CN110062379A - Identity identifying method based on channel state information under a kind of human body behavior scene - Google Patents
Identity identifying method based on channel state information under a kind of human body behavior scene Download PDFInfo
- Publication number
- CN110062379A CN110062379A CN201910301024.3A CN201910301024A CN110062379A CN 110062379 A CN110062379 A CN 110062379A CN 201910301024 A CN201910301024 A CN 201910301024A CN 110062379 A CN110062379 A CN 110062379A
- Authority
- CN
- China
- Prior art keywords
- data
- human body
- channel state
- movement
- state information
- 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
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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
- H04W12/60—Context-dependent security
- H04W12/65—Environment-dependent, e.g. using captured environmental data
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W12/00—Security arrangements; Authentication; Protecting privacy or anonymity
- H04W12/60—Context-dependent security
- H04W12/68—Gesture-dependent or behaviour-dependent
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Security & Cryptography (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Social Psychology (AREA)
- Environmental & Geological Engineering (AREA)
- Health & Medical Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The present invention relates to channel state information application fields, and in particular to the identity identifying method based on channel state information under a kind of human body behavior scene.The acquisition for carrying out data first, obtains initial human body behavioral data, then carries out data prediction, movement segmentation and data reduction are carried out to data, sparse matrix is obtained, finally according to dense convolutional neural networks model, obtains authentication result by being classified to data；Compared with the identity identifying method of other channel state informations, overall calculation of the invention is time-consuming lower and can guarantee high Average Accuracy in verification process.
Description
Technical field
The present invention relates to channel state information application fields, and in particular to channel status is based under a kind of human body behavior scene
The identity identifying method of information.
Background technique
Current century with the communication technology fast development, people for human-computer interaction form probing direction from equipment
It is transferred to centered on interaction centered on the behavior perception of people.Identity identifying technology is as in general fit calculation and field of human-computer interaction
The a key technology being concerned has become one of the hot topic that temperature in recent years remains high.
Since authentication generally requires through the strategy of high precision the authentication result obtained to the identity of people.In reality
In the life of border, authentication can provide many aspects such as the protection of the privacy of user, personal safety protection and safeguarding of assets
Technical guarantee.
Currently, direction exhibition of some researchs around the authentication for carrying out people to gait information by channel state information
It opens.2016, Zhang et al. proposed the identity identifying method of WIFI-ID a kind of, and this method passes through analysis WIFI signal pair
In the feedback of the gait behavior of people, wireless signal is demonstrated for the enforceability of authentication procedures.The same year, Yuan etc.
People proposes the identity identifying method of WiWho a kind of, and this method can pass through the gait information of people in family or office
The identity of people is identified.Then, Wang et al. proposes a kind of identity identifying method WIFIU by gait, this method
The relevant channel state information data of gait are acquired under los path by people, and are supported using Principal Component Analysis and SVM
Vector machine classifier authenticates the gait behavior of user.2017, Xin et al. proposed a kind of entitled Free sense's
Gait identity authentication method, this method complete the classification and identification to user's gait by k-nearest neighbor.Then, Shi et al. is mentioned
Go out a kind of identity identifying method, the activity variance feature that this method is reflected by people in analysis channel state information data,
And application deep neural network technology completes the verification process to people.
However, the correlative study of the authentication based on channel state information largely concentrate with to the gait information of people into
Row research is unfavorable for recognizing from whole angle to have ignored the signature analysis of the human body behavior including gesture to people
Know the relationship between the motor behavior of people and identity characteristic.Meanwhile these overall works of research in authentication procedures are time-consuming
It is longer, it can not be expanded and spread in more application scenarios.
Summary of the invention
The purpose of the present invention is to provide the identity identifying method based on channel state information under a kind of human body behavior scene,
By dense convolutional neural networks model to promote the accuracy rate of whole verification process, while it is time-consuming to improve identification to reduce calculating
Rate.
The embodiment of the present invention provides the identity identifying method based on channel state information under a kind of human body behavior scene, packet
It includes:
Step 1: the acquisition of data: according in radio local network environment, the sighting distance between receiving end and transmitting terminal is captured
The various gestures movement carried out under path obtains initial human body by collecting the data packet of the channel state information
Behavioral data；
Step 2: data prediction: the initial human body behavioral data is carried out at denoising by low-pass filter
Reason, the human body behavioral data after being denoised；Principal Component Analysis is used to the human body behavioral data after the denoising, is obtained
Pretreated data；
Step 3: movement segmentation and data reduction: the movement of pretreated data is segmented, obtains enlivening movement
The subcarrier data in section and the subcarrier data of movement interval section；The subcarrier data in the active movement section is protected
It stays, the subcarrier data for acting interval section sets 0, removes disturbing factor and obtains sparse matrix；
Step 4: the classification and certification of data: according to the dense convolutional neural networks model of Dense Net-BC structure,
By classifying to data, authentication result is obtained；
The step 1, comprising:
The acquisition of data: it according in radio local network environment, captures under the los path between receiving end and transmitting terminal
The various gestures movement carried out obtains initial human body behavior number by collecting the data packet of the channel state information
According to；
Wherein, the collection process of the data are as follows:
If transmitter has m antenna Tx, receiver is with n antenna Rx, the transmission of m × n item will be formed in communication process
Link, wherein every transmissions links contain 30 subcarrier informations, is each wrapped in the data packet containing channel state information
Include the data that matrix form is m × n × 30；Transmitting terminal has 3 transmitting antennas in the present invention, and receiving end has 1 and receives antenna,
Channel state information is 1 × 3 × 30 complex matrix form in communication process, and all has 90 sons in each data packet and carry
Wave；
The step 2, comprising:
Data prediction: denoising is carried out to the initial human body behavioral data by low-pass filter, is obtained
Human body behavioral data after denoising；Principal Component Analysis is used to the human body behavioral data after the denoising, is pre-processed
Data afterwards；
Wherein, the denoising method particularly includes:
Denoising is carried out to raw information using 5 rank Butterworth low-pass filters, and will be existed with low frequency form
The human body behavior including gesture action message retain；
Wherein, the specific steps of the Principal Component Analysis are as follows:
If the complex matrix form through past hot-tempered channel state information data are as follows:
Seek the column average value of matrix H
Average value is subtracted in original matrix, seeks normalized matrix R:
According to normalized matrix, the covariance S of data characteristics is sought:
The characteristic value V and feature vector P for seeking covariance S, characteristic value is arranged from big to small, and by corresponding feature vector
It is rearranged in the form of column vector and forms new matrix:
V=[v1, v2..., v90]T；P=[p1, p2..., p90]
It is added up to characteristic value with direction from big to small and gradually counts contribution rate G, until meeting the threshold value δ of setting and remembering
The current quantity x that adds up of record, sets δ=0.9 in the present invention,
Data Dimensionality Reduction: the preceding x of selection matrix P is arranged and is formed new data matrix D:
D=[p1, p2..., px]；
The step 3, comprising:
Movement segmentation and data reduction: being segmented the movement of pretreated data, obtains active movement section
The subcarrier data of subcarrier data and movement interval section；The subcarrier data in the active movement section is retained, is moved
The subcarrier data for making interval section sets 0, removes disturbing factor and obtains sparse matrix；
Wherein, the specific steps of the movement segmentation and data reduction are as follows:
Working frequency is 200Hz in the present invention, and setting window k is 0.05s,
Windowed segments: to column main component p each in matrix Di(i=1 ..., x), total duration is set as t, according to window
Value k is divided, SiRepresent the information of i-th of window, window sum N=t/k；
IfFor piConceptual data average value, calculate the variance of the data in each window:
Calculate the average variance of all windows:
Movement segmentation: by the variance E of each windowiWith whole average variance EavgIt is compared, works as EiCompare EavgHour, will
Window data SiIt is added in movement interval section set Q；
Acquisition time section: according to movement interval section set Q, p is obtainediMovement interval section in (i=1 ..., x)
Annual distribution set T, meets mapping relations between Q and T:
Data reduction: to the amplitude data of column each in matrix H, zero setting is carried out according to time set T, and obtain by letter
Change treated sparse matrix
The step 4, comprising:
The classification and certification of data: according to the dense convolutional neural networks model of Dense Net-BC structure, by right
Data are classified, and authentication result is obtained；
Wherein, the classification and certification of the data method particularly includes:
The present invention uses the dense network with 3 dense piece to carry out network training and feature to channel state information data
Extraction process；
The average pond of the overall situation: it for the output data Y of dense convolutional neural networks model, using the average pond of the overall situation and obtains
Must be averaged value set
The present invention solves the classification problem of K people using normalization exponential function, to the output result of neuron it
Between carry out the calculating of probability, and be respectively mapped in corresponding (0,1) section:
Loss function: carrying out difference to tag along sort result and the true tag of normalization exponential function and compare and assess,
The cross entropy loss function for combining L2 normal form is selected to assess as loss function C and to current network model:
The beneficial effects of the present invention are:
1. the present invention is compared with the authentication method of the channel state information under gait scene, application scenarios are more extensive, energy
It is enough that the study of deep layer is carried out to the continuous human body behavior including gesture and completes to authenticate, therefore can be to recognizing under office scene
Card method provides solution；
2. the present invention combines data reduction of the movement segmentation method to movement irrelevant portions using Principal Component Analysis.In master
In component analysis procedure, count the feature vector of descending sort contribution rate and, until reach preset fixed threshold and extract
Data source of the individual features vector as movement segmentation, which can significantly reduce the data complexity in calculating process simultaneously
It is time-consuming to reduce overall calculation.Meanwhile the Annual distribution of the movement interval section in these data can represent original whole number
Interval time in.In order to optimize calculating effect, the amplitude data zero setting of time corresponding in former data is constituted into sparse square
Battle array；
3. the present invention achievees the effect that Deep Learning by the dense convolutional neural networks of Dense Net-BC type.Simultaneously
Compared with the identity identifying method of other channel state informations, overall calculation of the invention is time-consuming lower and in verification process
It can guarantee high Average Accuracy.
Detailed description of the invention
Fig. 1 is the implement scene schematic diagram of human body behavior；
Fig. 2 is a kind of flow chart of the identity identifying method based on channel state information under human body behavior scene；
The structure organization figure that Fig. 3 is dense piece in the dense convolutional neural networks model of the present invention；
Specific embodiment
In order to make the foregoing objectives, features and advantages of the present invention clearer and more comprehensible, with reference to the accompanying drawing to the present invention
It is described further:
Fig. 1 is the implement scene schematic diagram of human body behavior；
Fig. 2 is a kind of flow chart of the identity identifying method based on channel state information under human body behavior scene；
The structure organization figure that Fig. 3 is dense piece in the dense convolutional neural networks model of the present invention.
The technical scheme of the present invention is realized as follows:
1, data acquire: in the collection process of data, if transmitter has m antenna (Tx) and receiver there is n
Antenna (Rx), then m × n transmissions links will be formed in communication process, wherein carrying containing 30 sons in every transmissions links
Wave information then each includes the data that matrix form is m × n × 30 in the data packet containing channel state information.In this hair
In bright, transmitting terminal has 3 transmitting antennas, and receiving end has 1 and receives antenna, therefore channel state information is 1 in communication process
× 3 × 30 complex matrix form, and 90 subcarriers are all had in each data packet.
2, data prediction: since indoor environment can generate a large amount of high frequency environment noise for channel state information, because
This amplitude information for directly analyzing the subcarrier in channel state information cannot obtain the relevant number of very intuitive human body behavior
It is believed that breath.In data preprocessing phase, the present invention believes original channel status using 5 rank Butterworth low-pass filters
Breath carries out denoising, and protects to the action message with the human body behavior existing for low frequency form including gesture
It stays.
3, using principal component analysis: in order to extract the strongest subcarrier data of human body behavior reaction ability to people, this
Invention is analyzed by the way of Principal Component Analysis by pretreated channel state information data.If hot-tempered through the past
The complex matrix form of channel state information data are as follows:
(1) the column average value of matrix H is sought
(2) average value is subtracted in original matrix, seeks normalized matrix R:
(3) according to normalized matrix, the covariance S of data characteristics is sought:
(4) the characteristic value V and feature vector P for seeking covariance S, characteristic value is arranged from big to small, and by corresponding feature
Vector is rearranged in the form of column vector to be formed new matrix and can obtain:
V=[v1, v2..., v90]T (5)
P=[p1, p2..., p90] (6)
(5) it is added up to characteristic value with direction from big to small and gradually counts contribution rate G, until meeting the threshold value δ of setting
And currently cumulative quantity x is recorded, δ=0.9 is set in the present invention
(6) Data Dimensionality Reduction: the preceding x of selection matrix P is arranged and is formed new data matrix D:
D=[p1, p2..., px] (8)
4, movement segmentation and data reduction: the working frequency applied in the present invention is 200Hz, therefore window k is arranged and is
0.05s。
(1) windowed segments: to each column main component p in matrix Di(i=1 ..., x), total duration is set as t, according to
Window value k is divided, SiRepresent the information of i-th of window, window sum N=t/k.
(2) it setsFor piConceptual data average value, calculate the variance of the data in each window:
(3) average variance of all windows is calculated:
(4) movement segmentation: by the variance E of each windowiWith whole average variance EavgIt is compared.Work as EiCompare EavgHour,
By window data SiIt is added in movement interval section set Q.
(5) acquisition time section: according to movement interval section set Q, p is obtainediInterval section is acted in (i=1 ..., x)
Annual distribution set T, mapping relations are met between Q and T:
(6) data reduction: to the amplitude data of column each in matrix H, zero setting is carried out according to time set T, and obtain warp
It crosses and simplifies treated sparse matrix
5, classification and certification: in order to promote certification effect, the dense network model of Dense Net-BC structure is selected to carry out
The classification of authentication based on human body behavior and verification process.
(1) dense convolutional neural networks: the present invention uses the dense network with 3 dense piece to channel state information number
It is as shown in Fig. 3 according to progress network training and characteristic extraction procedure, each dense piece of institutional framework.
(2) global average pond: for the output data Y of dense convolutional neural networks model, using the average pond of the overall situation
And obtain average value set
(3) normalize exponential function: the present invention solves the classification problem of K people using normalization exponential function, right
The calculating of probability is carried out between the output result of neuron, and is respectively mapped in corresponding (0,1) section:
(4) it loss function: is compared in order to which the tag along sort result to normalization exponential function carries out difference with true tag
And assessment, while the appearance of overfitting problem is prevented, the present invention, which selects, combines the cross entropy loss function of L2 normal form as loss
Function C simultaneously assesses current network model.
。
Claims (5)
1. the identity identifying method based on channel state information under a kind of human body behavior scene characterized by comprising
Step 1: the acquisition of data: according in radio local network environment, the los path between receiving end and transmitting terminal is captured
Lower carried out various gestures movement obtains initial human body behavior by collecting the data packet of the channel state information
Data；
Step 2: data prediction: denoising is carried out to the initial human body behavioral data by low-pass filter, is obtained
Human body behavioral data after to denoising；Principal Component Analysis is used to the human body behavioral data after the denoising, obtains pre- place
Data after reason；
Step 3: movement segmentation and data reduction: being segmented the movement of pretreated data, obtains active movement section
Subcarrier data and movement interval section subcarrier data；The subcarrier data in the active movement section is retained,
The subcarrier data of movement interval section sets 0, removes disturbing factor and obtains sparse matrix；
Step 4: the classification and certification of data: according to the dense convolutional neural networks model of Dense Net-BC structure, pass through
Classify to data, obtains authentication result.
2. the identity identifying method based on channel state information under a kind of human body behavior scene according to claim 1,
It is characterized in that, the step 1, comprising:
The acquisition of data: according in radio local network environment, capture lower of los path between receiving end and transmitting terminal into
Capable various gestures movement obtains initial human body behavioral data by collecting the data packet of the channel state information；
Wherein, the collection process of the data are as follows: set transmitter with m antenna Tx, receiver is with n antenna Rx, communicating
M × n transmissions links will be formed in the process, wherein every transmissions links contain 30 subcarrier informations, each contain channel
It include the data that matrix form is m × n × 30 in the data packet of status information；Transmitting terminal has 3 transmitting antennas in the present invention,
Receiving end has 1 and receives antenna, the complex matrix form that channel state information is 1 × 3 × 30 in communication process, and every number
According to all having 90 subcarriers in packet.
3. the identity identifying method based on channel state information under a kind of human body behavior scene according to claim 1,
It is characterized in that: the step 2, comprising:
Data prediction: denoising is carried out to the initial human body behavioral data by low-pass filter, is denoised
Human body behavioral data afterwards；Principal Component Analysis is used to the human body behavioral data after the denoising, is obtained pretreated
Data；
Wherein, the denoising method particularly includes: raw information is carried out using 5 rank Butterworth low-pass filters
Denoising, and will be with the reservation of the action message of the human body behavior existing for low frequency form including gesture；
Wherein, the specific steps of the Principal Component Analysis are as follows: set the complex matrix through past hot-tempered channel state information data
Form are as follows:
Seek the column average value of matrix H
Average value is subtracted in original matrix, seeks normalized matrix R:
According to normalized matrix, the covariance S of data characteristics is sought:
The characteristic value V and feature vector P for seeking covariance S, characteristic value are arranged from big to small, and by corresponding feature vector to arrange
The form of vector rearranges the new matrix of composition:
V=[v1,v2,…,v90]T；P=[p1,p2,…,p90]
It is added up to characteristic value with direction from big to small and gradually counts contribution rate G, until the threshold value δ and record of satisfaction setting work as
Preceding cumulative quantity x, the present invention in set δ=0.9,
Data Dimensionality Reduction: the preceding x of selection matrix P is arranged and is formed new data matrix D:
D=[p1,p2,…,px]。
4. the identity identifying method based on channel state information under a kind of human body behavior scene according to claim 1,
It is characterized in that: the step 3, comprising:
Movement segmentation and data reduction: being segmented the movement of pretreated data, and the son for obtaining active movement section carries
Wave number is according to the subcarrier data with movement interval section；The subcarrier data in the active movement section is retained, between movement
Subcarrier data between septal area sets 0, removes disturbing factor and obtains sparse matrix；
Wherein, the specific steps of the movement segmentation and data reduction are as follows: working frequency is 200Hz in the present invention, and window k is arranged
For 0.05s,
Windowed segments: to column main component p each in matrix Di(i=1 ..., x), total duration is set as t, according to window value k into
Row divides, SiRepresent the information of i-th of window, window sum N=t/k；
IfFor piConceptual data average value, calculate the variance of the data in each window:
Calculate the average variance of all windows:
Movement segmentation: by the variance E of each windowiWith whole average variance EavgIt is compared, works as EiCompare EavgHour, by the window
Mouth data SiIt is added in movement interval section set Q；
Acquisition time section: according to movement interval section set Q, p is obtainediThe time point of movement interval section in (i=1 ..., x)
Cloth set T, meets mapping relations between Q and T:
Data reduction: to the amplitude data of column each in matrix H, zero setting is carried out according to time set T, and is obtained at by simplifying
Sparse matrix after reason
5. the identity identifying method based on channel state information under a kind of human body behavior scene according to claim 1,
It is characterized in that: the step 4, comprising:
The classification and certification of data: according to the dense convolutional neural networks model of Dense Net-BC structure, by data
Classify, obtains authentication result；
Wherein, the classification and certification of the data method particularly includes: the present invention uses the dense network with 3 dense piece
Network training and characteristic extraction procedure are carried out to channel state information data；
The average pond of the overall situation: it for the output data Y of dense convolutional neural networks model, using the average pond of the overall situation and obtains flat
Equal value set
The present invention using normalization exponential function the classification problem of K people is solved, between the output result of neuron into
The calculating of row probability, and be respectively mapped in corresponding (0,1) section:
Loss function: difference is carried out to tag along sort result and the true tag of normalization exponential function and compares and assesses, is selected
As loss function C and current network model is assessed in conjunction with the cross entropy loss function of L2 normal form:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910301024.3A CN110062379B (en) | 2019-04-15 | 2019-04-15 | Identity authentication method based on channel state information under human behavior scene |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910301024.3A CN110062379B (en) | 2019-04-15 | 2019-04-15 | Identity authentication method based on channel state information under human behavior scene |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110062379A true CN110062379A (en) | 2019-07-26 |
| CN110062379B CN110062379B (en) | 2022-10-28 |
Family
ID=67319149
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910301024.3A Active CN110062379B (en) | 2019-04-15 | 2019-04-15 | Identity authentication method based on channel state information under human behavior scene |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110062379B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110598552A (en) * | 2019-08-09 | 2019-12-20 | 吉林大学 | Expression recognition method based on improved particle swarm optimization convolutional neural network optimization |
| CN111093163A (en) * | 2019-12-23 | 2020-05-01 | 安徽大学 | Passive target classification method based on channel state information |
| CN111652132A (en) * | 2020-06-03 | 2020-09-11 | 河北工程大学 | Non-line-of-sight identity recognition method and device based on deep learning and storage medium |
| CN112308042A (en) * | 2020-05-22 | 2021-02-02 | 哈尔滨工程大学 | Stranger action identification method based on channel state information |
| CN113055057A (en) * | 2019-12-27 | 2021-06-29 | 电子科技大学 | Physical layer authentication method based on millimeter wave multi-antenna channel sparse peak characteristics |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104021573A (en) * | 2014-06-13 | 2014-09-03 | 哈尔滨工程大学 | Human movement classification and identification method based on joint pose angles |
| CN106899968A (en) * | 2016-12-29 | 2017-06-27 | 南京航空航天大学 | A kind of active noncontact identity identifying method based on WiFi channel condition informations |
| US10116695B1 (en) * | 2016-02-03 | 2018-10-30 | Symantec Corporation | Systems and methods for verifying that operators are human based on operator gaze |
| CN108932479A (en) * | 2018-06-06 | 2018-12-04 | 上海理工大学 | A kind of human body anomaly detection method |
| CN109063697A (en) * | 2018-10-11 | 2018-12-21 | 浙江工业大学 | A kind of human body sitting posture detection method based on channel state information |
| CN109325399A (en) * | 2018-07-13 | 2019-02-12 | 哈尔滨工程大学 | A kind of stranger's gesture identification method and system based on channel state information |
-
2019
- 2019-04-15 CN CN201910301024.3A patent/CN110062379B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104021573A (en) * | 2014-06-13 | 2014-09-03 | 哈尔滨工程大学 | Human movement classification and identification method based on joint pose angles |
| US10116695B1 (en) * | 2016-02-03 | 2018-10-30 | Symantec Corporation | Systems and methods for verifying that operators are human based on operator gaze |
| CN106899968A (en) * | 2016-12-29 | 2017-06-27 | 南京航空航天大学 | A kind of active noncontact identity identifying method based on WiFi channel condition informations |
| CN108932479A (en) * | 2018-06-06 | 2018-12-04 | 上海理工大学 | A kind of human body anomaly detection method |
| CN109325399A (en) * | 2018-07-13 | 2019-02-12 | 哈尔滨工程大学 | A kind of stranger's gesture identification method and system based on channel state information |
| CN109063697A (en) * | 2018-10-11 | 2018-12-21 | 浙江工业大学 | A kind of human body sitting posture detection method based on channel state information |
Non-Patent Citations (4)
| Title |
|---|
| MATTEO GADALETA, MICHELE ROSSI: "IDNet: Smartphone-based Gait Recognition with Convolutional Neural Networks", 《PREPRINT SUBMITTED TO PATTERN RECOGNITION》 * |
| XUE DING, TING JIANG, WEIXIA ZOU: "A New Method of Dynamic Gesture Recognition Using Wi-Fi Signals Based on Adaboost", 《2017 17TH INTERNATIONAL SYMPOSIUM ON COMMUNICATIONS AND INFORMATION TECHNOLOGIES (ISCIT)》 * |
| 蒋思琪，王向勇: "基于信道状态信息的人体行为识别方法", 《通信电源技术》 * |
| 陈甜甜，姚璜，左明章，田元，杨梦婷: "基于深度信息的动态手势识别综述", 《计算机科学》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110598552A (en) * | 2019-08-09 | 2019-12-20 | 吉林大学 | Expression recognition method based on improved particle swarm optimization convolutional neural network optimization |
| CN111093163A (en) * | 2019-12-23 | 2020-05-01 | 安徽大学 | Passive target classification method based on channel state information |
| CN113055057A (en) * | 2019-12-27 | 2021-06-29 | 电子科技大学 | Physical layer authentication method based on millimeter wave multi-antenna channel sparse peak characteristics |
| CN113055057B (en) * | 2019-12-27 | 2022-05-20 | 电子科技大学 | Physical layer authentication method based on millimeter wave multi-antenna channel sparse peak characteristics |
| CN112308042A (en) * | 2020-05-22 | 2021-02-02 | 哈尔滨工程大学 | Stranger action identification method based on channel state information |
| CN111652132A (en) * | 2020-06-03 | 2020-09-11 | 河北工程大学 | Non-line-of-sight identity recognition method and device based on deep learning and storage medium |
| CN111652132B (en) * | 2020-06-03 | 2022-06-21 | 河北工程大学 | Non-line-of-sight identity recognition method and device based on deep learning and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110062379B (en) | 2022-10-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110062379A (en) | Identity identifying method based on channel state information under a kind of human body behavior scene | |
| CN106899968A (en) | A kind of active noncontact identity identifying method based on WiFi channel condition informations | |
| Li et al. | Learning human activities through Wi-Fi channel state information with multiple access points | |
| CN106407905A (en) | Machine learning-based wireless sensing motion identification method | |
| CN109447099B (en) | PCA (principal component analysis) dimension reduction-based multi-classifier fusion method | |
| CN110730473B (en) | WiFi activity recognition-oriented signal feature extraction method | |
| Ding et al. | Wihi: WiFi based human identity identification using deep learning | |
| CN101216884A (en) | A method and system for face authentication | |
| CN103093133B (en) | Biological identity authentication method facing institute of electrical and electronic engineers (IEEE) 802.15.6 | |
| Ding et al. | Device-free location-independent human activity recognition using transfer learning based on CNN | |
| Samadh et al. | Indoor localization based on channel state information | |
| Li et al. | Fingerprints in the air: Unique identification of wireless devices using RF RSS fingerprints | |
| CN103530657B (en) | A kind of based on weighting L2 extraction degree of depth study face identification method | |
| CN115294658B (en) | Personalized gesture recognition system and gesture recognition method for multiple application scenes | |
| Shang et al. | LSTM-CNN network for human activity recognition using WiFi CSI data | |
| Cao et al. | A lightweight deep learning algorithm for WiFi-based identity recognition | |
| Hu et al. | ResFi: WiFi-enabled device-free respiration detection based on deep learning | |
| Liu et al. | Behavior privacy preserving in RF sensing | |
| CN111652132B (en) | Non-line-of-sight identity recognition method and device based on deep learning and storage medium | |
| Zhong et al. | A climate adaptation device-free sensing approach for target recognition in foliage environments | |
| CN109728863A (en) | Personnel activity's duration estimation method, device and terminal device | |
| Jiang et al. | RF‐Gait: Gait‐Based Person Identification with COTS RFID | |
| CN109308782B (en) | Behavior detection method and system for target object | |
| Wu et al. | Device-free human activity recognition with identity-based transfer mechanism | |
| Yin et al. | Back-guard: Wireless backscattering based user activity recognition and identification with parallel attention model |
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 |