CN106507356A - A kind of wireless authentication method and its system - Google Patents
A kind of wireless authentication method and its system Download PDFInfo
- Publication number
- CN106507356A CN106507356A CN201611169078.1A CN201611169078A CN106507356A CN 106507356 A CN106507356 A CN 106507356A CN 201611169078 A CN201611169078 A CN 201611169078A CN 106507356 A CN106507356 A CN 106507356A
- Authority
- CN
- China
- Prior art keywords
- action
- network
- cutting
- wireless
- mobile terminal
- 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.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W12/00—Security arrangements; Authentication; Protecting privacy or anonymity
- H04W12/06—Authentication
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W12/00—Security arrangements; Authentication; Protecting privacy or anonymity
- H04W12/08—Access security
Landscapes
- Engineering & Computer Science (AREA)
- Computer Security & Cryptography (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention provides a kind of wireless authentication method and its system, wherein, include in the wireless authentication method：S1 mobile terminals detect auto- physical data；S2 mobile terminals carry out action cutting to the physical data for obtaining；S3 mobile terminals are classified to the action that cutting is obtained jointly with wireless apss；S4 wireless apss obtain corresponding password to be certified and carry out wireless authentication to which according to the result of classification.So, user is during wireless authentication is carried out, it is no longer necessary to which mode manually input password to be certified, only needs whipping mobile terminal, allows mobile terminal that deliberate action occurs, simple and convenient, provides the user facility, improves Consumer's Experience.
Description
Technical field
The present invention relates to communication technical field, more particularly to a kind of wireless authentication method and its system.
Background technology
Wireless technology is a kind of short range wireless transmission technology, and as its transmission speed is very fast, coverage is also long,
Therefore it is widely used in office space and family.The terminal with radio function accesses wireless apss (Access Point, access
Point) after, wireless terminal is authenticated by wireless apss, after certification passes through, it is allowed to which wireless terminal accesses the Internet.
At present, wireless apss are mainly the authentication mode of user name+password, i.e. wireless terminal to the authentication mode of wireless terminal
When accessing wireless apss, report for carrying out the username and password of authentication to wireless apss, if wireless apss are to the use that receives
Name in an account book and cipher authentication pass through, then it represents that the wireless terminal has the authority using wireless network, it is allowed to which which accesses the Internet；No
Then, represent that the wireless terminal does not have the authority using wireless network, limit which and access the Internet.In some home wireless systems
In, wireless terminal only need to report password to wireless apss, and be received according to locally stored user name by wireless router
Password is authenticated.But this authentication mode needs user to be manually entered password, relatively complicated, and Consumer's Experience is poor.
Content of the invention
For the problems referred to above, the invention provides a kind of wireless authentication method and its system, efficiently solve existing wireless
Relatively complicated technical problem during authentication method use.
The technical scheme that the present invention is provided is as follows：
A kind of wireless authentication method, the mobile terminal and wireless apss including communication connection, wraps in the wireless authentication method
Include：
S1 mobile terminals detect auto- physical data；
S2 mobile terminals carry out action cutting to the physical data for obtaining；
S3 mobile terminals are classified to the action that cutting is obtained jointly with wireless apss；
S4 wireless apss obtain corresponding password to be certified and carry out wireless authentication to which according to the result of classification.
In the technical program, mobile terminal is split to the action of user's whipping/mobile terminal, and combines wireless apss
Automatic decision goes out password to be certified, carries out wireless authentication with this wireless aps according to the password to be certified, if comparing successfully, completes
Wireless authentication, it is allowed to which user accesses.So, user is during wireless authentication is carried out, it is no longer necessary to mode manually
Password to be certified is input, whipping mobile terminal is only needed, allows mobile terminal that deliberate action occurs, simple and convenient, it is that user carries
For facility, Consumer's Experience is improved.
It is further preferred that also including before step S1：
S01 sets the directioin parameter of deliberate action；
S02 sets the threshold parameter of deliberate action, and the threshold parameter includes：Acceleration during mobile terminal generation action
Degree, angular velocity and displacement；
S03 sets up action cutting training network, and which is had according to the parameter set in step S01 and step S02
Supervised training, the action cutting training network include：Data Layer, the first full articulamentum, active coating, the second full articulamentum with
And Euclidean loss layer；
S04 sets up posture prediction network and carries out Training to which, and the posture prediction network is a depth net
Network；
The posture for training prediction network split is that prime predicts network and second prediction network by S05；
S06 sets the incidence relation between deliberate action and password to be certified, wherein, action correspondingly at least
Password to be certified.
In the technical program, before wireless authentication is carried out, first mobile terminal and wireless apss are configured, including
The relevant parameter and the incidence relation between deliberate action and password to be certified of deliberate action is set, action cutting training net is set up
According to configuration information, network and posture prediction network etc., judge that mobile terminal there occurs deliberate action with this mobile terminal, carry
Foundation for wireless authentication.
It is further preferred that step S1 is specially：Mobile terminal action is detected by the built-in sensor of mobile terminal
Physical data in journey；
Step S2 is specially：Net is implemented in the action cutting that the physical data input that step S1 is obtained by mobile terminal is trained
Network, realizes carrying out action cutting to the physical data for obtaining, and wherein, network is implemented in the action cutting to be included：Data Layer,
One full articulamentum, active coating and the second full articulamentum；
Step S3 is specially：By cut in step S2 obtain action data input mobile terminal in prime prediction
Network obtains eigenvalue and sends it to wireless apss, and the eigenvalue for obtaining input second is predicted real-time performance by wireless apss
Classification to action.
In the technical program, physical data of the enforcement network to getting is cut using action in mobile terminal and enters action
Cut, the action data input prime prediction network that afterwards cutting is obtained；Take up the ball prime prediction network meter
With this, the eigenvalue input second prediction network for obtaining, realizes that action is sorted out, substantially increases in mobile terminal
The accuracy judged by itself action.In addition, during being communicated between mobile terminal and wireless apss, mobile terminal will be counted
The eigenvalue for calculating is sent to wireless apss, is calculated by the action data that cutting is obtained by eigenvalue, is possessed uncertainty, no
Easily forge, during directly send to wireless apss for, greatly improve data transmission procedure
In security performance.
It is further preferred that specifically including in step s 4：
S41 wireless apss are according in the incidence relation between the deliberate action and password to be certified set in S05 and step S3
The action that classification is obtained obtains password to be certified accordingly；
The password to be certified for receiving is compared by S42 wireless apss with preset password, completes wirelessly recognizing for mobile terminal
Card.
It is further preferred that cutting in training network and action cutting enforcement network in action, the active coating is ReLU
(Rectified linear unit) layer；And/or,
In posture prediction network, the posture prediction network includes：Data Layer, multiple convolutional layers and full articulamentum
Multilamellar convolutional neural networks, also include SoftMax graders；The SoftMax graders are according to multilamellar convolutional neural networks
The eigenvalue of output is classified to action.
Present invention also offers a kind of wireless authentication system, including the wireless apss of communication of mobile terminal connection, wherein,
Mobile terminal, for detecting auto- physical data, carrying out action cutting for the physical data to obtaining
And the action that cutting is obtained is classified for associated wireless AP；
Wireless apss are additionally operable to obtain corresponding password to be certified and carry out wireless authentication to which according to the result of classification.
In the technical program, mobile terminal is split to the action of user's whipping/mobile terminal, and combines wireless apss
Automatic decision goes out password to be certified, carries out wireless authentication with this wireless aps according to the password to be certified, if comparing successfully, completes
Wireless authentication, it is allowed to which user accesses.So, user is during wireless authentication is carried out, it is no longer necessary to mode manually
Password to be certified is input, whipping mobile terminal is only needed, allows mobile terminal that deliberate action occurs, simple and convenient, it is that user carries
For facility, Consumer's Experience is improved.
It is further preferred that the mobile terminal includes：
Data acquisition module, for obtaining the physical data in mobile terminal course of action；
Action cutting module, the physical data for getting to data acquisition module carry out action cutting；
First classification of motion module, obtains eigenvalue for carrying out process to the action that the cutting of action cutting module is obtained；
Data transmission blocks, for sending the eigenvalue that the first classification of motion module is obtained to wireless apss；
The wireless apss include：
Information receiving module, for receiving the eigenvalue of mobile terminal transmission；
Second classification of motion module, the eigenvalue for being received according to information receiving module are classified to which；
Password acquisition module, for obtaining corresponding password to be certified according to the result of the second classification of motion module classification；
Password authentication module, for being authenticated to the password to be certified that password acquisition module is obtained.
It is further preferred that also including in the terminal：
First configuration module, for setting the incidence relation between deliberate action and password to be certified, setting deliberate action
Directioin parameter and set the threshold parameter of deliberate action, the threshold parameter includes：Adding during mobile terminal generation action
Speed, angular velocity and displacement；
Network training module, for setting up action cutting training network, and according to the parameter set in configuration module to which
Training is carried out, the action cutting training network includes：Data Layer, the first full articulamentum, active coating, second connect entirely
Connect layer and Euclidean loss layer；And for setting up posture prediction network and carrying out Training to which, the pre- survey grid of the posture
Network is a depth network；
Network cutting module, the posture prediction network split for training network training module are predicted for prime
Network and second prediction network.
In the technical program, before wireless authentication is carried out, first mobile terminal and wireless apss are configured, including
The relevant parameter and the incidence relation between deliberate action and password to be certified of deliberate action is set, action cutting training net is set up
According to configuration information, network and posture prediction network etc., judge that mobile terminal there occurs deliberate action with this mobile terminal, carry
Foundation for wireless authentication.
It is further preferred that in action cutting module, the physical data input that data acquisition module is obtained is trained
Action cutting implement network, realize carrying out action cutting to the physical data for obtaining, wherein, network is implemented in action cutting
Include：Data Layer, the first full articulamentum, active coating and the second full articulamentum；
In the first classification of motion module, for the action data input mobile terminal for obtaining the cutting of action cutting module
In prime prediction network obtain eigenvalue；
In the second classification of motion module, for the eigenvalue for obtaining input second prediction real-time performance is cut
The classification of the action that arrives.
In the technical program, physical data of the enforcement network to getting is cut using action in mobile terminal and enters action
Cut, the action data input prime prediction network that afterwards cutting is obtained；Take up the ball prime prediction network meter
With this, the eigenvalue input second prediction network for obtaining, realizes that action is sorted out, substantially increases in mobile terminal
The accuracy judged by itself action.In addition, during being communicated between mobile terminal and wireless apss, mobile terminal will be counted
The eigenvalue for calculating is sent to wireless apss, is calculated by the action data that cutting is obtained by eigenvalue, is possessed uncertainty, no
Easily forge, during directly send to wireless apss for, greatly improve data transmission procedure
In security performance.
It is further preferred that cutting in training network and action cutting enforcement network in action, the active coating is ReLU
Layer；And/or,
In posture prediction network, the posture prediction network includes：Data Layer, multiple convolutional layers and full articulamentum
Multilamellar convolutional neural networks, also include SoftMax graders；The SoftMax graders are according to multilamellar convolutional neural networks
The eigenvalue of output is classified to action.
In the technical program, after posture prediction network has carried out feature extraction to the physical data for getting, allow
Multilamellar convolutional neural networks oneself study select input combinations of features obtain characteristic vector to grader classify best to obtain
Classifying quality.Compared to traditional artificial selection's input feature vector, obtain after combination after output characteristic vector to point
Class device carries out classifying more intelligent, realizes more accurately classifying quality.And feature is carried in the multilamellar convolutional neural networks
Take part and Classification and Identification part is engaged togather, wherein feature learning part implicit expression is carried out, be largely focused on the multilamellar volume
Middle the hidden layer of product neutral net, i.e., while training training sample set, carry out feature learning, eliminate loaded down with trivial details manual
The process of feature and design feature is extracted, flow process is simplified, the time is saved.
Description of the drawings
Below by the way of clearly understandable, preferred implementation is described with reference to the drawings, to above-mentioned characteristic, technical characteristic,
Advantage and its implementation are further described.
Fig. 1 is a kind of embodiment schematic flow sheet of wireless authentication method in the present invention；
Fig. 2 is wireless authentication method another embodiment schematic flow sheet in the present invention；
Fig. 3 is wireless authentication system schematic diagram in the present invention.
Fig. 4 is mobile terminal schematic diagram in the present invention；
Fig. 5 is wireless apss schematic diagram in the present invention.
Drawing reference numeral explanation：
100- wireless authentication systems, 110- mobile terminals, 120- wireless apss, 111- data acquisition modules, 112- actions are cut
Cut module, 113- first classification of motion modules, 114- data transmission blocks, 121- information receiving modules, the second actions of 122- point
Generic module, 123- password acquisition modules, 124- password authentication modules.
Specific embodiment
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below by control description of the drawings
The specific embodiment of the present invention.It should be evident that drawings in the following description are only some embodiments of the present invention, for
For those of ordinary skill in the art, on the premise of not paying creative work, can be obtaining other according to these accompanying drawings
Accompanying drawing, and obtain other embodiments.
The wireless authentication method of present invention offer is provided, specifically includes the mobile terminal of communication connection and wireless
AP, it can be seen that include in the wireless authentication method：S1 mobile terminals detect auto- physical data；S2
Mobile terminal carries out action cutting to the physical data for obtaining；S3 mobile terminals are entered to the action that cutting is obtained jointly with wireless apss
Row classification；S4 wireless apss obtain corresponding password to be certified and carry out wireless authentication to which according to the result of classification.
Specifically, in the wireless authentication method, mobile terminal is split to the action of user's whipping/mobile terminal, and
Associated wireless AP automatic decisions go out password to be certified, carry out wireless authentication with this wireless aps according to the password to be certified, if comparing
Success, then complete wireless authentication, it is allowed to which user accesses.More particularly, above-mentioned mobile terminal can be smart mobile phone, flat board electricity
Brain etc., is not specifically limited here.
Based on this, also included before step S1：S01 sets the directioin parameter of deliberate action；S02 sets deliberate action
Threshold parameter, threshold parameter include：Acceleration, angular velocity and displacement during mobile terminal generation action；S03 sets up dynamic
Make cutting training network, and Training is carried out to which according to the parameter set in step S01 and step S02, action is cut
Training network includes：Data Layer, the first full articulamentum, active coating, the second full articulamentum and Euclidean loss layer；S04 sets up
Posture prediction network simultaneously carries out Training to which, and posture prediction network is a depth network；S05 will be pre- for the posture for training
Survey grid network is split as prime prediction network and second prediction network；S06 set deliberate action and password to be certified it
Between incidence relation, wherein, corresponding at least one password to be certified of action.
Specifically, for deliberate action, can be that the action of whipping can also be once closed for mobile action, such as left and right whipping
Connection numeral 1, upper and lower whipping once associates numeral 0；And for example, mobile circle association digital 2 clockwise, counterclockwise mobile circle association
Numeral 3 etc..Further, an action can associate a code to be certified, it is also possible to associate two codes to be certified, according to practical situation
Set；Such as, left and right whipping once associates numeral 01, and upper and lower whipping once associates numeral 23 etc., and here is not specifically limited.
For the directioin parameter of deliberate action, in one embodiment, with mobile terminal display screen as coordinate axess, laterally
For x-axis, longitudinal direction is y-axis, and vertical screen is z-axis, will be set as left and right whipping, association numeral 1 along x-axis whipping；To get rid of along y-axis
It is set as upper and lower whipping, association numeral 0；By mobile circle association digital 2 clockwise in x/y plane；To be moved in x/y plane counterclockwise
Dynamic circle association numeral 3 etc., by that analogy.
For the threshold parameter of deliberate action, at least include：Acceleration, angular velocity during mobile terminal generation action and
Displacement.In a kind of specific embodiment, the displacement threshold value for setting mobile terminal or so whipping/upper and lower whipping sets
It is 0.1m/s for 0.2m (rice), acceleration2, by that analogy.It is noted that we are to above-mentioned default incidence relation, side here
It is not especially limited to parameter and threshold parameter, can be set according to practical situation.In actual applications, by moving
The accelerometer of dynamic terminal built-in and gyroscope obtain the physical data of mobile terminal whipping/movement.
For training network is cut in action, including but not limited to：Data Layer, the first full articulamentum, active coating, second connect entirely
Connect layer and Euclidean loss layer (Euclidean Loss layers).When the training for having supervision is carried out to the action slicing network, will
Initial data input data layer, is returned using label (given data) in Euclidean Loss layers, with this, by instruction
Practice so that the error (loss) of whole action cutting training network is minimum.Specifically, in this process, input is original
Data can be arbitrary data, naturally it is also possible to be the physical data of the action of the mobile terminal arrived used in the present invention；?
The label that Euclidean Loss layers are used is the correct cutting result of the initial data of input.In an example, it is assumed that should
Action cutting training network supports 4 alphabetical combinations, i.e. then cut point position 3, if the data of input are write manually for user
ABCD, then the label that Euclidean Loss layers are used be correct A, B, C, D ways of writing, with this to the action
Cutting training network is trained, and until error is minimum, completes to train.Finally, it is to be noted that, above-mentioned active coating is ReLU
Layer, certainly can also be other activation primitives, according to practical situation depending on.
For posture predicts network, which is specially a depth network.In an example, wrap in the posture prediction network
Include：The multilamellar convolutional neural networks of data Layer, multiple convolutional layers and full articulamentum, also include SoftMax graders；
SoftMax graders are classified to action according to the eigenvalue that multilamellar convolutional neural networks are exported.Predicted using the posture
Training is carried out to which before network, to improve the accuracy that the posture predicts network final classification.Specifically, in training
During, the initial data of input can be arbitrary data, naturally it is also possible to be the dynamic of the mobile terminal that arrives used in the present invention
The physical data of work, does not specifically limit.After predicting network training well the posture, prime prediction is split as
Network and second prediction network, wherein, prime prediction network is placed in mobile terminal with action training slicing network,
Second prediction network is placed in wireless apss.Predict that network and second predict the journey of network split as prime
Degree, we are not specifically limited here, according to practical situation depending on, e.g., in an example, prime prediction network serve as reasons
The multilamellar convolutional neural networks that data Layer, multiple convolutional layers and full articulamentum are constituted, second prediction network is SoftMax
Grader.
This is based on, the wireless authentication method is specially：By the built-in sensor of mobile terminal, S1 detects that mobile terminal is moved
Physical data during work；Net is implemented in the action cutting that the physical data input that step S1 is obtained by S2 mobile terminals is trained
Network, realizes carrying out action cutting to the physical data for obtaining, and wherein, network is implemented in the action cutting to be included：Data Layer, first
Full articulamentum, active coating and the second full articulamentum；S3 by cut in step S2 obtain action data input mobile terminal in
Prime prediction network obtain eigenvalue and send it to wireless apss, wireless apss by the eigenvalue for obtaining be input into the second appearance
Gesture predicts classification of the real-time performance to action；S4 wireless apss obtain corresponding password to be certified and which are entered according to the result of classification
Row wireless authentication.
Specifically, in the course of the work, after the built-in sensor of mobile terminal detects physical data, action is inputted
Network is implemented in cutting, obtains the cut point between each action in the second full articulamentum, obtains each action；Afterwards, each by obtained
The parameter input prime prediction network of action obtains eigenvalue；Afterwards, the eigenvalue for obtaining is sent to wireless apss；Wirelessly
AP receives eigenvalue, is inputted the prediction probability that second prediction network exports each action classification, and maximum predicted is general
The classification of rate obtains final result as the classification for predicting；Afterwards, obtained according to the incidence relation of prediction to be certified accordingly
Password is simultaneously authenticated to which.
Above-mentioned embodiment is improved and obtains present embodiment, as shown in Fig. 2 in this embodiment, this is wireless
Authentication method includes：S1 mobile terminals detect auto- physical data；S2 mobile terminals enter to the physical data for obtaining
Action is cut；S3 mobile terminals are classified to the action that cutting is obtained jointly with wireless apss；S4 wireless apss are according to classification
As a result obtain corresponding password to be certified and wireless authentication is carried out to which；S41 wireless apss according in S05 set deliberate action with
The action for obtaining of classifying in incidence relation and step S3 between password to be certified obtains password to be certified accordingly；S42 is wireless
The password to be certified for receiving is compared by AP with preset password, completes the wireless authentication of mobile terminal.
Specifically, in the course of the work, after the built-in sensor of mobile terminal detects physical data, action is inputted
Network is implemented in cutting, obtains the cut point between each action in the second full articulamentum, obtains each action；Afterwards, each by obtained
The parameter input prime prediction network of action obtains eigenvalue；Afterwards, the eigenvalue for obtaining is sent to wireless apss；Wirelessly
AP receives eigenvalue, is inputted the prediction probability that second prediction network exports each action classification, and maximum predicted is general
The classification of rate obtains final result as the classification for predicting；Finally, wireless apss are obtained according to the incidence relation of prediction accordingly
Which is simultaneously compared by password to be certified with preset password, if the password to be certified for obtaining is consistent with preset password, is compared into
Work(, completes wireless authentication, it is allowed to user access network；Otherwise, authentification failure, does not allow user access network.By this process
In, between mobile terminal and wireless apss, transmission is eigenvalue, in fact action number of the prime prediction network according to well cutting
According to calculating, actual physical significance is had no, it is difficult to forge, improve the security performance in data transmission procedure with this；In addition,
The characteristic value data amount of transmission is very little, will not consume Radio Resource.
In one example, it is assumed that wireless cipher is 0011, and numeral 0 associate left and right whipping smart mobile phone, numeral 1 association on
Lower whipping smart mobile phone.In posture prediction network, left and right whipping is set as 0 class, upper and lower whipping is set as 1 class, by which
His posture is set as 2 classes.
It is mounted with that one can be authenticated to itself whipping action before wireless authentication is carried out on smart mobile phone
Terminal applies；After user carries smart mobile phone entrance wireless zone, the wireless switching of smart mobile phone is opened, nothing in association
Line AP.Afterwards, user successively up and down whipping smart mobile phone twice, left and right whipping smart mobile phone twice.
Application APP in smart mobile phone detects the action of user by built-in accelerometer and gyroscope, and will detection
To physical data input action cutting implement network physical data is cut；Afterwards, by the action of well cutting input the
One posture prediction network obtains corresponding eigenvalue and is sent to wireless apss, and wireless apss are inputted after receiving eigenvalue
Second predicts real-time performance to the classification to action.If checking classification 2, authentification failure；If the classification for checking is all
It is 0 and 1, then by the classification composition sequence for predicting, obtains password to be certified.
After obtaining password to be certified, which is compared by wireless apss with preset password, if comparing successfully, certification is described
Success, wireless apss are let pass the smart mobile phone, it is allowed to which which connects the Internet；Otherwise, authentification failure, it is desirable to which user re-starts and recognizes
Card.
It is illustrated in figure 3 the present invention and wireless authentication system is provided, it can be seen that in the wireless authentication system 100
Include the wireless apss 120 of the communication connection of mobile terminal 1 10, wherein, mobile terminal 1 10, for detecting auto- physics
Data, action cutting is carried out for the physical data to obtaining and the action that obtains of cutting is entered for associated wireless AP120
Row classification；Wireless apss 120 are additionally operable to obtain corresponding password to be certified and carry out wireless authentication to which according to the result of classification.
In the course of the work, mobile terminal 1 10 carries out action cutting to which after detecting auto- physical data, it
Afterwards, combining wireless AP120 classifies to the action that cutting is obtained, and obtains corresponding password to be certified according to the result of classification
And wireless authentication is carried out to which.
More particularly, as shown in figure 4, including in mobile terminal 1 10：Data acquisition module 111, action cutting module
112nd, the first classification of motion module 113 and data transmission blocks 114, wherein, action cutting module 112 and data acquisition module
111 connections, the first classification of motion module 113 are connected with action cutting module 112, and data transmission blocks 114 and the first action divide
Generic module 113 connects.In the course of the work, first, obtained in 10 course of action of mobile terminal 1 by data acquisition module 111
Physical data；Afterwards, action cutting module 112 carries out action cutting to the physical data that data acquisition module 111 gets；
Then, the first classification of motion module 113 carries out process to the action that the cutting of action cutting module 112 is obtained and obtains eigenvalue；Most
Afterwards, data transmission blocks 114 send the eigenvalue that the first classification of motion module 113 is obtained to wireless apss 120.
As shown in figure 5, including in wireless apss 120：Information receiving module 121, the second classification of motion module 122, password
Acquisition module 123 and password authentication module 124, wherein, the second classification of motion module 122 is connected with information receiving module 121,
Password acquisition module 123 is connected with the second classification of motion module 122, and password authentication module 124 is connected with password acquisition module 123
Connect.In the course of the work, first, information receiving module 121 receives the spy that data transmission blocks 114 send in mobile terminal 1 10
Value indicative；Afterwards, the second classification of motion module 122 is classified to which according to the eigenvalue that information receiving module 121 is received；Connect
, password acquisition module 123 obtains corresponding password to be certified according to the result that the second classification of motion module 122 is classified；Finally,
Password authentication module 124 is authenticated to the password to be certified that password acquisition module 123 is obtained, and completes the nothing of mobile terminal 1 10
Line certification.
Furthermore, it is understood that also including in mobile terminal 1 10：First configuration module, for setting deliberate action and waiting to recognize
The threshold parameter of incidence relation, the directioin parameter for setting deliberate action and setting deliberate action between card password, threshold value are joined
Number includes：Acceleration, angular velocity and displacement during 10 generation action of mobile terminal 1；
Specifically, for deliberate action, can be that the action of whipping can also be once closed for mobile action, such as left and right whipping
Connection numeral 1, upper and lower whipping once associates numeral 0；And for example, mobile circle association digital 2 clockwise, counterclockwise mobile circle association
Numeral 3 etc..Further, an action can associate a code to be certified, it is also possible to associate two codes to be certified, according to practical situation
Set；Such as, left and right whipping once associates numeral 01, and upper and lower whipping once associates numeral 23 etc., and here is not specifically limited.
For the directioin parameter of deliberate action, in one embodiment, with 10 display screen of mobile terminal 1 as coordinate axess,
It is laterally x-axis, longitudinal direction is y-axis, and vertical screen is z-axis, left and right whipping, association numeral 1 will be set as along x-axis whipping；Will be along y
Axle gets rid of and is set as upper and lower whipping, association numeral 0；By mobile circle association digital 2 clockwise in x/y plane；By the inverse time in x/y plane
One circle association numeral 3 of pin movement etc., by that analogy.
For the threshold parameter of deliberate action, at least include：Acceleration, angular velocity during 10 generation action of mobile terminal 1
And displacement.In a kind of specific embodiment, the displacement of 10 or so whipping of mobile terminal 1/upper and lower whipping is set
Threshold value is set as that 0.2m (rice), acceleration are 0.1m/s2, by that analogy.It is noted that we are to above-mentioned default association here
Relation, directioin parameter and threshold parameter are not especially limited, and can be set according to practical situation.In practical application
In, by the built-in accelerometer of mobile terminal 1 10 and the physical data of gyroscope acquisition 10 whipping of mobile terminal 1/movement.
In addition, also including network training module, for setting up action cutting training network, and set according in configuration module
Parameter Training carried out to which, action cutting training network includes：Data Layer, the first full articulamentum, active coating,
Two full articulamentums and Euclidean loss layer；And for setting up posture prediction network and carrying out Training to which, posture is predicted
Network is a depth network；Network cutting module, the posture prediction network split for training network training module is the
One posture prediction network and second prediction network.
For training network is cut in action, including but not limited to：Data Layer, the first full articulamentum, active coating, second connect entirely
Connect layer and Euclidean loss layer (Euclidean Loss layers).When the training for having supervision is carried out to the action slicing network, will
Initial data input data layer, is returned using label (given data) in Euclidean Loss layers, with this, by instruction
Practice so that the error (loss) of whole action cutting training network is minimum.Specifically, in this process, input is original
Data can be arbitrary data, naturally it is also possible to be the physical data of the action of the mobile terminal 1 10 arrived used in the present invention；?
The label that Euclidean Loss layers are used is the correct cutting result of the initial data of input.In an example, it is assumed that should
Action cutting training network supports 4 alphabetical combinations, i.e. then cut point position 3, if the data of input are write manually for user
ABCD, then the label that Euclidean Loss layers are used be correct A, B, C, D ways of writing, with this to the action
Cutting training network is trained, and until error is minimum, completes to train.Finally, it is to be noted that, above-mentioned active coating is ReLU
Layer, certainly can also be other activation primitives, according to practical situation depending on.
For posture predicts network, which is specially a depth network.In an example, wrap in the posture prediction network
Include：The multilamellar convolutional neural networks of data Layer, multiple convolutional layers and full articulamentum, also include SoftMax graders；
SoftMax graders are classified to action according to the eigenvalue that multilamellar convolutional neural networks are exported.Predicted using the posture
Training is carried out to which before network, to improve the accuracy that the posture predicts network final classification.Specifically, in training
During, the initial data of input can be arbitrary data, naturally it is also possible to be the mobile terminal 1 10 arrived used in the present invention
The physical data of action, does not specifically limit.After predicting network training well the posture, network cutting module is split
Network and second prediction network are predicted for prime, wherein, prime predicts network and action training slicing network
It is placed in mobile terminal 1 10, second prediction network is placed in wireless apss 120.Network and second is predicted as prime
Posture predict network split degree, we are not specifically limited here, according to practical situation depending on, e.g., in an example,
Prime prediction network is the multilamellar convolutional neural networks being made up of data Layer, multiple convolutional layers and full articulamentum, second
Posture prediction network is SoftMax graders.
This is based on, in action cutting module 112, the physical data input that data acquisition module 111 is obtained is trained
Action cutting implement network, realize to obtain physical data carry out action cutting, wherein, action cutting implement network in wrap
Include：Data Layer, the first full articulamentum, active coating and the second full articulamentum；In the first classification of motion module 113, for inciting somebody to action
Prime prediction network in the action data input mobile terminal 1 10 that the cutting of action cutting module 112 is obtained obtains feature
Value；In the second classification of motion module 122, for the eigenvalue for obtaining input second prediction real-time performance cutting is obtained
Action classification.
Specifically, in the course of the work, after the data acquisition module 111 in mobile terminal 1 10 detects physical data,
Action cutting is inputted in action cutting module 112 and implements network, obtain the cutting between each action in the second full articulamentum
Point, obtains each action；Afterwards, will be pre- for the parameter input prime of each action for obtaining in the first classification of motion module 113
Survey grid network obtains corresponding eigenvalue, and sends it to wireless apss 120.After wireless apss 120 receive this feature value,
By eigenvalue input second prediction network in two classification of motion modules 122, the classification to action is realized, final output is each dynamic
Make the prediction probability of classification, using the classification of maximum predicted probability as the classification for predicting, obtain final result；Afterwards, password
Acquisition module 123 obtains password to be certified accordingly and is authenticated which to which according to the incidence relation of prediction, if to be certified
Password is consistent with preset password, then compare successfully, complete wireless authentication, it is allowed to user access network；Otherwise, authentification failure, no
Allow user access network.
It should be noted that above-described embodiment can independent assortment as needed.The above is only the preferred of the present invention
Embodiment, it is noted that for those skilled in the art, in the premise without departing from the principle of the invention
Under, some improvements and modifications can also be made, these improvements and modifications also should be regarded as protection scope of the present invention.
Claims (10)
1. a kind of wireless authentication method, it is characterised in that including the mobile terminal that communicates to connect and wireless apss, the wireless authentication
Method includes：
S1 mobile terminals detect auto- physical data；
S2 mobile terminals carry out action cutting to the physical data for obtaining；
S3 mobile terminals are classified to the action that cutting is obtained jointly with wireless apss；
S4 wireless apss obtain corresponding password to be certified and carry out wireless authentication to which according to the result of classification.
2. wireless authentication method as claimed in claim 1, it is characterised in that also included before step S1：
S01 sets the directioin parameter of deliberate action；
S02 sets the threshold parameter of deliberate action, and the threshold parameter includes：Acceleration, angle during mobile terminal generation action
Speed and displacement；
S03 sets up action cutting training network, and has carried out supervision according to the parameter set in step S01 and step S02 to which
Training, the action cutting training network include：Data Layer, the first full articulamentum, active coating, the second full articulamentum and Europe
Family name's loss layer；
S04 sets up posture prediction network and carries out Training to which, and the posture prediction network is a depth network；
The posture for training prediction network split is that prime predicts network and second prediction network by S05；
S06 sets the incidence relation between deliberate action and password to be certified, wherein, action correspondingly at least one to be certified
Password.
3. wireless authentication method as claimed in claim 2, it is characterised in that
Step S1 is specially：The physical data in mobile terminal course of action is detected by the built-in sensor of mobile terminal；
Step S2 is specially：Network is implemented in the action cutting that the physical data input that step S1 is obtained by mobile terminal is trained,
Realize carrying out action cutting to the physical data for obtaining, wherein, network is implemented in the action cutting to be included：Data Layer, first
Full articulamentum, active coating and the second full articulamentum；
Step S3 is specially：By cut in step S2 obtain action data input mobile terminal in prime prediction network
Obtain eigenvalue and send it to wireless apss, wireless apss are by the eigenvalue for obtaining input second prediction real-time performance to dynamic
The classification of work.
4. wireless authentication method as claimed in claim 2 or claim 3, it is characterised in that specifically include in step s 4：
S41 wireless apss are according to classification in the incidence relation between the deliberate action and password to be certified set in S05 and step S3
The action for obtaining obtains password to be certified accordingly；
The password to be certified for receiving is compared by S42 wireless apss with preset password, completes the wireless authentication of mobile terminal.
5. wireless authentication method as claimed in claim 3, it is characterised in that cut training network in action and action cutting is real
Apply in network, the active coating is ReLU layers；And/or,
In posture prediction network, the posture prediction network includes：Data Layer, multiple convolutional layers and full articulamentum many
Layer convolutional neural networks, also include SoftMax graders；The SoftMax graders are exported according to multilamellar convolutional neural networks
Eigenvalue action is classified.
6. a kind of wireless authentication system, it is characterised in that the wireless authentication system includes the nothing of communication of mobile terminal connection
Line AP, wherein,
Mobile terminal, for detect auto- physical data, for obtain physical data carry out action cutting and
The action that cutting is obtained is classified for associated wireless AP；
Wireless apss are additionally operable to obtain corresponding password to be certified and carry out wireless authentication to which according to the result of classification.
7. wireless authentication method as claimed in claim 6, it is characterised in that
The mobile terminal includes：
Data acquisition module, for obtaining the physical data in mobile terminal course of action；
Action cutting module, the physical data for getting to data acquisition module carry out action cutting；
First classification of motion module, obtains eigenvalue for carrying out process to the action that the cutting of action cutting module is obtained；
Data transmission blocks, for sending the eigenvalue that the first classification of motion module is obtained to wireless apss；
The wireless apss include：
Information receiving module, for receiving the eigenvalue of mobile terminal transmission；
Second classification of motion module, the eigenvalue for being received according to information receiving module are classified to which；
Password acquisition module, for obtaining corresponding password to be certified according to the result of the second classification of motion module classification；
Password authentication module, for being authenticated to the password to be certified that password acquisition module is obtained.
8. wireless authentication system as claimed in claim 7, it is characterised in that
Also include in the terminal：
First configuration module, for the side for setting the incidence relation between deliberate action and password to be certified, setting deliberate action
To parameter and the threshold parameter of setting deliberate action, the threshold parameter includes：Acceleration during mobile terminal generation action,
Angular velocity and displacement；
Network training module, for setting up action cutting training network, and is carried out to which according to the parameter set in configuration module
Training, the action cutting training network include：Data Layer, the first full articulamentum, active coating, the second full articulamentum
And Euclidean loss layer；And for setting up posture prediction network and carrying out Training to which, the posture prediction network is
One depth network；
Network cutting module, the posture prediction network split for training network training module is that prime predicts network
Network is predicted with second.
9. wireless authentication system as claimed in claim 8, it is characterised in that
In action cutting module, the physical data that data acquisition module is obtained is input into the action cutting enforcement net for training
Network, realizes carrying out action cutting to the physical data for obtaining, and wherein, network is implemented in the action cutting to be included：Data Layer,
One full articulamentum, active coating and the second full articulamentum；
In the first classification of motion module, for the action data that the cutting of action cutting module is obtained is input in mobile terminal
Prime prediction network obtains eigenvalue；
In the second classification of motion module, for the eigenvalue for obtaining is input into what second prediction real-time performance cutting was obtained
The classification of action.
10. wireless authentication system as claimed in claim 9, it is characterised in that
Training network is cut in action and action cutting is implemented in network, the active coating is ReLU layers；And/or,
In posture prediction network, the posture prediction network includes：Data Layer, multiple convolutional layers and full articulamentum many
Layer convolutional neural networks, also include SoftMax graders；The SoftMax graders are exported according to multilamellar convolutional neural networks
Eigenvalue action is classified.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611169078.1A CN106507356A (en) | 2016-12-16 | 2016-12-16 | A kind of wireless authentication method and its system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611169078.1A CN106507356A (en) | 2016-12-16 | 2016-12-16 | A kind of wireless authentication method and its system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106507356A true CN106507356A (en) | 2017-03-15 |
Family
ID=58330203
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611169078.1A Pending CN106507356A (en) | 2016-12-16 | 2016-12-16 | A kind of wireless authentication method and its system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106507356A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020000966A1 (en) * | 2018-06-27 | 2020-01-02 | 上海连尚网络科技有限公司 | Method for generating wireless access point information, device, and computer readable medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101789862A (en) * | 2010-01-25 | 2010-07-28 | 中兴通讯股份有限公司 | Encryption and decryption device and method based on gravity acceleration |
| CN105072611A (en) * | 2015-07-08 | 2015-11-18 | 广东欧珀移动通信有限公司 | Verification method and terminal |
| CN105160310A (en) * | 2015-08-25 | 2015-12-16 | 西安电子科技大学 | 3D (three-dimensional) convolutional neural network based human body behavior recognition method |
| CN105740892A (en) * | 2016-01-27 | 2016-07-06 | 北京工业大学 | High-accuracy human body multi-position identification method based on convolutional neural network |
-
2016
- 2016-12-16 CN CN201611169078.1A patent/CN106507356A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101789862A (en) * | 2010-01-25 | 2010-07-28 | 中兴通讯股份有限公司 | Encryption and decryption device and method based on gravity acceleration |
| CN105072611A (en) * | 2015-07-08 | 2015-11-18 | 广东欧珀移动通信有限公司 | Verification method and terminal |
| CN105160310A (en) * | 2015-08-25 | 2015-12-16 | 西安电子科技大学 | 3D (three-dimensional) convolutional neural network based human body behavior recognition method |
| CN105740892A (en) * | 2016-01-27 | 2016-07-06 | 北京工业大学 | High-accuracy human body multi-position identification method based on convolutional neural network |
Non-Patent Citations (1)
| Title |
|---|
| 吴军等: "基于深度卷积神经网络的人体动作识别", 《华中科技大学学报》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020000966A1 (en) * | 2018-06-27 | 2020-01-02 | 上海连尚网络科技有限公司 | Method for generating wireless access point information, device, and computer readable medium |
| US11863995B2 (en) | 2018-06-27 | 2024-01-02 | Shanghai Lianshang Network Technology Co., Ltd. | Method for generating wireless access point information, device, and computer readable medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11038879B2 (en) | Identity authentication method and apparatus, and user equipment | |
| CN107122641B (en) | Intelligent equipment owner identification method and intelligent equipment owner identification device based on use habit | |
| CN107742069A (en) | terminal control method, device and storage medium | |
| CN105989333A (en) | Fingerprint authentication method, fingerprint authentication system and terminal supporting fingerprint authentication function | |
| CN108121210A (en) | Permission distribution method and device for household appliance, storage medium and processor | |
| CN104820924A (en) | Online safe payment system based on handwriting authentication | |
| CN104618316B (en) | Safe verification method, device and system | |
| CN106790129A (en) | A kind of identity authentication method and device | |
| CN107851190A (en) | Slave unit is electrically connected to the communicator of main equipment | |
| CN104022878B (en) | A kind of pad pasting SIM card and corresponding weight discriminating system and method for authenticating | |
| CN105429969A (en) | User identity verification method and equipment | |
| CN107042803A (en) | The Vehicular system of wearable device control | |
| CN105488875B (en) | A kind of gate inhibition's verification method and device | |
| CN106470192B (en) | Identity verification method, device and system | |
| CN108495253A (en) | Public transport localization method and device, storage medium and electronic device | |
| CN108600057A (en) | A kind of method and system perceiving control household electrical appliance based on WiFi | |
| CN107358432A (en) | Mobile terminal is swiped the card method, apparatus and computer-readable recording medium | |
| CN104808940A (en) | Method and device for transmitting data between terminals | |
| CN107480513A (en) | A kind of certification implementation method and device, computer installation, readable storage medium storing program for executing | |
| CN108762659A (en) | Control method, device, storage medium and the electronic device of cloud terminal | |
| CN106156560A (en) | Unlocking method and a device for touch panel device | |
| CN106507356A (en) | A kind of wireless authentication method and its system | |
| CN106600759A (en) | Fingerprint verifying method, server and system | |
| CN106254942A (en) | A kind of control the method for television system, television system and mobile terminal | |
| CN107294981A (en) | A kind of method and apparatus of certification |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170315 |