CN101711709A - Method for controlling electrically powered artificial hands by utilizing electro-coulogram and electroencephalogram information - Google Patents
Method for controlling electrically powered artificial hands by utilizing electro-coulogram and electroencephalogram information Download PDFInfo
- Publication number
- CN101711709A CN101711709A CN200910154966A CN200910154966A CN101711709A CN 101711709 A CN101711709 A CN 101711709A CN 200910154966 A CN200910154966 A CN 200910154966A CN 200910154966 A CN200910154966 A CN 200910154966A CN 101711709 A CN101711709 A CN 101711709A
- Authority
- CN
- China
- Prior art keywords
- signal
- eeg signals
- reference signal
- hand
- formula
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Abstract
The invention relates to a method for controlling electrically powered artificial hands by utilizing electro-coulogram and electroencephalogram information. The traditional electrically powered hand control methods are not suitable for nerve paralysis people and paralytic people with seriously degenerated muscles. In the method, a scalp pick-up electrode in an electroencephalogram pick-up sensor is placed at the Fp1 or the Fp2 position of the forehead cerebelli anterior, determined by the international electroencephalogram association standard 10-20 lead system, and a reference electrode is placed at the pinna position; an original signal enters a micro processor after being processed; a determined component analysis method based on partical swarm optimization is applied to build the reference signal, extract an electroencephalogram signal containing electro-coulogram information and identify a hand movement pattern, and the micro processor outputs a corresponding control signal according to the identified result to control the electrically powered artificial hands to move. The method adopts an eye and brain coordination mode to express the hand movement consciousness and utilizes the useful information contained in an electro-coulogram signal to enhance the features of the electroencephalogram signal generated by the same movement consciousness; the identification correct rate of the hand movement pattern is high, and the electrically powered hand control is reliable.
Description
Technical field
The invention belongs to information and control technology field, relate to a kind of useful information that utilizes in the electro-ocular signal to be comprised, strengthen motion imagination EEG signals feature, improve under the motion imagination pattern technology based on the limb action recognition effect of EEG signals, it specifically is a kind of control information source that utilizes eye electricity and EEG signals as electrically powered hand, by the EEG signals that contains an electrical information is extracted, is analyzed, identification obtains a plurality of patterns of action of doing evil through another person, and then realizes doing evil through another person multivariant real-time control.
Background technology
(electro-oculogram EOG) is a kind of bioelectrical signals of following ocular movement to produce to electro-ocular signal.Eyeball is an ambipolar spheroid, and the cornea district presents positive polarity, and retinal field presents negative polarity, and potential difference between the two forms an electric field in the front portion of head.When eyeball rotated, this electric field space orientation changed.Near the electrode that electrical variation can be placed the eyes comprises that the scalp electrode for encephalograms detects.At the electrode detection of eye electricity to voltage signal be exactly electro-ocular signal.Eyeball all can be by corresponding electro-ocular signal record in the motion of vertical direction or horizontal direction.Action nictation can cause the variation of enclosing current potential near the eyes, thereby also can produce corresponding electro-ocular signal.
(electroencephalogram is the potential change that is caused by cerebral cortex neural cell group synapse transmission signal EEG) to EEG signals, can reflect the conscious activity that brain is autonomous or bring out, and is closely related with the action behavior of reality.At present, adopting EEG signals is brain-machine interaction (braincomputer interface as the control information source of artificial limb, BCI) field hot research problem, promptly carry out analyzing and processing when carrying out the limb motion imagination by the EEG signals that the people with disability is sent, identify the limb motion pattern that the people with disability will realize, produce the corresponding control signal that drives the artificial limb action in view of the above.
The collection of EEG signals is had two kinds of intrusive mood (invasive) and non-intrusion types (non-invasive), and the intrusive mood eeg signal acquisition uses implanted electrode, and non-intrusion type then adopts scalp electrode.Intrusive mood eeg signal acquisition method is by microsurgery microelectrode to be implanted a kind of method of measuring EEG signals in the cranial cavity, and polarization is good, the signal to noise ratio height, but have a series of technical difficulty and bigger clinical risk.At present existing achievement in research proof adopts the feasibility of scalp EEG signals as artificial limb control information source, but because such signal is not directly to extract in brain neuron, but extract from scalp, weak output signal, and noise is big, signal to noise ratio is very low, uses it and carries out very difficulty of limb motion imagination pattern recognition.
In scalp eeg signal acquisition process, by the electro-ocular signal that eye movement causes, a part can be propagated along skull, produces with EEG signals and merges, and the EEG signals feature is changed.In traditional signal processing, electro-ocular signal is used as the in addition filtering of main interfering signal usually, so be known as " the electric artefact of eye ".But the human most information that obtained from surrounding come from vision, and retina receives external information, passes to nerve centre through pathways for vision, and outside stimulus causes the thinking activities of brain.It should be noted that especially people in the process of carrying out hand exercise, often need the cooperation of brain, eye, hands, can naturally and understandably follow some eye motions.When for example people wanted to grasp a certain object, eyes also can turn to this object naturally.The electro-ocular signal that eye movement this moment causes has comprised the movable information of hands, and natural fusion takes place at the skull place EEG signals that this electro-ocular signal and same motion imagination process produce.The hand exercise information that electro-ocular signal comprises, the hand exercise information that EEG signals is comprised is enhanced, and signal characteristic is strengthened.
Learn that from above analysis the motion of eyes and the activity of human brain have confidential relation, thereby eye is electric and EEG signals is closely related on physiology.Has very high research value although get rid of the brain electricity analytical of the electric artefact of eye, but in some concrete practical application, the brain-computer interactive system that electric control is done evil through another person as brain, there is no need painstakingly to pursue with the filtering of eye electricity artefact, even the eye electrical information can be used, the limb motion imagination mode that adopts eye, brain to coordinate utilizes electro-ocular signal that motion imagination EEG signals feature is strengthened, in the hope of obtaining the better recognition effect.
(Independent Component Analysis ICA) as a kind of information source decomposition technique of rising in recent years, has obtained using comparatively widely in fields such as EEG signals place, speech recognition, communication, Flame Image Process in the independent component analysis.The basic ideas of ICA algorithm are under hybrid matrix and source signal condition of unknown, utilize independent this hypothesis of statistics between source signal, by optimizing matrix of a linear transformation of algorithm searching observation signal is carried out linear transformation, obtain output vector, make it to approach as much as possible source signal, become a estimation source signal.The independent component parser is a lot, but most methods need isolating independent component number identical with detected mixed signal number, be that signal composition or the various interfering signal composition that needs all will be extracted out in the mixed signal, and the signal composition put in order uncertain.Extract required signal composition, also need a large amount of postposition to handle.So traditional independent component analytical method contains in this class higher-dimension application of an electrical information EEG signals in extraction, will comprise a large amount of redundant computation, expend ample resources, reduced the quality that signal recovers, far can not satisfy the requirement that electrically powered hand is handled in real time.
Existing scholar proposes the priori of required extraction signal is used for the independent component parser, to improve the accuracy of algorithm, realizes this algorithm more applications.But such algorithm only utilizes reference signal and minimum this class second order criterion of estimated signal mean square deviation to determine to optimize the weight matrix component at present; And adopt non-protruding, non-linear objective functions such as algorithm optimization negentropy based on gradient, mutual information, sheath degree.In the higher-dimension optimization problem of reality, it is accurate to determine to optimize the weight matrix short weight based on the second order criterion; Gradient descent algorithm makes the optimization result be very easy to converge to a local optimum point, causes this type of algorithm only can obtain some locally optimal solutions, and the signal element that needs to extract may be left in the basket.
(particle swarm optimization is a kind of parallel, global optimization approach at random PSO) to particle swarm optimization algorithm, is a kind of general utility tool that solves optimization problem.This method utilizes probabilistic search to replace exhaustive search; Utilize a plurality of particle parallel computations to reduce computation time; The information that utilization feeds back is constantly revised the direction and the amplitude of next step search of particle, approaches global optimum's point gradually, thereby reaches the target of global optimization.Character such as this algorithm does not require that optimised function has can be little, can lead, continuous, fast convergence rate, the simple programming easily of algorithm realizes.The application of in computer science and engineering problem, having succeeded of PSO algorithm, from traditional problems such as the optimization of initial complicated multimodal nonlinear function, multiple-objection optimizations, to image segmentation, pattern recognition, neural network weight training, robot path ruleization or the like in real time.Particularly there is document to show that (magnetoencephalography, the MEG) application of succeeding in the data analysis shows the application prospect of PSO algorithm in the EEG Processing field to the PSO algorithm in magneticencephalogram in the recent period.
By above analysis as can be known, if can will utilize the priori that needs to extract signal to carry out the signal extraction problem that independent component is analyzed, be converted into the optimization problem that suitable PSO algorithm is found the solution, and then carry out optimizing with the PSO algorithm and calculate, what can solve such ICA algorithm existence at present only can obtain some locally optimal solutions, the uncared-for problem of signal element possibility that causes needs to extract is guaranteed the high efficiency of algorithm and accuracy.
Summary of the invention
Purpose of the present invention is exactly at the deficiencies in the prior art, a kind of 3-degree-of-freedom electrical prosthetic hand control method that adopts eye electricity, EEG signals as the control information source is provided, it is finished open, close up, wrist is stretched, wrist is bent, wrist outward turning, these six kinds actions of doing evil through another person of wrist inward turning.This kind the bionical paralysis personage who is particularly suitable for nerve, muscle serious degradation that does evil through another person use.
Can specifically be expressed as: in eeg signal acquisition and processing procedure, electro-ocular signal regarded as a kind of available signal source, the corresponding eye motion that when doing hand motion, is aided with naturally with reference to people in the daily life, adopt the eye brain coordination mode imagination of moving, the useful information that utilizes in the electro-ocular signal to be comprised strengthens same sports consciousness and produces the EEG signals feature; (determination component analysis based on particle swarmoptimization algorithm DCA-PSO) contains an electrical information EEG signals extraction and the pattern recognition of the hand exercise imagination to utilize the definite analysis of components algorithm based on particle group optimizing that is proposed; And then the control signal that obtains doing evil through another person is exported to drive circuit, the control that drive motors drive electrically powered hand is done evil through another person to Three Degree Of Freedom by people with disability's motion wish realization.
Concrete control method of the present invention is that the scalp power-collecting electrode that the brain electricity picks up in the electric transducer is placed on the brain frons F that international electroencephalography meeting standard 10-20 lead system is determined
P1Or F
P2Position, reference electrode are placed on the auricle position.Pick up the primary signal that electric transducer picks up by the brain electricity and convert data acquisition (sophisticated signal processing technology in the eeg signal acquisition is adopted in elementary amplification, after-treatment, A/D conversion) through elementary amplification, after-treatment, A/D, enter microprocessor, application contains based on definite analysis of components method of particle group optimizing that an electrical information EEG signals is extracted and the motor pattern classification, identify open, close up, wrist is stretched, wrist is bent, wrist outward turning, these six kinds of hand exercise patterns of wrist inward turning and hand attonity pattern.When recognition result was certain hand exercise pattern, the output corresponding control signal was given drive circuit, and drive motors drives the sports consciousness action of electrically powered hand by the people with disability.When recognition result is hand attonity pattern, output zero setting, the expression people with disability does not have hand exercise consciousness, does evil through another person and is failure to actuate, and is in relaxation state.Microprocessor output corresponding control signal is given drive circuit, and drive motors drives electrically powered hand action employing existing mature technology.
Traditional independent component analytical method makes up a separation matrix W=(w by the gradient descent algorithm optimizing
Ij)
N * n,, isolate and the same number of independent component of detected mixed signal according to u=Wx (wherein x picks up the detection signal of electric transducer for the brain electricity).But extract required signal composition, also want a large amount of postposition to handle; And owing to adopt gradient descent algorithm to make the optimization result be very easy to converge to a local optimum point, the signal element that causes needs to extract may be left in the basket.Can only extract the required definite signal relevant and use the definite analysis of components method that is proposed with reference signal based on particle group optimizing, promptly replace optimizing estimation entire n * n separation matrix W, only need application particle group optimizing method finds certain w of delegation among the W, thereby obtain expecting the signal composition that extracts by output signal u=wx, this can effectively improve efficiency of algorithm and accuracy.
Application comprises that based on definite analysis of components method of particle group optimizing structure reference signal, extraction contain an electrical information EEG signals and motor pattern classification.
Use this method and at first make up reference signal (comprising signal extraction reference signal and Classification and Identification reference signal).Specifically: at standard compliant EEG signals laboratory, gather that hands grasps, hands opens, turn on the wrist, turn under the wrist, the original detection signal of wrist inward turning, six kinds of true hand motion correspondences of wrist outward turning, adopt existing maturation method that it is carried out off-line analysis and handle, obtain corresponding with six kinds of hand exercise patterns containing an electrical information EEG signals as sample signal.
This sample signal as the Classification and Identification reference signal, is carried out motor pattern identification according to a Classification and Identification reference signal and a dependency of electrical information EEG signals that contains that is extracted.
(Projection Pursuit, PP) method makes up the signal extraction reference signal to adopt projection pursuit.Contain an electrical information EEG signals sample according to six classes and all have the characteristics of unimodal pulse, select for use unimodal pulse signal to make up the initial model of reference signal as need; Select six kinds to contain a kind of in the electrical information EEG signals sample again, according to Hall projection index, with the sample signal data projection on the one-dimensional space, find out with initial model and differ maximum projection, the structure that does not reflect in the initial model that this projection is comprised merges on the initial model new model that is improved.From then on new model sets out again, reselects a kind ofly from six kinds of sample signals, repeats above step, and model is revised once more, contains an electrical information EEG signals sample until six kinds and all selects to dispose by above-mentioned steps.Six classes that comprise that the correction model of this moment is required structure contain a signal extraction reference signal of electrical information EEG signals sample prior information.
The signal extraction reference signal is relevant with output signal as the required fitness function of constraints constitution optimization algorithm, specifically:
Adopt negentropy maximization object function, the negentropy J of output signal u (u) is
J(u)＝H(u
g)-H(u) (1)
U in the formula
gBe and the identical Gaussian random vector of output signal u variance; The negentropy of H (.) expression signal; To be it remain unchanged to any linear transformation of u the characteristics of negentropy, and total right and wrong are minus, and having only when u is Gauss distribution just is zero.Based on this point, negentropy is a good object function.Make system that the maximization of output composition negentropy can cause the separation of signal.
People such as A.Hyvarinen have proved that in the article of being delivered " New approximations of differentialentropy for independent component analysis and projection pursuit " value of J (u) can carry out reliable approximate evaluation by through type (2).
J(u)≈c[E{G(u)}-E{G(v)}]
2 (2)
G is any non-quadratic function in the formula, and c is positive constant, and v is a gaussian variable with zero-mean, unit variance.
To estimate to be correlated with as constraints between output u and the corresponding reference signal r, combine, be used to optimize solution procedure with the negentropy maximal criterion.Relevant through type (3) expression between output u and the corresponding reference signal r.
f(u)＝ε(u，r)＝Cov(u，r)＝E{u，r}-{E(u)E(r)
T}＞0 (3)
Formula (3) as constraints, is added object function, disposable accurate extraction is contained the optimization problem that a problem of electrical information EEG signals is converted into the inequality constraints of a band.
Maximize J
G(u)
subject?to?f(u)＝ε(u，r)＞0 (4)
Be to guarantee the signal composition of independent component for needing that extracted, promptly institute's composition that extracts reaches some suitable degree with the dependency of reference signal, also is simplification optimization algorithm simultaneously, and introducing vector z is converted to equality constraint with inequality constraints, makes
h(u)＝f(u)-z
2＝0 (5)
Then problem further is converted into the optimization problem of band equality constraint, as the formula (6).
Maximize?J
G(u)
subject?to h(u)＝ε(u，r)-z
2＝0 (6)
Use method of Lagrange multipliers and obtain corresponding Augmented Lagrangian Functions formula, with the fitness function of this functional expression, as the formula (7) as particle swarm optimization algorithm.
F(w，μ，z)＝J
G(u)+μh(u) (7)
μ is that positive Lagrangian number is taken advantage of vector in the formula; U=wx, x are a primary electrical information EEG signals that contains.
Utilize particle swarm optimization algorithm to carry out optimizing and calculate, obtain the vectorial w of row of required separation matrix W, thereby obtain expecting the signal composition that extracts by output signal u=wx.Particle swarm optimization algorithm is a kind of sophisticated parallel, global optimization approach at random, is a kind of general utility tool that solves optimization problem.In the DCA-PSO algorithm, each of optimization problem is separated and is called a particle, and i particle is expressed as κ
i=(w
i, μ
i, z
i), each particle has its direction and speed separately in the search volume, and the experience of foundation oneself and the experience of other particle are sought optimal location in problem search space flight.The behavior of each particle is assessed by defined fitness function.Finish the algorithm operation by setting maximum iteration time.
The flow process of carrying out optimizing calculating is as follows:
I) if optimize first, then determine individual number of initial population and relevant parameter, for i individuality, it has given at random position and speed.
Ii), calculate each individual fitness function of optimizing according to formula (7).
Iii) fitness function that each individuality is tried to achieve and the individual optimal solution in its record compare, if current optimal result better (the fitness function value is bigger), the then individual optimal solution of with replacement before of separating.In addition, if separating of trying to achieve at present is better than colony's optimal solution, then colony's optimal solution is reset to present result.
The iv) parameter value w that colony's optimal solution is tried to achieve, μ, z, the value before replacing.
V) revise each individual position and speed in the population according to formula (8), formula (9).
Δ m in the formula
i J+1The speed that is called particle i is characterized in the change of particle position in the j time iteration.m
i jBe characterized in the current location of particle i in the j time iteration.m
i J+1Be characterized in the correction position of particle i in the j time iteration.p
Id jBe characterized in the j time iteration the desired positions that particle is previous.p
Gd jBe characterized in the j time iteration the optimal location that current all particles have reached.
Be the positive acceleration coefficient, γ is called inertia power, r
1 i, r
2 jIt is the uniform random number between [0,1].
Obtain containing the EEG signals of an electrical information according to u=wx.
According to Classification and Identification reference signal and the method for being extracted that a dependency of electrical information EEG signals carries out motor pattern identification that contains be:
Foundation
Calculate respectively and extract signal and six kinds of Classification and Identification reference signal r
iBetween the size of correlation coefficient, u is by being extracted signal, r in the formula
iBe in six kinds of Classification and Identification reference signals.Selecting one group of the correlation coefficient maximum makes comparisons with preset threshold, if greater than the threshold value Δ, with the pairing hand exercise pattern of Classification and Identification reference signal of the maximum group of correlation coefficient as recognition result, export the motion control instruction of doing evil through another person accordingly,, illustrate that this moment, the people with disability did not have hand motion consciousness if maximum correlation coefficient is less than or equal to the threshold value Δ, differentiate and be hand attonity pattern, controller is exported zero setting, do evil through another person and be failure to actuate, 0.5≤Δ≤1.
The inventive method adopts eye, brain coordination mode to express the hand exercise wish, and the useful information that utilizes electro-ocular signal to comprise strengthens the feature of EEG signals that same sports consciousness produces; Use the definite analysis of components algorithm (DCA-PSO) that is proposed and extract, analyze, handle, realize the identification of hand multi-locomotion mode containing an electrical information EEG signals based on particle group optimizing.Because the DCA-PSO method is in the signal extraction process, to comprise needs the reference signal of extraction signal priori relevant with output signal as constraints, make up the optimization problem model of belt restraining condition, and application has parallel, at random, the particle swarm optimization algorithm of global optimization characteristic carries out optimizing and calculates, thereby guarantee the disposable desired signal composition that from original detection signal, extracts, and can overcome in the present independent component analytical method owing to the optimization algorithm that adopts based on gradient, make when solving actual higher-dimension optimization problem, converge to a local optimum point easily, rather than converge to the problem of global optimum's point.Since as bionical electrically powered hand signal source to contain an electrical information EEG signals feature obvious, thereby the accuracy height of hand exercise pattern recognition, the action control of doing evil through another person is reliable, realized under the highly reliable discrimination Three Degree Of Freedom real-time control of six actions of doing evil through another person, and the paralysis personage who is specially adapted to nerve, muscle serious degradation uses.
Description of drawings
Fig. 1 among the present invention based on the sketch map of definite analysis of components method of particle group optimizing;
Fig. 2 is a mode identification method sketch map among Fig. 1.
The specific embodiment
The present invention selects the brain electricity to pick up the EEG signals that the electric transducer collection comprises an electrical information.Each brain electricity picks up that electric transducer comprises the scalp power-collecting electrode and as for the reference electrode of ear, reach the elementary amplifying circuit that is connected with reference electrode with power-collecting electrode, the outfan of elementary amplifying circuit connects with corresponding after-treatment circuit input end respectively, and the after-treatment circuit comprises trap circuit, back level amplifying circuit, the compensating circuit (being used to eliminate common-mode signal) of 50Hz.The after-treatment circuit output end is connected with the input of A/D change-over circuit.Three motors that 3-degree-of-freedom electrical is done evil through another person connect with corresponding drive circuit respectively.Microprocessor is connected with the outfan of A/D change-over circuit, the input end signal of drive circuit respectively.
The drive circuit of the elementary amplifying circuit in this device, after-treatment circuit (containing back level amplifying circuit, filter circuit etc.), A/D change-over circuit, microprocessor, motor all adopts bionical corresponding circuit and the device of doing evil through another person of existing single-degree-of-freedom.
The electrically powered hand control method of utilizing eye electricity and brain electric information is that the scalp power-collecting electrode that the brain electricity picks up in the electric transducer is placed on the brain frons F that international electroencephalography meeting standard 10-20 lead system is determined
P1Or F
P2The position, reference electrode is placed on the auricle position, when wearing the eye that people with disability that 3-degree-of-freedom electrical does evil through another person adopts nature, the brain coordination mode carries out the hand exercise imagination, during the expressive movement will, pick up primary signal that electric transducer picks up through elementary amplification by the brain electricity, after-treatment, A/D converts data acquisition, enter microprocessor, application contains an electrical information EEG signals extraction and motor pattern classification based on definite analysis of components method of particle group optimizing, identifies and opens, close up, wrist is stretched, wrist is bent, the wrist outward turning, these six kinds of hand exercise patterns of wrist inward turning and hand attonity pattern; Hand exercise pattern output corresponding control signal according to identification is given drive circuit, drives three motors of doing evil through another person through drive circuit, finishes certain action that 3-degree-of-freedom electrical is done evil through another person; When recognition result was hand attonity pattern, output zero setting was done evil through another person and is failure to actuate.
As shown in Figure 1, the definite analysis of components method based on particle group optimizing comprises that structure reference signal, extraction contain an electrical information EEG signals and motor pattern classification.
(1) reference signal comprises signal extraction reference signal and Classification and Identification reference signal, and the concrete grammar that makes up reference signal is:
1. at standard compliant EEG signals laboratory, gather that hands grasps, hands opens, turn on the wrist, turn under the wrist, the original detection signal of wrist inward turning, six kinds of true hand motion correspondences of wrist outward turning, carry out off-line analysis and handle, obtain corresponding with six kinds of hand exercise patterns containing an electrical information EEG signals as sample signal.
2. with this sample signal directly as the Classification and Identification reference signal.
3. adopt projection Pursuit Method to make up the signal extraction reference signal, specifically: contain a unimodal pulse signal that electrical information EEG signals sample all comprises as the initial model that makes up reference signal according to six classes; Select six kinds to contain a kind of in the electrical information EEG signals sample again, according to Hall projection index, with the sample signal data projection on the one-dimensional space, find out with initial model and differ maximum projection, the structure that does not reflect in the initial model that this projection is comprised merges on the initial model new model that is improved; Again from new model, from six kinds of sample signals, reselect a kind ofly, repeat above step, model is revised once more, contain an electrical information EEG signals sample until six kinds and all select to dispose by above-mentioned steps; Final correction model is the signal extraction reference signal.
(2) extract and to contain a concrete grammar of electrical information EEG signals and be:
A. at first that signal extraction reference signal r is relevant with output signal u as the required fitness function of constraints constitution optimization algorithm, specifically:
Adopt negentropy maximization object function, the negentropy J of output signal u (u) is
J(u)＝H(u
g)-H(u) (1)
U in the formula
gBe the Gaussian random vector identical, the negentropy of H (.) expression signal with output signal u variance; Formula (1) is reduced to
J(u)≈c[E{G(u)}-E{G(v)}]
2 (2)
G is any non-quadratic function in the formula, and c is positive constant, and v is a gaussian variable with zero-mean, unit variance.
With relevant between output signal u and the signal extraction reference signal r, be expressed as constraints
ε(u，r)＝Cov(u，r)＝E{u，r}-{E(u)E(r)
T}＞0 (3)
Formula (3) as constraints, is added negentropy maximization object function, introduce vector z simultaneously, extraction is contained the optimization model that a problem of electrical information EEG signals is converted into the band equality constraint, be expressed as
Maximize J
G(u)
subject?to?h(u)＝ε(u，r)-z
2＝0 (4)
Use method of Lagrange multipliers and obtain corresponding Augmented Lagrangian Functions formula,, be expressed as the fitness function of this functional expression as particle swarm optimization algorithm
F(w，μ，z)＝J
G(u)+μh(u) (5)
μ is that positive Lagrangian number is taken advantage of vector in the formula; U=wx, x are that original detection signal, w are the row vector of separation matrix W.
B. utilize particle swarm optimization algorithm to carry out optimizing and calculate, obtain the vectorial w of row of required separation matrix W, thereby obtain expecting the signal composition that extracts by output signal u=wx; Finish the algorithm operation by setting maximum iteration time.The flow process of carrying out optimizing calculating is as follows:
I) if optimize first, then determine individual number of initial population and relevant parameter, for i individuality, it has given at random position and speed, and i particle is expressed as κ
i=(w
i, μ
i, z
i);
Ii), calculate each individual fitness function F (w that optimizes according to formula (5)
i, μ
i, z
i);
Fitness function F (the w that iii) each individuality is tried to achieve
i, μ
i, z
i) with the record in this individual adaptive optimal control degree functional value compare, if current fitness function value is greater than this previous individual adaptive optimal control degree functional value, then with current fitness function value as the record in adaptive optimal control degree functional value, if current fitness function value is smaller or equal to this previous individual adaptive optimal control degree functional value, then this individual adaptive optimal control degree functional value in the record is constant; If current adaptive optimal control degree functional value that should individuality is greater than the adaptive optimal control degree functional value of colony, then with current adaptive optimal control degree functional value that should individuality as the adaptive optimal control degree functional value of colony; If the current little adaptive optimal control degree functional value that equals colony greatly of adaptive optimal control degree functional value that should individuality, then the adaptive optimal control degree functional value of colony is constant;
The iv) parameter value w that colony's optimal solution is tried to achieve, μ, z, the value before replacing;
V) revise each individual position and speed in the population according to formula (6) and formula (7);
Δ m in the formula
i j+ 1 is the speed of particle i, i.e. the change of particle position in the j time iteration; m
i jCurrent location for particle i in the j time iteration; m
i J+1Correction position for particle i in the j time iteration; p
Id jFor in the j time iteration, the desired positions that particle is previous; p
Gd jFor in the j time iteration, the optimal location that current all particles have reached;
Be the positive acceleration coefficient, γ is an inertia power, r
1 i, r
2 jIt is the uniform random number between [0,1];
C. obtain output signal u according to u=wx, promptly contain the EEG signals of an electrical information;
(3) as shown in Figure 2, the motor pattern classification is to carry out motor pattern identification according to a Classification and Identification reference signal and a dependency of electrical information EEG signals that contains that is extracted, and concrete grammar is:
Foundation
Calculate respectively and extract signal and six kinds of Classification and Identification reference signal r
iBetween the size of correlation coefficient, u is by being extracted signal, r in the formula
iBe in six kinds of Classification and Identification reference signals; Selecting one group of the correlation coefficient maximum makes comparisons with preset threshold, if greater than the threshold value Δ, with the pairing hand exercise pattern of Classification and Identification reference signal of the maximum group of correlation coefficient as recognition result, export the motion control instruction of doing evil through another person accordingly,, illustrate that this moment, the people with disability did not have hand motion consciousness if maximum correlation coefficient is less than or equal to the threshold value Δ, differentiate and be hand attonity pattern, controller is exported zero setting, do evil through another person and be failure to actuate, 0.5≤Δ≤1.
Concrete work process is:
When wearing people with disability that 3-degree-of-freedom electrical does evil through another person when needing the artificial hand controlled action, brain, eye, the Handball Association of being accustomed to adopting with reference to people transfers in the pattern of capable hand exercise, adopt eye brain coordination mode to carry out the hand exercise imagination, the electrically powered hand system will be according to an electrical information EEG signals that contains of corresponding hand motion, on analysis, base of recognition, artificial hand controlled is finished corresponding action: turn over motor control consciousness on the wrist, do evil through another person to finish synchronously and stretch the wrist action; Turn over motor control consciousness under the wrist, do evil through another person to finish synchronously and bend the wrist action; Wrist outward turning motor control consciousness is done evil through another person and is finished the wrist outward turning synchronously; Wrist inward turning motor control consciousness is done evil through another person and is finished the wrist inward turning synchronously; The five fingers stretching control consciousness, the hand of doing evil through another person opens; The motor control of clenching fist consciousness, the hand of doing evil through another person closes up; Hand loosens consciousness, does evil through another person and does not do any action.Avoid the multiple freedom degrees hand-prosthesis of conventional commercial need be, realized the do evil through another person real-time control of six actions of Three Degree Of Freedom by repeatedly switching the situation just can reach multiple freedom degrees hand-prosthesis control.
Claims (1)
1. utilize the electrically powered hand control method of eye electricity and brain electric information, it is characterized in that this control method is that the scalp power-collecting electrode that the brain electricity picks up in the electric transducer is placed on the brain frons F that international electroencephalography meeting standard 10-20 lead system is determined
P1Or F
P2Position, reference electrode are placed on the auricle position, pick up the primary signal that electric transducer picks up by the brain electricity and convert data acquisition through elementary amplification, after-treatment, A/D, enter microprocessor, application contains based on definite analysis of components method of particle group optimizing that an electrical information EEG signals is extracted and the motor pattern classification, identify open, close up, wrist is stretched, wrist is bent, wrist outward turning, these six kinds of hand exercise patterns of wrist inward turning and hand attonity pattern; Hand exercise pattern output corresponding control signal according to identification is given drive circuit, and drive motors drives the electrically powered hand action; When recognition result was hand attonity pattern, output zero setting was done evil through another person and is failure to actuate;
Described definite analysis of components method based on particle group optimizing comprises that structure reference signal, extraction contain an electrical information EEG signals and motor pattern classification;
(1) described reference signal comprises signal extraction reference signal and Classification and Identification reference signal, and the concrete grammar that makes up reference signal is:
1. at standard compliant EEG signals laboratory, gather that hands grasps, hands opens, turn on the wrist, turn under the wrist, the original detection signal of wrist inward turning, six kinds of true hand motion correspondences of wrist outward turning, carry out off-line analysis and handle, obtain corresponding with six kinds of hand exercise patterns containing an electrical information EEG signals as sample signal;
2. with this sample signal directly as the Classification and Identification reference signal;
3. adopt projection Pursuit Method to make up the signal extraction reference signal, specifically: contain a unimodal pulse signal that electrical information EEG signals sample all comprises as the initial model that makes up reference signal according to six classes; Select six kinds to contain a kind of in the electrical information EEG signals sample again, according to Hall projection index, with the sample signal data projection on the one-dimensional space, find out with initial model and differ maximum projection, the structure that does not reflect in the initial model that this projection is comprised merges on the initial model new model that is improved; Again from new model, from six kinds of sample signals, reselect a kind ofly, repeat above step, model is revised once more, contain an electrical information EEG signals sample until six kinds and all select to dispose by above-mentioned steps; Final correction model is the signal extraction reference signal;
(2) extract and to contain a concrete grammar of electrical information EEG signals and be:
A. at first that signal extraction reference signal r is relevant with output signal u as the required fitness function of constraints constitution optimization algorithm, specifically:
Adopt negentropy maximization object function, the negentropy J of output signal u (u) is
J(u)＝H(u
g)-H(u) (1)
U in the formula
gBe the Gaussian random vector identical, the negentropy of H (.) expression signal with output signal u variance; Formula (1) is reduced to
J(u)≈c[E{G(u)}-E{G(v)}]
2 (2)
G is any non-quadratic function in the formula, and c is positive constant, and v is a gaussian variable with zero-mean, unit variance;
With relevant between output signal u and the signal extraction reference signal r, be expressed as constraints
ε(u，r)＝Cov(u，r)＝E{u，r}-{E(u)E(r)
T}＞0(3)
Formula (3) as constraints, is added negentropy maximization object function, introduce vector z simultaneously, extraction is contained the optimization model that a problem of electrical information EEG signals is converted into the band equality constraint, be expressed as
Maximize J
G(u)
subject?to?h(u)＝ε(u，r)-z
2＝0 (4)
Use method of Lagrange multipliers and obtain corresponding Augmented Lagrangian Functions formula,, be expressed as the fitness function of this functional expression as particle swarm optimization algorithm
F(w，μ，z)＝J
G(u)+μh(u) (5)
μ is that positive Lagrangian number is taken advantage of vector in the formula; U=wx, x are that original detection signal, w are the row vector of separation matrix W;
B. utilize particle swarm optimization algorithm to carry out optimizing and calculate, obtain the vectorial w of row of required separation matrix W, thereby obtain expecting the signal composition that extracts by output signal u=wx; Finish the algorithm operation by setting maximum iteration time;
The flow process of carrying out optimizing calculating is as follows:
I) if optimize first, then determine individual number of initial population and relevant parameter, for i individuality, it has given at random position and speed, and i particle is expressed as κ
i=(w
i, μ
i, z
i);
Ii), calculate each individual fitness function F (w that optimizes according to formula (5)
i, μ
i, z
i);
Fitness function F (the w that iii) each individuality is tried to achieve
i, μ
i, z
i) with the record in this individual adaptive optimal control degree functional value compare, if current fitness function value is greater than this previous individual adaptive optimal control degree functional value, then with current fitness function value as the record in adaptive optimal control degree functional value, if current fitness function value is smaller or equal to this previous individual adaptive optimal control degree functional value, then this individual adaptive optimal control degree functional value in the record is constant; If current adaptive optimal control degree functional value that should individuality is greater than the adaptive optimal control degree functional value of colony, then with current adaptive optimal control degree functional value that should individuality as the adaptive optimal control degree functional value of colony; If the current little adaptive optimal control degree functional value that equals colony greatly of adaptive optimal control degree functional value that should individuality, then the adaptive optimal control degree functional value of colony is constant;
The iv) parameter value w that colony's optimal solution is tried to achieve, μ, z, the value before replacing;
V) revise each individual position and speed in the population according to formula (6) and formula (7);
Δ m in the formula
i J+1Be the speed of particle i, i.e. the change of particle position in the j time iteration; m
i jCurrent location for particle i in the j time iteration; m
i J+1Correction position for particle i in the j time iteration; p
Id jFor in the j time iteration, the desired positions that particle is previous; p
Gd jFor in the j time iteration, the optimal location that current all particles have reached;
Be the positive acceleration coefficient, γ is an inertia power, r
1 i, r
2 jIt is the uniform random number between [0,1];
C. obtain output signal u according to u=wx, promptly contain the EEG signals of an electrical information;
(3) described motor pattern classification is to carry out motor pattern identification according to a Classification and Identification reference signal and a dependency of electrical information EEG signals that contains that is extracted, and concrete grammar is:
Foundation
Calculate respectively and extract signal and six kinds of Classification and Identification reference signal r
iBetween the size of correlation coefficient, μ is by being extracted signal, r in the formula
iBe in six kinds of Classification and Identification reference signals; Selecting one group of the correlation coefficient maximum makes comparisons with preset threshold, if greater than the threshold value Δ, with the pairing hand exercise pattern of Classification and Identification reference signal of the maximum group of correlation coefficient as recognition result, export the motion control instruction of doing evil through another person accordingly,, illustrate that this moment, the people with disability did not have hand motion consciousness if maximum correlation coefficient is less than or equal to the threshold value Δ, differentiate and be hand attonity pattern, controller is exported zero setting, do evil through another person and be failure to actuate, 0.5≤Δ≤1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910154966XA CN101711709B (en) | 2009-12-07 | 2009-12-07 | Method for controlling electrically powered artificial hands by utilizing electro-coulogram and electroencephalogram information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910154966XA CN101711709B (en) | 2009-12-07 | 2009-12-07 | Method for controlling electrically powered artificial hands by utilizing electro-coulogram and electroencephalogram information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101711709A true CN101711709A (en) | 2010-05-26 |
| CN101711709B CN101711709B (en) | 2012-05-23 |
Family
ID=42415973
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200910154966XA Expired - Fee Related CN101711709B (en) | 2009-12-07 | 2009-12-07 | Method for controlling electrically powered artificial hands by utilizing electro-coulogram and electroencephalogram information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101711709B (en) |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102133139A (en) * | 2011-01-21 | 2011-07-27 | 华南理工大学 | Artificial hand control system and method |
| CN102521505A (en) * | 2011-12-08 | 2012-06-27 | 杭州电子科技大学 | Brain electric and eye electric signal decision fusion method for identifying control intention |
| CN102697493A (en) * | 2012-05-03 | 2012-10-03 | 北京工业大学 | Method for rapidly and automatically identifying and removing ocular artifacts in electroencephalogram signal |
| CN103235872A (en) * | 2013-04-03 | 2013-08-07 | 浙江工商大学 | Projection pursuit dynamic cluster method for multidimensional index based on particle swarm optimization |
| CN103338265A (en) * | 2013-07-10 | 2013-10-02 | 安徽大学 | Information interaction system and information interaction method for combining brain electricity and eye electricity |
| CN103690281A (en) * | 2014-01-04 | 2014-04-02 | 张江杰 | Brain wave controlled artificial limb system |
| CN104825256A (en) * | 2015-04-30 | 2015-08-12 | 南京信息工程大学 | Artificial limb system with perception feedback function |
| CN105708587A (en) * | 2016-01-25 | 2016-06-29 | 电子科技大学 | Lower-limb exoskeleton training method and system triggered by brain-computer interface under motion imagination pattern |
| CN105943207A (en) * | 2016-06-24 | 2016-09-21 | 吉林大学 | Intelligent artificial limb movement system based on idiodynamics and control methods thereof |
| CN106066909A (en) * | 2016-05-27 | 2016-11-02 | 电子科技大学 | Torpedo Shape multidisciplinary design optimization based on ALMABC algorithm |
| CN106155296A (en) * | 2015-04-20 | 2016-11-23 | 北京智谷睿拓技术服务有限公司 | Control method and equipment |
| CN106293013A (en) * | 2015-04-30 | 2017-01-04 | 北京智谷睿拓技术服务有限公司 | Headwork determines method and apparatus |
| CN106293049A (en) * | 2015-08-19 | 2017-01-04 | 北京智谷睿拓技术服务有限公司 | Input information determines method and apparatus |
| CN108175547A (en) * | 2018-01-30 | 2018-06-19 | 浙江省公众信息产业有限公司 | Control method, device and the computer readable storage medium of artificial prosthesis |
| CN108319928A (en) * | 2018-02-28 | 2018-07-24 | 天津大学 | A kind of deep learning model and application based on Multi-objective PSO optimization |
| CN108523883A (en) * | 2018-03-19 | 2018-09-14 | 天津大学 | A kind of continuous Mental imagery identifying system of left and right index finger based on actual act modeling |
| CN109199410A (en) * | 2018-06-28 | 2019-01-15 | 浙江大学 | A kind of Speech perception appraisal procedure based on eye motion |
| CN109521880A (en) * | 2018-11-27 | 2019-03-26 | 东南大学 | A kind of Teleoperation Systems and method based on the driving of mixed biologic electric signal |
| CN113595569A (en) * | 2021-07-29 | 2021-11-02 | 中国人民解放军国防科技大学 | ICA-R algorithm-based 2FSK signal anti-interference method and device |
| CN113807402A (en) * | 2021-08-23 | 2021-12-17 | 天津大学 | System for inhibiting MIs-triggering of MI-BCI system and training and testing method thereof |
| CN114089831A (en) * | 2021-11-17 | 2022-02-25 | 杭州电力设备制造有限公司 | Control method of automatic lifting device for assisting disassembly and assembly of mutual inductor in box transformer substation |
| CN114327048A (en) * | 2021-12-07 | 2022-04-12 | 重庆邮电大学 | Mechanical arm control method and system based on electroencephalogram signals and electro-oculogram signals |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100515366C (en) * | 2007-04-24 | 2009-07-22 | 杭州电子科技大学 | Real time control device and control method by two-degrees-of freedom myoelectricity artificial hand |
| CN100546553C (en) * | 2007-05-18 | 2009-10-07 | 天津大学 | Adopt the prosthetic hand and the control method thereof of myoelectricity and brain electricity Collaborative Control |
| CN201223477Y (en) * | 2008-06-13 | 2009-04-22 | 天津三星电子有限公司 | Mechanical prosthetic limb controlled by thought of human brain |
| CN101455596A (en) * | 2008-12-18 | 2009-06-17 | 西安交通大学苏州研究院 | Nerve artificial limb hand driven and controlled by brain-computer interface and control method thereof |
| CN101474070B (en) * | 2009-01-21 | 2010-08-04 | 电子科技大学 | Method for removing ocular artifacts in brain-electrical signal |
-
2009
- 2009-12-07 CN CN200910154966XA patent/CN101711709B/en not_active Expired - Fee Related
Cited By (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102133139A (en) * | 2011-01-21 | 2011-07-27 | 华南理工大学 | Artificial hand control system and method |
| CN102521505A (en) * | 2011-12-08 | 2012-06-27 | 杭州电子科技大学 | Brain electric and eye electric signal decision fusion method for identifying control intention |
| CN102521505B (en) * | 2011-12-08 | 2014-11-26 | 杭州电子科技大学 | Brain electric and eye electric signal decision fusion method for identifying control intention |
| CN102697493A (en) * | 2012-05-03 | 2012-10-03 | 北京工业大学 | Method for rapidly and automatically identifying and removing ocular artifacts in electroencephalogram signal |
| CN103235872A (en) * | 2013-04-03 | 2013-08-07 | 浙江工商大学 | Projection pursuit dynamic cluster method for multidimensional index based on particle swarm optimization |
| CN103338265A (en) * | 2013-07-10 | 2013-10-02 | 安徽大学 | Information interaction system and information interaction method for combining brain electricity and eye electricity |
| CN103338265B (en) * | 2013-07-10 | 2016-03-30 | 安徽大学 | A kind of in conjunction with brain electricity and the information interaction system of eye electricity and information interacting method |
| CN103690281A (en) * | 2014-01-04 | 2014-04-02 | 张江杰 | Brain wave controlled artificial limb system |
| CN106155296A (en) * | 2015-04-20 | 2016-11-23 | 北京智谷睿拓技术服务有限公司 | Control method and equipment |
| US10613623B2 (en) | 2015-04-20 | 2020-04-07 | Beijing Zhigu Rui Tuo Tech Co., Ltd | Control method and equipment |
| CN104825256A (en) * | 2015-04-30 | 2015-08-12 | 南京信息工程大学 | Artificial limb system with perception feedback function |
| US10646141B2 (en) | 2015-04-30 | 2020-05-12 | Beijing Zhigu Rui Tuo Tech Co., Ltd | Method and device for determining head movement |
| CN106293013A (en) * | 2015-04-30 | 2017-01-04 | 北京智谷睿拓技术服务有限公司 | Headwork determines method and apparatus |
| CN106293049A (en) * | 2015-08-19 | 2017-01-04 | 北京智谷睿拓技术服务有限公司 | Input information determines method and apparatus |
| CN105708587B (en) * | 2016-01-25 | 2017-11-03 | 电子科技大学 | A kind of the lower limb exoskeleton training method and system of the triggering of Mental imagery pattern brain-computer interface |
| CN105708587A (en) * | 2016-01-25 | 2016-06-29 | 电子科技大学 | Lower-limb exoskeleton training method and system triggered by brain-computer interface under motion imagination pattern |
| CN106066909A (en) * | 2016-05-27 | 2016-11-02 | 电子科技大学 | Torpedo Shape multidisciplinary design optimization based on ALMABC algorithm |
| CN105943207A (en) * | 2016-06-24 | 2016-09-21 | 吉林大学 | Intelligent artificial limb movement system based on idiodynamics and control methods thereof |
| CN108175547A (en) * | 2018-01-30 | 2018-06-19 | 浙江省公众信息产业有限公司 | Control method, device and the computer readable storage medium of artificial prosthesis |
| CN108319928A (en) * | 2018-02-28 | 2018-07-24 | 天津大学 | A kind of deep learning model and application based on Multi-objective PSO optimization |
| CN108319928B (en) * | 2018-02-28 | 2022-04-19 | 天津大学 | Deep learning method and system based on multi-target particle swarm optimization algorithm |
| CN108523883A (en) * | 2018-03-19 | 2018-09-14 | 天津大学 | A kind of continuous Mental imagery identifying system of left and right index finger based on actual act modeling |
| CN109199410A (en) * | 2018-06-28 | 2019-01-15 | 浙江大学 | A kind of Speech perception appraisal procedure based on eye motion |
| CN109521880A (en) * | 2018-11-27 | 2019-03-26 | 东南大学 | A kind of Teleoperation Systems and method based on the driving of mixed biologic electric signal |
| CN113595569A (en) * | 2021-07-29 | 2021-11-02 | 中国人民解放军国防科技大学 | ICA-R algorithm-based 2FSK signal anti-interference method and device |
| CN113807402A (en) * | 2021-08-23 | 2021-12-17 | 天津大学 | System for inhibiting MIs-triggering of MI-BCI system and training and testing method thereof |
| CN114089831A (en) * | 2021-11-17 | 2022-02-25 | 杭州电力设备制造有限公司 | Control method of automatic lifting device for assisting disassembly and assembly of mutual inductor in box transformer substation |
| CN114327048A (en) * | 2021-12-07 | 2022-04-12 | 重庆邮电大学 | Mechanical arm control method and system based on electroencephalogram signals and electro-oculogram signals |
| CN114327048B (en) * | 2021-12-07 | 2024-04-19 | 山东华数信息技术股份有限公司 | Mechanical arm control method and system based on electroencephalogram signals and electrooculogram signals |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101711709B (en) | 2012-05-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101711709B (en) | Method for controlling electrically powered artificial hands by utilizing electro-coulogram and electroencephalogram information | |
| CN106726030B (en) | Brain machine interface system and its application based on Clinical EEG Signals control robot movement | |
| CN103793058B (en) | A kind of active brain-computer interactive system Mental imagery classification of task method and device | |
| CN107037883A (en) | A kind of mixing brain machine interface system and method based on Mental imagery | |
| CN110179643A (en) | A kind of neck rehabilitation training system and training method based on annulus sensor | |
| CN103955269A (en) | Intelligent glass brain-computer interface method based on virtual real environment | |
| CN111110982A (en) | Hand rehabilitation training method based on motor imagery | |
| Birch et al. | Brain-computer interface research at the Neil Squire Foundation | |
| CN110534180B (en) | Deep learning human-computer interaction motor imagery brain-computer interface system and training method | |
| CN103699217A (en) | Two-dimensional cursor motion control system and method based on motor imagery and steady-state visual evoked potential | |
| CN111584031B (en) | Brain-controlled intelligent limb rehabilitation system based on portable electroencephalogram acquisition equipment and application | |
| CN102096468A (en) | Brain-computer interface (BCI)-based home appliance remote control device and method | |
| CN103258215A (en) | Multi-lead correlation analysis electroencephalo-graph (EEG) feature extraction method | |
| Zhang et al. | Asynchronous brain-computer interface shared control of robotic grasping | |
| CN113143676B (en) | Control method of external limb finger based on brain-muscle-electricity cooperation | |
| Borisoff et al. | Brain interface research for asynchronous control applications | |
| CN102306303A (en) | Electroencephalography signal characteristic extraction method based on small training samples | |
| Birch et al. | Current trends in brain-computer interface research at the Neil Squire Foundation | |
| CN103294192A (en) | LED lamp switch control device and control method thereof based on motor imagery | |
| CN113208593A (en) | Multi-modal physiological signal emotion classification method based on correlation dynamic fusion | |
| CN110074779A (en) | A kind of EEG signal identification method and device | |
| Millán et al. | Non-invasive brain-actuated control of a mobile robot | |
| WO2022099807A1 (en) | Robot natural control method based on electromyographic signal and error electroencephalographic potential | |
| CN109498362A (en) | A kind of hemiplegic patient's hand movement function device for healing and training and model training method | |
| CN113408397A (en) | Domain-adaptive cross-subject motor imagery electroencephalogram signal identification system and method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| ASS | Succession or assignment of patent right |
Owner name: NANTONG HENGLI HEAVY INDUSTRY MACHINERY CO., LTD. Free format text: FORMER OWNER: HANGZHOU ELECTRONIC SCIENCE AND TECHNOLOGY UNIV Effective date: 20131021 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| COR | Change of bibliographic data |
Free format text: CORRECT: ADDRESS; FROM: 310018 HANGZHOU, ZHEJIANG PROVINCE TO: 226600 NANTONG, JIANGSU PROVINCE |
|
| TR01 | Transfer of patent right |
Effective date of registration: 20131021 Address after: 226600, No. seven, Renmin Road, Hu Hu Town, Haian County, Jiangsu, Nantong Patentee after: Nantong Hengli Heavy Industries Machinery Co., Ltd. Address before: Hangzhou City, Zhejiang province 310018 Jianggan District Xiasha Higher Education Park No. 2 street Patentee before: Hangzhou Electronic Science and Technology Univ |
|
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120523 Termination date: 20181207 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |