CN106940592A - A kind of brain control visual human avoidance obstacle method - Google Patents
A kind of brain control visual human avoidance obstacle method Download PDFInfo
- Publication number
- CN106940592A CN106940592A CN201710088805.XA CN201710088805A CN106940592A CN 106940592 A CN106940592 A CN 106940592A CN 201710088805 A CN201710088805 A CN 201710088805A CN 106940592 A CN106940592 A CN 106940592A
- Authority
- CN
- China
- Prior art keywords
- button
- brain
- flicker
- eeg signals
- score value
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/011—Arrangements for interaction with the human body, e.g. for user immersion in virtual reality
- G06F3/015—Input arrangements based on nervous system activity detection, e.g. brain waves [EEG] detection, electromyograms [EMG] detection, electrodermal response detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
- G06F18/24155—Bayesian classification
Abstract
The invention discloses a kind of brain control visual human avoidance obstacle method, including off-line training step：System initialization；Start interactive interface；User watches the button flashed accordingly attentively according to prompting, and the EEG signals of user are gathered by electrode cap；Handle the EEG signals collected；It is determined that the setup parameter of the threshold value and maximum button flicker wheel number, minimum button flicker wheel number of the maximum difference that score value is returned with second largest normalization as on-line training stage self adaptation Bayes classifier；On-line Control step：Relevant parameter is reinitialized, starts interactive interface；User decides the button for watching flicker attentively in its sole discretion, while electrode cap gathers EEG signals；Handle the EEG signals collected；Control command is generated according to the classification results of output, motion of virtual human is controlled.The inventive method can control visual human to carry out moving avoiding barrier accordingly in virtual scene, improve the training effect to brain control, allow user to obtain more preferable experience.
Description
Technical field
The present invention relates to brain-computer interface and field of virtual reality, more particularly to a kind of brain control visual human avoidance obstacle method.
Background technology
Patient with exercise functional disorder itself can not well contact with the external world, brain-computer interface (brain-
Computer interface, BCI) technology can allow these patients directly to control brain-computer interface equipment by brain signal, real
Now exchanged and the operation of object to external world with extraneous.The essence of brain-computer interface technology is that one is set up between human brain and computer
Kind of external channel, this passage need not the participation of nerve or musculature can just realize external communication and control.
Last century, people substitute some missings of patient to the research of brain-computer interface mainly around control external equipment
Function, such as control artificial limb, wheelchair, spelling typewriting etc..And achieve important breakthrough in terms of this century brain computer interface application field：Will
Brain-computer interface technology is combined with virtual reality (virtual reality, VR), constitutes the brain-computer interface based on virtual reality new
Form.The most basic mode of virtual reality is the virtual world that a three dimensions is produced using computer simulation.Virtual reality is not
Only to the simulation of vision, while the also simulation of the sense organ such as the sense of hearing, tactile, so that allowing user to have plants a sense on the spot in person
Feel, user can experience the things in three dimensions in real time, without limitation.And the brain-computer interface system based on virtual reality
System be exactly simulate a reality environment true to nature and design one can real-time, interactive system, so as to allow the user can be with
Perceive the things in virtual scene.
Brain-computer interface technology based on virtual reality can either retain the effect of brain-computer interface, also absorb virtual reality
Advantage, with extremely wide application prospect.On the one hand, brain-computer interface can be used as the brain machine interface system based on virtual reality
Input equipment, has expanded virtual reality system input mode；Brain-computer interface input is than traditional virtual reality system input mode (ratio
Such as keyboard, mouse, joystick) it is more directly perceived, it is not necessary to control by hand but by idea the change of virtual scene；Separately
On the one hand, virtual reality technology can as the highly useful feedback of the information mode of brain machine interface system, virtual reality feedback with
Traditional simple feedback pattern of brain-computer interface is compared, can be provided for brain-computer interface user it is more positive, more rich it is colorful, more
The situation feedback model of incentive.Virtual reality technology, which is incorporated into brain-computer interface, can shorten the brain-computer interface training time.By
In user it is trained and is tested in the scene true to nature of simulation, and be no longer mechanical brain-computer interface equipment, this is very big
Ground reduces the security risk and expense of experiment, but can but obtain identical effect.
The technology that brain-computer interface is combined with virtual reality at this stage achieves certain progress, occurs in that a small amount of based on fortune
The dynamic imagination or P300 brain control dummy scene roaming, brain control virtual game etc.；However, there is presently no occur controlling by P300
The system of visual human's avoiding barrier processed.It is simply simple to use and the virtual environment of research system is relatively dull before
Virtual environment is observed in camera movement, the visual human of control can only some position of body do mechanicalness and specifically act, it appears it is slow-witted
Plate, interactivity are poor.
The content of the invention
It is an object of the invention to the shortcoming for overcoming prior art, there is provided a kind of brain control visual human avoidance obstacle side with deficiency
Method, can control visual human to carry out moving avoiding barrier accordingly in virtual scene, improve the training effect to brain control, allow
User obtain more preferable experience sense by.
A kind of brain control visual human avoidance obstacle method, including off-line training step and On-line Control step；
A off-line trainings：
S1, system initialization；
S2, startup interactive interface, interactive interface show the button and prompt window flashed at random；
S3, user watch the button flashed accordingly attentively according to the instruction of prompt window, and user is gathered by electrode cap
EEG signals；
The EEG signals that S4, processing are collected, including：Pretreatment, signal characteristic abstraction and is entered using Bayes classifier
Row classification based training, obtains corresponding normalize of different buttons and returns score value, it is to be considered as to make that normalization, which returns the maximum button of score value,
The button that user watches attentively；
It is determined that the threshold value of the maximum difference that score value is returned with second largest normalization, and maximum button flicker wheel number, minimum
Button flicker wheel number as on-line training stage self adaptation Bayes classifier setup parameter；
B On-line Controls：
S5, reinitialize relevant parameter, start interactive interface, prompt window is no longer shown；
S6, user decide the button for watching flicker attentively in its sole discretion according to Obstacle Position and visual human position, while electrode cap
Gather the EEG signals of user；
S7, collect after EEG signals, handle the EEG signals collected, including：Pretreatment, signal characteristic abstraction and profit
Classification based training is carried out with self adaptation Bayes classifier；
S8, according to the classification results of output generate control command, control motion of virtual human.
It is preferred that, each user carries out 20 groups of tests during off-line training, and button of the every group of test bag containing 10 wheels dodges
It is bright.
It is preferred that, it is determined that maximum button flicker wheel number, the method for minimum button flicker wheel number are：Multigroup test is carried out, often
Then group test bag takes the average accuracy A taken turns as m containing many wheel button flickers to the preceding m accuracy taken turnsm；Accuracy AmWith
Certain change is presented in the increase for number of rounds of tests, draws accuracy AmWith the graph of a relation of button flicker wheel number, by analysis, it is ensured that
Maximum button flicker wheel number k is selected under conditions of certain accuracymaxWith minimum button flicker wheel number kmin。
Specifically, minimum button flicker wheel number is 3, maximum button flicker wheel number is 8.
It is preferred that, it is determined that the threshold method of the maximum difference for returning score value with second largest normalization is：Collect K wheel brains
During electric signal, the preceding K characteristic vectors taken turns are averaged, the characteristic vector after being then averaged is classified with Bayes classifier
Return, the different buttons of acquisition are corresponding to return score value Si, maximum normalization returns score value and is designated as Smax, second largest normalization
Return score value and be designated as Ssec, define Smax-Ssec=Δ θ (0 ＜ Δ θ ＜ 1), sets different Δ θ threshold θ0；
Judge Δ θ ＞ θ0Whether set up；If so, output S after then button has flashedmaxCorresponding button sign conduct
Classification results；Continue to gather next round EEG signals if invalid；
The Δ θ ＞ θ after reaching button flicker setting wheel number0Still invalid, then the characteristic vector to all wheels takes
It is average, export SmaxCorresponding button sign is used as classification results；
Accuracy and rate of information transmission are drawn according to the classification results of output；Obtain accuracy and θ0Relation figure line with
And rate of information transmission and θ0Relation figure line, then by comparing, ensureing that accuracy and system information transmissions rate all meet need
In the case of asking, one optimum value of selection is used as threshold θ0。
It is preferred that, step S7's concretely comprises the following steps：
When gathering EEG signals first, button flicker wheel number k=0 is set；After each round collection EEG signals, button is judged
Flicker wheel number k >=kminWhether set up, if not, then proceed next round eeg signal acquisition；If so, then to collection
To EEG signals handled；
The process of processing includes：Pretreatment and signal characteristic abstraction, are then averaged to the preceding k characteristic vectors taken turns, will be flat
Characteristic vector after carries out classification recurrence with Bayes classifier, obtained after classification recurrence that different buttons are corresponding to be returned
Score value μi', i represents i-th of button, by μi' be normalized and obtain corresponding normalization and return score value Si', maximum normalization
Return score value and be designated as Smax', second largest normalization returns score value and is designated as Ssec'；
Judge Smax'-Ssec' ＞ θ0Whether set up, if so, then export Smax' corresponding button sign be used as classification knot
Really；If not, then determine whether button flicker wheel number k≤kmaxWhether set up；
Proceed next round eeg signal acquisition if setting up；S is exported if invalidmax' corresponding button sign work
For classification results.
It is preferred that, system initialization includes the initialization of system hardware equipment and the initialization of relevant parameter.
It is preferred that, system hardware equipment includes eeg signal acquisition instrument and be connected with eeg signal acquisition instrument 32 are logical
Electrode on the electrode for encephalograms cap in road, electrode cap is in the electrode position of standard, before collection brain signal, in the electricity of electrode cap
Conducting resinl is injected between pole and user's brain.
It is preferred that, interactive interface includes electrical brain stimulation interface and human-computer interaction interface；Electrical brain stimulation interface includes several
Three-dimensional flash button, marks different symbols respectively, and several buttons represent several action commands respectively, and they are located at same
Row equidistantly arrangement, and stimulate user to produce the EEG signals for including P300 patterns in the way of flashing at random；Man-machine friendship
Mutual interface includes a three-dimensional street scene, and one can complete the visual human of different action commands, and be distributed in not
With the various barriers of position.
Further, the concrete mode that flashes at random is：Several three-dimensional flash buttons are often taken turns in a random order
Once, each each duration of a scintillation 0.2s of button, the gap that each button flashes is 0.2s for each flicker, between often taking turns during rest
Between at intervals of 2s.
It is preferred that, pretreatment specifically refers to EEG signals carrying out bandpass filtering and removes artefact, Ran Houyong in step S4
Amplifier is amplified to EEG signals；Signal characteristic abstraction refers specifically to be every No. four button flickers of wheel from 32 electrode channels
In each button flicker extract one section of initial characteristicses signal, after being handled through down-sampling, the signal of 32 passages is connected into one
The characteristic vector of individual P300 patterns, has the characteristic vector of several P300 patterns in so often taking turns；Classification based training is specifically referred to
The characteristic vector for testing the P300 patterns often taken turns collected is trained with Bayes classifier, obtained after classification recurrence
Score value μ is returned to several knobs are correspondingi, i represents i-th of button, by μiIt is normalized and obtains corresponding recurrence score value
Si, it is to be considered as the button that user watches attentively to return the maximum button of score value.
The present invention compared with prior art, has the following advantages that and beneficial effect：
1st, virtual reality and brain-computer interface are combined by the present invention, and visual human can be controlled to be carried out in virtual scene accordingly
Motion avoiding barrier, improve to the training effect of brain control, obtain more preferable experience sense by.
2nd, the present invention carries out online classification using self adaptation Bayes classifier, greatly improves brain control visual human's
Efficiency, significantly reduces the reaction time, control visual human more link up, with stronger experience sense and interactivity.
3rd, the present invention creates virtual scene by OSG, can flexibly assign visual human different attitude actions, and energy
Enough state of the Real Time Observation visual human in complicated virtual environment.
Brief description of the drawings
Fig. 1 is embodiment method flow diagram；
Fig. 2 is the flow chart that control stage handles EEG signals method；
Fig. 3 is interactive interface schematic diagram in off-line training step；
Fig. 4 is accuracy AmWith the relation figure line of button flicker wheel number；
Fig. 5 is accuracy and threshold θ0Relation figure line and rate of information transmission and threshold θ0Relation figure line.
Embodiment
With reference to embodiment and accompanying drawing, the present invention is described in further detail, but embodiments of the present invention are not limited
In this.
Such as Fig. 1, a kind of brain control visual human avoidance obstacle method, including off-line training step and On-line Control step；
Specifically, off-line training comprises the following steps：
S1, system initialization：System initialization includes the initialization of system hardware equipment and the initialization of relevant parameter.
System hardware equipment includes eeg signal acquisition instrument, the brain electricity electricity for 32 passages being connected with eeg signal acquisition instrument
Electrode on polar cap, electrode cap is in the electrode position of standard.Collection brain signal before, need to electrode cap electrode with using
Conducting resinl is injected between person's brain, it is ensured that electrode for encephalograms cap collects good EEG signals.To user with upper electrode cap, and
Relevant parameter is initialized.
S2, startup interactive interface as shown in Figure 3, interactive interface include electrical brain stimulation interface and human-computer interaction interface.
Electrical brain stimulation interface includes 4 three-dimensional flash buttons, and they mark " L ", " C ", " W ", " R " respectively.This four are pressed
Button represents respectively, it is couchant walk, advance and turn right, they are located at equidistantly arranges with a line, and the side to flash at random
Formula stimulates user to produce the EEG signals for including P300 patterns.The above-mentioned concrete mode that flashes at random is：4 three-dimensional flickers are pressed
Button often takes turns each flicker in a random order once, each each duration of a scintillation 0.2s of button, the gap of each button flicker
For 0.2s, the time of having a rest is at intervals of 2s between often taking turns.
Human-computer interaction interface includes three-dimensional street scene, one can turn left, it is couchant walk, advance and turn right
Visual human, and it is distributed in the various barriers of diverse location.
In off-line training step, interactive interface also includes prompt window, can show including " L ", " C ", " W ",
The different instructions of " R ".And in On-line Control step, the prompt window is no longer shown.
Then EEG signals are carried out off-line training by S3, the scalp EEG signals by electrode cap collection user.
The letter that user shows in the prompt window in interactive interface watches the corresponding button in electrical brain stimulation interface attentively,
Such as prompt window shows " L " that the eyes of user, which are just watched attentively, stimulates the button that " L " is indicated in interface.For each user
Carry out 20 groups of tests, button flicker of the every group of test bag containing 10 wheels.In one group of test, human eye stares at a character always, and 4 are pressed
Button constantly flashes, and human eye immovable eyes before one group of wheel test of test namely 10 terminates go to see other characters.
Until the flicker of this group collection terminates, eyes return the next instruction that prompt window sees prompting, repeat to walk above
Suddenly.
The EEG signals that S4, processing are collected, the process of processing includes：Pretreatment, signal characteristic abstraction and utilizes pattra leaves
This grader carries out classification based training.
Pretreatment specifically refers to EEG signals carrying out bandpass filtering and removes artefact, then with amplifier to EEG signals
It is amplified.
Signal characteristic abstraction refers specifically to be each button flicker in every No. four button flickers of wheel from 32 electrode channels
One section of initial characteristicses signal is extracted, after being handled through down-sampling, the data of 32 passages are connected into the feature of a P300 pattern
Vector, extracts the characteristic vector of four P300 patterns in so often taking turns.
Classification based training specifically refers to the characteristic vector Bayes classifier by the P300 patterns often taken turns collected are tested
It is trained, carries out obtaining the corresponding recurrence score value μ of 4 buttons after classification recurrencei, i represents i-th of button, by μiCarry out normalizing
Change obtains corresponding normalization and returns score value Si, maximum normalization returns score value and is designated as Smax, second largest normalization, which is returned, to divide
Value is designated as Ssec, SmaxCorresponding button is to be considered as the button that user watches attentively.
20 groups of training datas of user are gathered, in 20 groups of tests, average work is taken to preceding m (1≤m≤10) accuracy taken turns
The accuracy A taken turns for mm.For example, accuracy of the accuracy as the 1st wheel is calculated for 20 the 1st wheels in 20 groups, it is then right
20 the 2nd wheels calculate accuracy of the accuracy as the 2nd wheel, calculate successively.Accuracy AmPresented with the increase of number of rounds of tests
Certain change, draws accuracy AmWith the graph of a relation of button flicker wheel number, by the analysis to data, it is ensured that certain accuracy
Under the conditions of, it is the minimum and maximum flicker wheel number of control stage setting.Minimum flicker wheel number is set in the present embodiment as 3 wheels,
Maximum flicker wheel number is 8 wheels.
In order to improve systematic function, accuracy and system information transmissions rate are taken into account, the maximum normalization of definition returns score value
SmaxScore value S is returned with second largest normalizationsecDifference Smax-Ssec=Δ θ (0 ＜ Δ θ ＜ 1).To the training data of user
Ten folding cross validations are carried out, 20 group data sets are divided into 10 parts, in turn using 9 parts as training data, 1 part is used as test data.
The different Δ θ of setting threshold θ0(0 ＜ θ0＜ 1), according to steps of processing：Collect K (K≤10) wheel brains
During electric signal, the preceding K characteristic vectors taken turns are averaged, the characteristic vector after being then averaged is classified with Bayes classifier
Return, obtain the corresponding recurrence score value S of 4 buttonsi, maximum normalization returns score value and is designated as Smax, second largest normalization returns
Score value is returned to be designated as Ssec, then judge Δ θ ＞ θ0Whether set up；If so, then button exports S after having flashed 10 wheelsmaxCorrespondence
Button sign be used as classification results；
Continue to gather next round EEG signals if invalid, the Δ θ ＞ θ after the wheel of button flicker 100Still not into
Vertical, then the characteristic vector to 10 wheels is averaged, and exports SmaxCorresponding button sign is used as classification results.
Accuracy and rate of information transmission are drawn according to the classification results of output；Obtain accuracy and θ0Relation figure line with
And rate of information transmission and θ0Relation figure line, then by comparing, ensureing that accuracy and system information transmissions rate all meet need
In the case of asking, one best difference of selection returns score value S as maximum normalizationmaxReturn and divide with second largest normalization
Value SsecDifference Smax-SsecThreshold θ0。
Above-mentioned flicker wheel number and threshold θ0, the setting ginseng of on-line training stage self adaptation Bayes classifier will be used as
Number.
After off-line training terminates, into control stage, specifically comprise the following steps：
S5, reinitialize relevant parameter, start interactive interface, mode and off-line training that this stage button flashes at random
Random flashing mode it is the same, but this stage prompt window will not be to corresponding prompting.
Which button S6, user watch attentively when deciding flicker in its sole discretion according to Obstacle Position and visual human position, simultaneously
Electrode cap gathers the EEG signals of user.Corresponding button is watched in the action that user wants to complete according to oneself attentively always, directly
Visual human has action to realize on to interactive interface, is then transferred into and watches next button attentively.
S7, collect after EEG signals, EEG signals are handled.
When gathering EEG signals first, button flicker wheel number k=0 is set.After each round collection EEG signals, button is judged
Whether flicker wheel number k >=3 set up, if not, then proceed next round eeg signal acquisition；If so, then to collecting
EEG signals handled.
The process of processing includes：Pretreatment and signal characteristic abstraction, the step of this step is with off-line training step are identical.It is right
EEG signals are carried out after pretreatment and signal characteristic abstraction, and the preceding k characteristic vectors taken turns are averaged, then will it is average after spy
Levy vector and classification recurrence is carried out with Bayes classifier, carry out obtaining the corresponding recurrence score value μ of 4 buttons after classification recurrencei', i
I-th of button is represented, by μi' be normalized and obtain corresponding normalization and return score value Si', maximum normalization returns score value
It is designated as Smax', second largest normalization returns score value and is designated as Ssec'。
Judge Smax'-Ssec' ＞ θ0Whether set up, if so, then export Smax' corresponding button sign be used as classification knot
Really；If not, then determine whether whether button flicker wheel number k≤8 set up；
If button flicker wheel number k≤8 are set up, proceed next round eeg signal acquisition；Exported if invalid
Smax' corresponding button sign is used as classification results.
S8, according to the classification results of output generate control command, control motion of virtual human.Pass through OSG (Open Scene
Graph virtual scene) is created, the complicated motion of personage and the change of virtual environment is flexibly realized.
So far, processing procedure terminates, and button flicker wheel number k=0, repeat step S6-S8 is then made, until user's control
Visual human completes task, collection EEG signals cut-off.
Above-described embodiment is preferably embodiment, but embodiments of the present invention are not by above-described embodiment of the invention
Limitation, other any Spirit Essences without departing from the present invention and the change made under principle, modification, replacement, combine, simplification,
Equivalent substitute mode is should be, is included within protection scope of the present invention.
Claims (10)
1. a kind of brain control visual human avoidance obstacle method, it is characterised in that including off-line training step and On-line Control step：
A off-line trainings：
S1, system initialization；
S2, startup interactive interface, interactive interface show the button and prompt window flashed at random；
S3, user watch the button flashed accordingly attentively according to the instruction of prompt window, and the brain of user is gathered by electrode cap
Electric signal；
The EEG signals that S4, processing are collected, including：Pretreatment, signal characteristic abstraction and is divided using Bayes classifier
Class is trained, and is obtained corresponding normalize of different buttons and is returned score value, it is to be considered as user that normalization, which returns the maximum button of score value,
The button watched attentively；
It is determined that the threshold value of the maximum difference that score value is returned with second largest normalization, and maximum button flicker wheel number, minimum button
Flicker wheel number as on-line training stage self adaptation Bayes classifier setup parameter；
B On-line Controls：
S5, reinitialize relevant parameter, start interactive interface, prompt window is no longer shown；
S6, user decide the button for watching flicker attentively in its sole discretion according to Obstacle Position and visual human position, while electrode cap is gathered
The EEG signals of user；
S7, collect after EEG signals, handle the EEG signals collected, including：Pretreatment, signal characteristic abstraction and utilization are certainly
Adapt to Bayes classifier and carry out classification based training；
S8, according to the classification results of output generate control command, control motion of virtual human.
2. brain control visual human avoidance obstacle method according to claim 1, it is characterised in that pre-processed in step S4 specific
Refer to EEG signals carrying out bandpass filtering and remove artefact, then EEG signals are amplified with amplifier；Signal characteristic
Extract and refer specifically to be each button flicker one section of initial characteristics of extraction in the multiple button flicker of every wheel from 32 electrode channels
Signal, after being handled through down-sampling, the signal of 32 passages is connected into the characteristic vector of a P300 pattern, meeting in so often taking turns
There is the characteristic vector of multiple P300 patterns；Classification based training specifically refer to test the feature of the P300 patterns often taken turns that collects to
Amount is trained with Bayes classifier, carries out obtaining the corresponding recurrence score value μ of multiple knobs after classification recurrencei, i represents i-th
Individual button, by μiIt is normalized and obtains corresponding recurrence score value Si, it is to be considered as what user watched attentively to return the maximum button of score value
Button.
3. brain control visual human avoidance obstacle method according to claim 1, it is characterised in that each in off-line training step
User carries out 20 groups of tests, button flicker of the every group of test bag containing 10 wheels.
4. brain control visual human avoidance obstacle method according to claim 1, it is characterised in that determine that maximum is pressed in step S4
Button flicker wheel number, the method for minimum button flicker wheel number are：Multigroup test is carried out, every group of test bag flashes containing many wheel buttons, so
The average accuracy A taken turns as m is taken to the preceding m accuracy taken turns afterwardsm；Accuracy AmAs the increase of number of rounds of tests is presented certain
Change, draws accuracy AmWith the graph of a relation of button flicker wheel number, by analysis, it is ensured that selected most under conditions of certain accuracy
Big button flicker wheel number kmaxWith minimum button flicker wheel number kmin。
5. brain control visual human avoidance obstacle method according to claim 1, it is characterised in that minimum button flicker wheel number is
3, maximum button flicker wheel number is 8.
6. brain control visual human avoidance obstacle method according to claim 2, it is characterised in that determined in step S4 it is maximum with
The threshold method of difference that second largest normalization returns score value is：
When collecting K wheel EEG signals, the preceding K characteristic vectors taken turns are averaged, the characteristic vector after being then averaged uses shellfish
This grader of leaf carries out classification recurrence, and the different buttons of acquisition are corresponding to return score value Si, maximum normalization returns score value and is designated as
Smax, second largest normalization returns score value and is designated as Ssec, define Smax-Ssec=Δ θ (0 ＜ Δ θ ＜ 1), sets different Δ θ's
Threshold θ0；
Judge Δ θ ＞ θ0Whether set up, if so, output S after then button has flashedmaxCorresponding button sign is tied as classification
Really；Continue to gather next round EEG signals if invalid；
The Δ θ ＞ θ after reaching button flicker setting wheel number0Still invalid, then the characteristic vector to all wheels is averaged,
Export SmaxCorresponding button sign is used as classification results；
Accuracy and rate of information transmission are drawn according to the classification results of output；Obtain accuracy and θ0Relation figure line and letter
Cease transfer rate and θ0Relation figure line, then by comparing, ensureing that accuracy and system information transmissions rate all meet demand
In the case of, one optimum value of selection is used as threshold θ0。
7. brain control visual human avoidance obstacle method according to claim 3, it is characterised in that step S7 specific steps
For：
When gathering EEG signals first, button flicker wheel number k=0 is set；After each round collection EEG signals, judge that button flashes
Take turns number k >=kminWhether set up, if not, then proceed next round eeg signal acquisition；If so, then to collecting
EEG signals are handled；
The process of processing includes：Pretreatment and signal characteristic abstraction, are then averaged to the preceding k characteristic vectors taken turns, after average
Characteristic vector carry out classification recurrence with Bayes classifier, obtained after classification recurrence the corresponding recurrence score value of different buttons
μi', i represents i-th of button, by μi' be normalized and obtain corresponding normalization and return score value Si', maximum normalization is returned
Score value is designated as Smax', second largest normalization returns score value and is designated as Ssec'；
Judge Smax'-Ssec' ＞ θ0Whether set up, if so, then export Smax' corresponding button sign is used as classification results；If no
Set up, then determine whether button flicker wheel number k≤kmaxWhether set up；
If k≤kmaxEstablishment then proceeds next round eeg signal acquisition；S is exported if invalidmax' corresponding button sign
It is used as classification results.
8. brain control visual human avoidance obstacle method according to claim 1, it is characterised in that system initialization includes system
The initialization of hardware device and the initialization of relevant parameter；System hardware equipment includes eeg signal acquisition instrument and electric with brain
Electrode on the electrode for encephalograms cap of 32 connected passages of signal sampler, electrode cap is in the electrode position of standard, in collection
Before brain signal, conducting resinl is injected between the electrode and user's brain of electrode cap.
9. brain control visual human avoidance obstacle method according to claim 1, it is characterised in that interactive interface includes brain electricity thorn
Swash interface and human-computer interaction interface；Electrical brain stimulation interface includes several three-dimensional flash buttons, and different symbols are marked respectively, if
A dry button represents several action commands respectively, and they are located at equidistantly arranges with a line, and in the way of flashing at random
User is stimulated to produce the EEG signals for including P300 patterns；Human-computer interaction interface includes a three-dimensional street scene, one
It is individual to complete the visual human of different action commands, and it is distributed in the various barriers of diverse location.
10. brain control visual human avoidance obstacle method according to claim 1, it is characterised in that the tool flashed at random
Body mode is：Several three-dimensional flash buttons often take turns each flicker in a random order once, when each button flashes lasting every time
Between 0.2s, the gap of each button flicker is 0.2s, and the time of having a rest is at intervals of 2s between often taking turns.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710088805.XA CN106940592B (en) | 2017-02-20 | 2017-02-20 | A kind of brain control visual human avoidance obstacle method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710088805.XA CN106940592B (en) | 2017-02-20 | 2017-02-20 | A kind of brain control visual human avoidance obstacle method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106940592A true CN106940592A (en) | 2017-07-11 |
| CN106940592B CN106940592B (en) | 2019-10-18 |
Family
ID=59468593
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710088805.XA Active CN106940592B (en) | 2017-02-20 | 2017-02-20 | A kind of brain control visual human avoidance obstacle method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106940592B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112099619A (en) * | 2020-08-11 | 2020-12-18 | 东南大学 | Time window length self-adaptive selection method of mixed sight brain-computer interface |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103955270A (en) * | 2014-04-14 | 2014-07-30 | 华南理工大学 | Character high-speed input method of brain-computer interface system based on P300 |
| CN103955269A (en) * | 2014-04-09 | 2014-07-30 | 天津大学 | Intelligent glass brain-computer interface method based on virtual real environment |
| CN104965584A (en) * | 2015-05-19 | 2015-10-07 | 西安交通大学 | Mixing method for brain-computer interface based on SSVEP and OSP |
| CN106406534A (en) * | 2016-09-28 | 2017-02-15 | 雷嘉兴 | Virtual reality game design technology involving brain waves in control |
-
2017
- 2017-02-20 CN CN201710088805.XA patent/CN106940592B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103955269A (en) * | 2014-04-09 | 2014-07-30 | 天津大学 | Intelligent glass brain-computer interface method based on virtual real environment |
| CN103955270A (en) * | 2014-04-14 | 2014-07-30 | 华南理工大学 | Character high-speed input method of brain-computer interface system based on P300 |
| CN104965584A (en) * | 2015-05-19 | 2015-10-07 | 西安交通大学 | Mixing method for brain-computer interface based on SSVEP and OSP |
| CN106406534A (en) * | 2016-09-28 | 2017-02-15 | 雷嘉兴 | Virtual reality game design technology involving brain waves in control |
Non-Patent Citations (2)
| Title |
|---|
| 刘甜，李远清，王洪涛，胡友盼: "《基于超声波传感器的脑控轮椅避障系统的研究》", 《计算机测量与控制》 * |
| 陆盛林: "《基于脑机接口的虚拟现实三维目标的控制研究及应用》", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112099619A (en) * | 2020-08-11 | 2020-12-18 | 东南大学 | Time window length self-adaptive selection method of mixed sight brain-computer interface |
| CN112099619B (en) * | 2020-08-11 | 2024-02-06 | 东南大学 | Time window length self-adaptive selection method for mixed vision brain-computer interface |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106940592B (en) | 2019-10-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109992113A (en) | A kind of MI-BCI system and its control method induced based on more scenes | |
| Kus et al. | Asynchronous BCI based on motor imagery with automated calibration and neurofeedback training | |
| CN103955269A (en) | Intelligent glass brain-computer interface method based on virtual real environment | |
| CN104360730B (en) | Man-machine interaction method supported by multi-modal non-implanted brain-computer interface technology | |
| Pinheiro et al. | Wheelchair simulator game for training people with severe disabilities | |
| CN108446020A (en) | Merge Mental imagery idea control method and the application of Visual Graph and deep learning | |
| CN107433021A (en) | A kind of VR rehabilitation systems based on mirror neuron | |
| CN107308638A (en) | A kind of entertaining rehabilitation training of upper limbs system and method for virtual reality interaction | |
| CN102866775A (en) | System and method for controlling brain computer interface (BCI) based on multimode fusion | |
| CN103955270B (en) | Character high-speed input method of brain-computer interface system based on P300 | |
| CN109310561A (en) | Move teaching system and movement teaching method | |
| CN104548347A (en) | Pure idea nerve muscle electrical stimulation control and nerve function evaluation system | |
| CN106774847A (en) | 3D vision P300 Speller systems based on virtual reality technology | |
| Legény et al. | Navigating in virtual worlds using a self-paced SSVEP-based brain–computer interface with integrated stimulation and real-time feedback | |
| CN103258215A (en) | Multi-lead correlation analysis electroencephalo-graph (EEG) feature extraction method | |
| CN110442244A (en) | A kind of reality-virtualizing game exchange method and system based on brain-computer interface | |
| CN112465059A (en) | Multi-person motor imagery identification method based on cross-brain fusion decision and brain-computer system | |
| Chen et al. | CreativeBioMan: a brain-and body-wearable, computing-based, creative gaming system | |
| CN112244774A (en) | Brain-computer interface rehabilitation training system and method | |
| CN103300852A (en) | Training method for controlling remotely-controlled trolley in electrocerebral way based on motor imagery | |
| CN107479696A (en) | Based on P300 normal form virtual reality brain machine interface systems and implementation method | |
| CN117389441B (en) | Writing imagination Chinese character track determining method and system based on visual following assistance | |
| CN107544675A (en) | Brain control formula virtual reality method | |
| CN109799909A (en) | A kind of olfactory analog system and method based on Virtual Reality | |
| CN107122050A (en) | Stable state of motion VEP brain-machine interface method based on CSFL GDBN |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |