilizing the verification set, judging whether the trained model is optimal on the verification set or not, and if so, inputting the data of the testset into the optimal model to obtain a classification result. According to the method, the inception module and the convolutional neural network are combined, the inception deep neural network modelis established, calculation resources can be efficiently utilized, the calculation complexity is reduced, and the classification effect of the neural network is improved.

The invention discloses a common spatial pattern and deep learning method based on brain-computer interface auxiliary rehabilitation therapy. The method comprises the following steps that: S1: preprocessing an electroencephalogram signal to obtain the electroencephalogram signal of which the noise is filtered; S2: through an improved OVR-CSP (One Versus Rest- Common Spatial Pattern) algorithm, carrying out feature extraction on various categories of motor imagery EEG (electroencephalogram) signals of which the noise is filtered to obtain the feature of each category of motor imagery electroencephalogram signal, forming one-dimensional feature data, and meanwhile, taking a variance as the input of a classifier; and S3: utilizing a transformed CNN (Convolutional Neural Network) which is suitable for a one-dimensional input sample to carry out secondary feature extraction and classification. When the technical scheme of the invention is adopted, the movement position and the limb movementstate of a patient can be more accurately judged so as to provide an objective data support for realizing the scale evaluation of the rehabilitation degree of a patient.

The invention provides a DWT- and Parametric t-SNE-based characteristic extracting method of motor imagery EEG(Electroencephalogram) signals. First, effective time and frequency ranges of EEG characteristics are determined by using a Wigner-Ville distribution and power spectrum; the EEG signals in a specific time and frequency segment is subjected to three-layer discrete wavelet decomposition and statistical characteristic quantity including the average value, the energy average value, the mean square error and the like are calculated and are taken as the time frequency characteristic of the EEG signals; at the same time, a parameterization t-SNE algorithm is utilized for performing non-linear characteristic mapping on said wavelet coefficients and embedded coordinates corresponding to a low-dimensional space are taken as the non-linear characteristic; the two characteristics are standardized and a characteristic vector including both the time frequency information and the non-linear information of the EEG signals in the specific time frequency segment is obtained. According to the invention, EEG characteristics of compactness and completeness are obtained and a method for solving a problem of poor generalization performance of a traditional manifold learning algorithm in pattern classification application through fitting a multilayer forward propagation neural network to nonlinear mapping is proposed, so that accuracy of pattern classification of MI-EEG signals is improved further.

The invention discloses a two-dimensional cursor movement control system and method based on motor imagery and coded modulation visual evoked potential. The system comprises an electrode cap, an electroencephalography acquisition instrument, a system control unit, a data processing module, a cursor control module and a visual stimulator, wherein the visual stimulator is provided for a user in an interface display way. The method comprises the following steps that: the user simultaneously executes a motor imagery task and a visual attention task according to a work interface instruction; the electrode cap collects an electroencephalography signal; the electroencephalography acquisition instrument carries out amplification, filtering and analog-digital conversion on the electroencephalography signal; the system control unit separates the electroencephalography data generated by the motor imagery and the visual attention, and then, the electroencephalography data is stored; the data processing module carries out preprocessing, feature extraction and classification identification on two classes of electroencephalography data in sequence; and the cursor control module controls a cursor to carry out continuous two-dimensional movement according to classification identification. The system has the advantages of being high in control accuracy, good in robustness and the like, the continuous movement of a two-dimensional cursor can be realized, and the system can be used for the movement control of a computer mouse.

The invention discloses a motor imagery EEG pattern recognition method based on the time-frequency parameter optimization of an artificial bee colony. The method comprises the steps of conducting the leads selection based on the linear decision rule, selecting time-domain and frequency-domain optimal parameters based on the artificial bee colony algorithm, extracting features based on the common spacial pattern algorithm, and finally classifying features based on the linear discriminant analysis algorithm. The result of the method shows that, a lead channel of larger inter-class distinction degree can be effectively selected based on the lead selection algorithm. At the same time, based on the time-frequency parameter optimization algorithm of the artificial bee colony, a time window and a frequency band of larger inter-class distinction degree can be automatically selected, so that a better classification effect is obtained. The method is capable of effectively recognizing different motor imagery modes. Compared with the traditional parameter manual selection method and the frequency-domain parameter automatic selection algorithm, global optimal parameters can be automatically searched in both time domain and frequency domain at the same time based on the above method. Therefore, the feature extraction and feature classification effect for motor imagery EEG signals is improved.

The invention discloses a motor imagery electroencephalogram signal processing method and a device, and a storage medium. The method comprises the following steps: respectively inputting a source MI electroencephalogram signal belonging to a source domain and a target MI electroencephalogram signal belonging to a target domain into an initial feature extraction model to obtain a first source MI feature and a first target MI feature; inputting the first source MI feature into an initial classification model to obtain a first classification result output by the initial classification model, thefirst classification result is used for representing an action expected to be executed by the source MI electroencephalogram signal; under the condition that a first known action corresponding to thesource MI electroencephalogram signal is inconsistent with an action expected to be executed by the source MI electroencephalogram signal represented by the first classification result or the similarity between the feature distribution of the first source MI feature and the feature distribution of the first target MI feature is smaller than a first preset threshold value, determining that the first target MI feature is not consistent with the first target MI feature; and adjusting model parameters of the initial feature extraction model and/or model parameters of the initial classification model to obtain a target feature extraction model and a target classification model.

The invention relates to a dual-tree complex wavelet and common spatial pattern combined electroencephalogram characteristic extraction method. The method comprises the steps that electroencephalogram signals with suitable channels are selected at first, up-sampling or down-sampling is conducted on the original frequency according to the characteristics of dual-tree complex wavelet frequency partitions, frequency bands corresponding to the frequency ranges of delta, theta, alpha and beta rhythm waves are obtained by means of dual-tree complex wavelet multi-scale decomposition, signal reconstruction is conducted in the scales, a plurality of reconstruction signals in the corresponding frequency bands are obtained, the same decomposition and reconstruction are conducted on the suitable channels, reconstruction signals of all the frequency bands of all the channels are combined to be input into a spatial filter, a six-dimensional feature vector is obtained, and motor imagery task classification is accomplished by means of a support vector machine at last. The dual-tree complex wavelet and common spatial pattern combined electroencephalogram characteristic extraction method not only conducts frequency information analysis on motor imagery electroencephalogram signals, but also can effectively overcome electrode selection defects.

The invention belongs to the field of a brain machine connector and provides a vague discrete event shared control method of a brain-controlled robotic system. The vague discrete event shared control method of the brain-controlled robotic system utilizes a method of combing a human brain control command and robot autonomous control based on the vague discrete event system to distinguish a motor imagery brain electroencephalogram online and take the motor imagery brain electroencephalogram as a control command with the highest priority degree to control a robot to advance, turn left and turn right. When no brain control command exists, the autonomous control module which is based on the vague discrete event system is operated to blur autonomous control states of obstacle avoidance and traveling close to a wall of the robot and forms the vague discrete event system aiming at the size of a barrier in a route, the length of distance and the like. The vague discrete event shared control method of the brain controlled robotic system makes up the problems that information transmission speed of the brain machine connector is low, distinguish error rate is high, control is delayed and the like through the shared control method and strengthens the adaptive capacity of the robot in a complex environment.