CN112765550B - Target behavior segmentation method based on Wi-Fi channel state information - Google Patents
Target behavior segmentation method based on Wi-Fi channel state information Download PDFInfo
- Publication number
- CN112765550B CN112765550B CN202110076299.9A CN202110076299A CN112765550B CN 112765550 B CN112765550 B CN 112765550B CN 202110076299 A CN202110076299 A CN 202110076299A CN 112765550 B CN112765550 B CN 112765550B
- Authority
- CN
- China
- Prior art keywords
- data
- matrix
- searching
- fitting
- indexrt
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 37
- 230000011218 segmentation Effects 0.000 title claims abstract description 29
- 230000009471 action Effects 0.000 claims abstract description 41
- 230000008569 process Effects 0.000 claims abstract description 5
- 239000011159 matrix material Substances 0.000 claims description 40
- 230000006399 behavior Effects 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 7
- 238000009825 accumulation Methods 0.000 claims description 6
- 238000012163 sequencing technique Methods 0.000 claims description 6
- 230000003044 adaptive effect Effects 0.000 claims description 5
- 230000001174 ascending effect Effects 0.000 claims description 4
- 238000012847 principal component analysis method Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 3
- 238000007405 data analysis Methods 0.000 claims description 3
- 238000000354 decomposition reaction Methods 0.000 claims description 3
- 230000037081 physical activity Effects 0.000 claims description 3
- 230000009467 reduction Effects 0.000 claims description 3
- 230000003068 static effect Effects 0.000 claims description 3
- 238000005259 measurement Methods 0.000 abstract description 2
- 230000000694 effects Effects 0.000 description 10
- 230000006870 function Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Image Analysis (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention provides a target behavior segmentation method based on Wi-Fi channel state Information (CHANNEL STATE Information, CSI). Firstly, the method of removing trend by segment fitting is utilized to process the time sequence, and the data fluctuation of the action occurrence period and the data fluctuation of the non-action occurrence period still keep larger difference while eliminating the overall trend of the sequence; then, a sliding variance is obtained for the time series after trending, and a sliding window variance comparison method is proposed according to the variation of the sliding variance, thereby extracting a CSI data segment only including motion information. The actual measurement result shows that compared with the existing behavior segmentation method, the target behavior segmentation method designed by the invention can extract the data segment when the action happens from the original CSI data more accurately. In addition, the method can perform efficient segmentation even if the original CSI data includes a plurality of discontinuous actions.
Description
Technical Field
The invention belongs to a data processing method, and particularly relates to a target behavior segmentation method based on channel state information under a Wi-Fi system.
Background
In recent years, human activity recognition systems based on Wi-Fi wireless sensing technology are popular with more and more people because of the characteristics of non-invasiveness, insensitivity to light and the like. Researchers realize behavior recognition by monitoring channel state Information (CHANNEL STATE Information, CSI) Information of echo signals and extracting signal characteristics of different actions. Therefore, in activity recognition, it is extremely important to extract from the entire data stream only the data segment at which the action occurred.
For CSI data, the variance of the amplitude of the CSI in the presence of activity is much larger than the variance in the absence of activity, and the variance of the amplitude of the CSI in the presence of activity is sensitive to changes in the surrounding environment information. Based on this, some researchers have performed action segmentation by setting an optimal or adaptive threshold. For example, in the TW-see system, researchers first propose a normalized variance sliding window algorithm using the characteristic that CSI amplitude has large fluctuation variation during the occurrence of an activity; detection of the start and end time points of the action is then achieved by setting an optimal threshold. In Wi-Multi systems, researchers have designed a two-step algorithm to detect motion. The first step, dividing the whole signal into a plurality of sub-data segments with the same length; then calculating the variance of each sub data segment and carrying out ascending sort; and finally, calculating the difference value between the sum of the larger half and the sum of the smaller half in all variances, comparing the difference value with an adaptive threshold value, and if the difference value is larger than the threshold value, indicating that an action exists in the sub-data segment. And secondly, respectively calculating the maximum eigenvalues of the CSI amplitude autocorrelation matrix and the phase autocorrelation matrix, and carrying out secondary verification on the motion data segment determined in the first step by a machine learning method.
In addition to the threshold-based motion segmentation algorithm, the researcher divides the waveform of a complete motion into four states with the same length, which are respectively defined as a stationary state, a starting state, a motion state and an ending state. The offline stage generates an action state model through a convolutional neural network, and in the action segmentation stage, the action state model is used for judging the activity state of the sliding window data, so that the starting time and the ending time of a single activity are determined.
The above-described methods can achieve division to some extent, but various disadvantages are unavoidable. In the TW-see system, the threshold or other parameters are changed, and the segmentation performance of the system is correspondingly changed. The Wi-Multi system divides the whole CSI data into a plurality of data segments with the length corresponding to the packet sending rate, and then performs activity determination on each data segment, if the length of the data segment is too long, the accuracy of the division of the motion is reduced. The segmentation method based on deep learning requires a lot of time for model training, and the judgment performance of the model can greatly influence the subsequent action segmentation. In addition, if the scene changes, the effectiveness of these algorithms may decrease. Therefore, the invention provides a segmentation fitting trend-removing and window variance comparison method to realize the segmentation of actions, thereby improving the instantaneity and the effectiveness of a behavior segmentation algorithm.
Disclosure of Invention
The invention aims to provide a target behavior segmentation method based on Wi-Fi channel state information under a Wi-Fi system, which can accurately and effectively extract data segments when actions occur from original data.
The invention discloses a Wi-Fi-based target behavior segmentation method, which specifically comprises the following steps:
Step one: preprocessing the CSI data;
assuming that the CSI data acquired by the receiver is subjected to data analysis and channel correction to obtain a matrix H, and preprocessing the matrix to obtain a matrix A
Where N is the number of received data packets, N tx and N rx are the number of transmit antennas and receive antennas, respectively, and N sub is the number of subcarriers. The specific flow of pretreatment is as follows:
firstly, taking a module of each data in the matrix H, removing phase information in the data, then, carrying out decentralization processing on each row vector, eliminating static components of signals, and calculating the following equation:
Wherein, |·| represents modulo arithmetic;
Step two: the matrix A is subjected to dimension reduction by using a principal component analysis method, noise is suppressed, and meanwhile, the CSI amplitude containing the action information is more prominent, and the method is as follows;
Firstly, calculating covariance matrix of matrix A, then utilizing eigenvalue decomposition to obtain eigenvalue of said covariance matrix and correspondent eigenvector, finally utilizing transposed matrix A to multiply ith eigenvector to obtain ith principal component,
hi＝AT×Ei
Wherein h i is the ith principal component of the matrix A, the dimension is Nx1, E i is the eigenvector corresponding to the ith eigenvalue of the covariance matrix, and the dimension is N sub·Ntx·Nrx x 1. In the algorithm, a second principal component h 2 of the matrix a is mainly used;
Step three: processing the second principal component h 2 obtained in the step three by using a piecewise fitting trending method to obtain a vector T, The method comprises the following specific steps of;
First, the second principal component h 2 is divided into a plurality of data segments with a length of M, namely
h2＝[h2(W1),h2(W2),…,h2(Wk),…,h2(WN/M)]
Where W k is an array [ kM, kM+1, …, (k+1) M-1], representing the index value of each data in the kth data segment. Then fitting each segment of data respectively, wherein the following formula is as follows:
T(Wk)＝h2(Wk)-fit(h2(Wk))
wherein fit (·) is a polynomial fit based on least squares method, resulting in:
T＝[T(W1),T(W2),…,T(Wk),…,T(WN/M)]
the following is the implementation process of polynomial fitting:
assuming m points in the data segment, each data point is (x i,yi), the fitting result is:
where a 0,…,an is the unknown to be solved and n is the highest power of the polynomial. The sum of squares of the deviations is minimized according to the least square method, i.e
Solving the above coefficients can be converted into an extremum problem of i=i (a 0,a1,…,an), and solving the extremum according to the requirement of the multiple function:
I.e.
The representation in matrix form is then:
When the point to be fitted and the highest degree of polynomial fitting are given, each coefficient of the fitting polynomial can be obtained, and then the fitting residual is obtained by subtracting the fitted point from the point before fitting, so as to form a vector T. The subsequent data processing is based on this data.
Step four: setting a sliding window with the length of K, and calculating a sliding variance V of T, wherein V= [ V l]1×(N-K+1);
Assuming that the data contained in the first sliding window is [ T l,Tl+1,…,Tl+K-1 ], the variance of the first sliding window data can be expressed as:
Step five: calculating a threshold M and setting a window length W;
According to step four, we find the sliding variance V, and then calculate an adaptive threshold M for motion segmentation based on the sliding variance. The specific calculation idea is as follows:
Firstly, carrying out ascending arrangement on V, and searching an original index value set c with the smaller half of values before sequencing after sequencing; then searching the longest continuous index value sequence segment in the set c, wherein the sliding variance data segment corresponding to the sequence segment represents the most stable data segment which does not contain human body activity information in the original signal; and finally, calculating probability density distribution and probability accumulation integral functions of the sliding variance data segment by using kernel density estimation, wherein when the value of the probability accumulation integral reaches 0.975, the corresponding integral upper limit is the optimal threshold M. The window length W is set to half the packet sending rate R S;
Step six: and fifthly, obtaining a sliding variance V, and dividing the motion by the self-adaptive threshold M and the window length W. The segmentation algorithm comprises the following steps:
Step1: searching the maximum value MAX in the sliding variance, comparing the maximum value MAX with a threshold value M, if MAX is smaller than M, indicating that no action occurs in the data segment, otherwise, indicating that the data segment has actions, and performing Step2;
Step2: the index indexMax of MAX is recorded while indexRt = indexLt = indexMax is defined. Searching an index value set L 1 corresponding to V l greater than a threshold M in a section [ indexRt, indexRt +N ], if the set L 1 is empty, indicating that the operation is finished, and the finishing time point is indexRt, if the set is not empty, searching the largest element iMax in the set, assigning the value of iMax to indexRt, and searching an index value set L 1 corresponding to V l greater than the threshold M in a new section [ indexRt, indexRt +N ]. After this step, the end time point indexRt of the action can be found;
Step3: searching an index value set L 2 corresponding to V l being greater than a threshold value M in a section [ indexLt-N, indexLt ], if the set L 2 is empty, indicating that the operation starts, and the starting time point is indexLt, if the set is not empty, searching the smallest element iMin in the set, assigning the value of iMin to indexLt, and searching the index value set L 2 corresponding to V l being greater than the threshold value M in a new section [ indexLt-N, indexLt ]. After this step, the starting time point indexLt of the action can be found out;
step4: setting V l corresponding to the interval [ indexLt, indexRt ] to 0, and then carrying out Step1, step2 and Step3;
Through the steps, the data segment when the action occurs can be effectively extracted from the original data, and meanwhile, the algorithm is still effective in dividing a plurality of discontinuous actions contained in the original data.
Advantageous effects
Firstly, the invention provides a method for removing trend by utilizing piecewise fitting to process a time sequence, so that the data fluctuation of an action occurrence period and the data fluctuation of a non-action occurrence period still have larger difference while eliminating the overall trend of the sequence; then, a sliding variance is obtained for the time series after trend removal, and a window variance comparison method is proposed according to fluctuation change of the sliding variance, so that a CSI data segment only comprising action information is extracted. The actual measurement result shows that the target behavior segmentation algorithm designed by the invention can accurately and effectively extract the data segment when the action occurs from the original data, and meanwhile, the segmentation of a plurality of discontinuous actions contained in the original data by using the algorithm is still effective.
Drawings
FIG. 1 is a flow chart showing the implementation of target behavior segmentation.
Fig. 2 is a schematic diagram of a segmentation result.
Detailed description of the preferred embodiments
Step one: preprocessing the CSI data;
assuming that CSI data acquired by a receiver is subjected to data analysis and channel correction to obtain a matrix H, and preprocessing the matrix to obtain a matrix A;
Where N is the number of received data packets, N tx and N rx are the number of transmit antennas and receive antennas, respectively, and N sub is the number of subcarriers. The specific flow is as follows:
firstly, taking a module of each data in the matrix H, removing phase information in the data, then, carrying out decentralization processing on each row vector, eliminating static components of signals, and calculating the following equation:
Wherein, |·| represents modulo arithmetic;
Step two: the matrix A is subjected to dimension reduction by using a principal component analysis method, noise is suppressed, and meanwhile, the CSI amplitude containing the action information is more prominent, and the method is as follows;
Firstly, calculating covariance matrix of matrix A, then utilizing eigenvalue decomposition to obtain eigenvalue of said covariance matrix and correspondent eigenvector, finally utilizing transposed matrix A to multiply ith eigenvector to obtain ith principal component,
hi＝AT×Ei
Wherein h i is the ith principal component of the matrix A, the dimension is Nx1, E i is the eigenvector corresponding to the ith eigenvalue of the covariance matrix, and the dimension is N sub·Ntx·Nrx x 1. In the algorithm, a second principal component h 2 of the matrix a is mainly used;
Step three: processing the second principal component h 2 obtained in the step three by using a piecewise fitting trending method to obtain a vector The method comprises the following specific steps of;
First, the second principal component h 2 is divided into a plurality of data segments with a length of M, namely
h2＝[h2(W1),h2(W2),…,h2(Wk),…,h2(WN/M)]
Where W k is an array [ kM, kM+1, …, (k+1) M-1], representing the index value of each data in the kth data segment. Then fitting each segment of data respectively, wherein the following formula is as follows:
T(Wk)＝h2(Wk)-fit(h2(Wk))
wherein fit (·) is a polynomial fit based on least squares method, resulting in:
T＝[T(W1),T(W2),…,T(Wk),…,T(WN/M)]
When the point to be fitted and the highest degree of polynomial fitting are given, each coefficient of the fitting polynomial can be obtained, and then the fitting residual is obtained by subtracting the fitted point from the point before fitting, so as to form a vector T. The subsequent data processing is based on this data.
Step four: setting a sliding window with the length of K, and calculating a sliding variance V of T, wherein V= [ V l]1×(N-K+1);
Assuming that the data contained in the first sliding window is [ T l,Tl+1,…,Tl+K-1 ], the variance of the first sliding window data can be expressed as:
Step five: calculating a threshold M and setting a window length W;
According to step four, we find the sliding variance V, and then calculate an adaptive threshold M for motion segmentation based on the sliding variance. The specific calculation idea is as follows:
Firstly, carrying out ascending arrangement on V, and searching an original index value set c with the smaller half of values before sequencing after sequencing; then searching the longest continuous index value sequence segment in the set c, wherein the sliding variance data segment corresponding to the sequence segment represents the most stable data segment which does not contain human body activity information in the original signal; and finally, calculating probability density distribution and probability accumulation integral functions of the sliding variance data segment by using kernel density estimation, wherein when the value of the probability accumulation integral reaches 0.975, the corresponding integral upper limit is the optimal threshold M. The window length W is set to half the packet sending rate R S;
Step six: and fifthly, obtaining a sliding variance V, and dividing the motion by the self-adaptive threshold M and the window length W. The segmentation algorithm comprises the following steps:
Step1: searching the maximum value MAX in the sliding variance, comparing the maximum value MAX with a threshold value M, if MAX is smaller than M, indicating that no action occurs in the data segment, otherwise, indicating that the data segment has actions, and performing Step2;
Step2: the index indexMax of MAX is recorded while indexRt = indexLt = indexMax is defined. Searching an index value set L 1 corresponding to V l greater than a threshold M in a section [ indexRt, indexRt +N ], if the set L 1 is empty, indicating that the operation is finished, and the finishing time point is indexRt, if the set is not empty, searching the largest element iMax in the set, assigning the value of iMax to indexRt, and searching an index value set L 1 corresponding to V l greater than the threshold M in a new section [ indexRt, indexRt +N ]. After this step, the end time point indexRt of the action can be found;
Step3: searching an index value set L 2 corresponding to V l being greater than a threshold value M in a section [ indexLt-N, indexLt ], if the set L 2 is empty, indicating that the operation starts, and the starting time point is indexLt, if the set is not empty, searching the smallest element iMin in the set, assigning the value of iMin to indexLt, and searching the index value set L 2 corresponding to V l being greater than the threshold value M in a new section [ indexLt-N, indexLt ]. After this step, the starting time point indexLt of the action can be found out;
step4: setting V l corresponding to the interval [ indexLt, indexRt ] to 0, and then carrying out Step1, step2 and Step3;
Through the steps, the data segment when the action occurs can be effectively extracted from the original data, and meanwhile, the algorithm is still effective in dividing a plurality of discontinuous actions contained in the original data.
Claims (1)
1. A target behavior segmentation method based on Wi-Fi channel state information comprises the following steps:
Step one: preprocessing the CSI data;
Assuming that the CSI data acquired by the receiver is subjected to data analysis and channel correction to obtain a matrix H, and preprocessing the matrix to obtain a matrix A
Wherein N is the number of received data packets, N tx and N rx are the number of transmitting antennas and receiving antennas, respectively, N sub is the number of subcarriers, and the specific process of preprocessing is as follows:
firstly, taking a module of each data in the matrix H, removing phase information in the data, then, carrying out decentralization processing on each row vector, eliminating static components of signals, and calculating the following equation:
Wherein, |·| represents modulo arithmetic;
Step two: the matrix A is subjected to dimension reduction by using a principal component analysis method, noise is suppressed, and meanwhile, the CSI amplitude containing the action information is more prominent, and the method is as follows:
Firstly, calculating covariance matrix of matrix A, then utilizing eigenvalue decomposition to obtain eigenvalue of said covariance matrix and correspondent eigenvector, finally utilizing transposed matrix A to multiply ith eigenvector to obtain ith principal component,
hi＝AT×Ei
Wherein h i is the ith principal component of matrix A, the dimension is Nx1, E i is the eigenvector corresponding to the ith eigenvalue of the covariance matrix, the dimension is N sub·Ntx·Nrx x 1, and in the principal component analysis method, the second principal component h 2 of matrix A is used;
Step three: processing the second principal component h 2 obtained in the step two by using a piecewise fitting trending method to obtain a vector T,
Step four: setting a sliding window with the length of K, and calculating a sliding variance V of T, wherein V= [ V l]1×(N-K+1);
Assuming that the data contained in the first sliding window is [ T l,Tl+1,…,Tl+K-1 ], the variance of the first sliding window data can be expressed as:
Step five: calculating a threshold M and setting a window length W;
according to the fourth step, the sliding variance V is obtained, and an adaptive threshold M is calculated according to the sliding variance, for motion segmentation, and the specific calculation method is as follows:
Firstly, carrying out ascending arrangement on V, and searching an original index value set c with the smaller half of values before sequencing after sequencing; then searching the longest continuous index value sequence segment in the set c, wherein the sliding variance data segment corresponding to the sequence segment represents the most stable data segment which does not contain human body activity information in the original signal; finally, calculating probability density distribution and probability accumulation integral function of the sliding variance data segment by using kernel density estimation, wherein when the value of the probability accumulation integral reaches 0.975, the corresponding upper integral limit is the optimal threshold M, and the window length W is set to be half of the packet sending rate R S;
Step six: according to the fifth step, the sliding variance V is obtained, the motion is segmented by the self-adaptive threshold M and the window length W, and the method comprises the following steps:
Step1: searching the maximum value MAX in the sliding variance, comparing the maximum value MAX with a threshold value M, if MAX is smaller than M, indicating that no action occurs in the data segment, otherwise, indicating that the data segment has actions, and performing Step2;
Step2: recording an index indexMax of MAX, defining indexRt = indexLt = indexMax, searching an index value set L 1 corresponding to V l being larger than a threshold M in a section [ indexRt, indexRt +N ], if the set L 1 is empty, indicating that the operation is finished, and the ending time point is indexRt, if the set is not empty, searching the largest element iMax in the set, assigning the value of iMax to indexRt, searching an index value set L 1 corresponding to V l being larger than the threshold M in a new section [ indexRt, indexRt +N ], and searching the ending time point indexRt of the operation after the step;
Step3: searching an index value set L 2 corresponding to V l being greater than a threshold value M in a section [ indexLt-N, indexLt ], if the set L 2 is empty, indicating that the action starts, and the starting time point is indexLt, if the set is not empty, searching the smallest element iMin in the set, assigning the value of iMin to indexLt, searching an index value set L 2 corresponding to V l being greater than the threshold value M in a new section [ indexLt-N, indexLt ], and searching the starting time point indexLt of the action after the step;
step4: setting V l corresponding to the interval [ indexLt, indexRt ] to 0, and then carrying out Step1, step2 and Step3;
Through the steps, the data segment when the action occurs can be effectively extracted from the original data, and meanwhile, the steps are used for dividing the original data into a plurality of discontinuous actions, so that the segmentation is still effective;
The third step is as follows:
Firstly, dividing a main component h 2 into a plurality of data segments with equal length, then carrying out polynomial fitting on each data segment by using a least square method, finally, making a difference between each data segment and corresponding fitting data, and sequentially recombining fitting error data to form a vector T, wherein the specific flow is as follows:
First, the second principal component h 2 is divided into a plurality of data segments with a length of M, namely
h2＝[h2(W1),h2(W2),…,h2(Wk),…,h2(WN/M)]
Wherein W k is an array [ kM, kM+1, …, (k+1) M-1] representing the index value of each data in the kth data segment, and then fitting the data of each segment separately, as follows:
T(Wk)＝h2(Wk)-fit(h2(Wk))
wherein fit (·) is a polynomial fit based on least squares method, resulting in:
T＝[T(W1),T(W2),…,T(Wk),…,T(WN/M)]
the following is the implementation process of polynomial fitting:
assuming m points in the data segment, each data point is (x i,yi), the fitting result is:
Wherein a 0,…,an is the unknown to be solved, n is the highest power of the polynomial, and the sum of squares of the deviations is minimized according to the least square method, i.e
Solving the above coefficients can be converted into an extremum problem of i=i (a 0,a1,…,an), and solving the extremum according to the requirement of the multiple function:
I.e.
The representation in matrix form is then:
when the point to be fitted and the highest degree of polynomial fitting are given, each coefficient of the fitting polynomial can be obtained, then the fitting residual is obtained by subtracting the fitted point from the point before fitting, a vector T is formed, and the subsequent data processing is based on the data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110076299.9A CN112765550B (en) | 2021-01-20 | 2021-01-20 | Target behavior segmentation method based on Wi-Fi channel state information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110076299.9A CN112765550B (en) | 2021-01-20 | 2021-01-20 | Target behavior segmentation method based on Wi-Fi channel state information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112765550A CN112765550A (en) | 2021-05-07 |
| CN112765550B true CN112765550B (en) | 2024-05-07 |
Family
ID=75701753
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110076299.9A Active CN112765550B (en) | 2021-01-20 | 2021-01-20 | Target behavior segmentation method based on Wi-Fi channel state information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112765550B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114465678A (en) * | 2022-04-13 | 2022-05-10 | 齐鲁工业大学 | Complex activity WIFI perception method based on deep learning |
| CN114900257B (en) * | 2022-05-26 | 2024-05-14 | Oppo广东移动通信有限公司 | Baseband chip, channel estimation method, data processing method and equipment |
| CN115219067B (en) * | 2022-09-20 | 2023-01-03 | 金乡县成启仓储服务有限公司 | Real-time state monitoring method for garlic storage |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015054419A1 (en) * | 2013-10-08 | 2015-04-16 | University Of Washington Through Its Center For Commercialization | Devices, systems, and methods for controlling devices using gestures |
| CN106658590A (en) * | 2016-12-28 | 2017-05-10 | 南京航空航天大学 | Design and implementation of multi-person indoor environment state monitoring system based on WiFi channel state information |
| CN109657572A (en) * | 2018-12-04 | 2019-04-19 | 重庆邮电大学 | Goal behavior recognition methods after a kind of wall based on Wi-Fi |
| CN109947238A (en) * | 2019-01-17 | 2019-06-28 | 电子科技大学 | A method of the non-cooperative gesture identification based on WIFI |
| CN110062378A (en) * | 2019-04-15 | 2019-07-26 | 哈尔滨工程大学 | Identity identifying method based on channel state information under a kind of gesture scene |
| CN110110580A (en) * | 2019-03-12 | 2019-08-09 | 西北大学 | A kind of network struction of sign language isolated word recognition and classification method towards Wi-Fi signal |
| WO2020093278A1 (en) * | 2018-11-07 | 2020-05-14 | 华为技术有限公司 | Multi-antenna based gesture recognition method and device |
| CN111901028A (en) * | 2020-07-24 | 2020-11-06 | 辽宁工程技术大学 | Human body behavior identification method based on CSI (channel State information) on multiple antennas |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107633227B (en) * | 2017-09-15 | 2020-04-28 | 华中科技大学 | CSI-based fine-grained gesture recognition method and system |
-
2021
- 2021-01-20 CN CN202110076299.9A patent/CN112765550B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015054419A1 (en) * | 2013-10-08 | 2015-04-16 | University Of Washington Through Its Center For Commercialization | Devices, systems, and methods for controlling devices using gestures |
| CN106658590A (en) * | 2016-12-28 | 2017-05-10 | 南京航空航天大学 | Design and implementation of multi-person indoor environment state monitoring system based on WiFi channel state information |
| WO2020093278A1 (en) * | 2018-11-07 | 2020-05-14 | 华为技术有限公司 | Multi-antenna based gesture recognition method and device |
| CN109657572A (en) * | 2018-12-04 | 2019-04-19 | 重庆邮电大学 | Goal behavior recognition methods after a kind of wall based on Wi-Fi |
| CN109947238A (en) * | 2019-01-17 | 2019-06-28 | 电子科技大学 | A method of the non-cooperative gesture identification based on WIFI |
| CN110110580A (en) * | 2019-03-12 | 2019-08-09 | 西北大学 | A kind of network struction of sign language isolated word recognition and classification method towards Wi-Fi signal |
| CN110062378A (en) * | 2019-04-15 | 2019-07-26 | 哈尔滨工程大学 | Identity identifying method based on channel state information under a kind of gesture scene |
| CN111901028A (en) * | 2020-07-24 | 2020-11-06 | 辽宁工程技术大学 | Human body behavior identification method based on CSI (channel State information) on multiple antennas |
Non-Patent Citations (3)
| Title |
|---|
| A Survey on Behavior Recognition Using WiFi Channel State Information;Siamak Yousefi;《IEEE Communications Magazine》;第55卷(第10期);全文 * |
| Activity Segmentation by Using Window Variance Comparison Method;Xinxing Tang;《International Conference in Communications, Signal Processing, and Systems》;全文 * |
| Research on Human Activity Recognition Technology under the Condition of Through-the-wall;Ruoyu Cao;《2020 IEEE/CIC International Conference on Communications in China (ICCC)》;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112765550A (en) | 2021-05-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112765550B (en) | Target behavior segmentation method based on Wi-Fi channel state information | |
| CN106658590B (en) | Design and implementation of multi-person indoor environment state monitoring system based on WiFi channel state information | |
| JP2019133639A5 (en) | ||
| CN110865357B (en) | Laser radar echo signal noise reduction method based on parameter optimization VMD | |
| CN109063615B (en) | Sign language identification method and system | |
| CN102819043B (en) | Array signal random noise adaptive model denoising method | |
| CN111008991B (en) | Background-aware related filtering target tracking method | |
| CN110929842B (en) | Accurate intelligent detection method for non-cooperative radio signal burst time region | |
| CN110062379B (en) | Identity authentication method based on channel state information under human behavior scene | |
| CN113259031B (en) | Bottom noise fitting and eliminating method for satellite spectrum situation generation | |
| CN109902554B (en) | Sign language identification method based on commercial Wi-Fi | |
| CN111901028B (en) | Human body behavior identification method based on CSI (channel State information) on multiple antennas | |
| CN110702986B (en) | Method and system for dynamically generating self-adaptive signal search threshold in real time | |
| CN105656577B (en) | Towards the cluster-dividing method and device of channel impulse response | |
| CN111882057A (en) | Multi-stage deep learning method oriented to space-time sequence data modeling and application | |
| CN116980824B (en) | Lightweight weighting integrated learning indoor CSI positioning method | |
| CN114680879B (en) | Method and device for detecting body motion artifact signal | |
| CN115577246A (en) | Method for detecting anti-vibration performance of gas cylinder protective cover | |
| CN110176968B (en) | Jump phenomenon correction method for WiFi human behavior recognition | |
| CN113824518A (en) | Non-cooperative signal detection method based on frequency spectrum | |
| CN114900406B (en) | Blind modulation signal identification method based on twin network | |
| CN110376436A (en) | A kind of multiple dimensioned line-spectrum detection method for noise power spectra | |
| CN110623662B (en) | Adaptive threshold iterative algorithm suitable for electrocardiosignal detection | |
| CN117584991B (en) | Mining area unmanned vehicle outside personnel safety protection method and system | |
| CN117194892A (en) | Method and device for determining peak value and valley value of signal and electronic equipment |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20240201 Address after: 1003, Building A, Zhiyun Industrial Park, No. 13 Huaxing Road, Henglang Community, Dalang Street, Longhua District, Shenzhen City, Guangdong Province, 518000 Applicant after: Shenzhen Wanzhida Technology Transfer Center Co.,Ltd. Country or region after: China Address before: 400065 No. 2, Chongwen Road, Nan'an District, Chongqing Applicant before: CHONGQING University OF POSTS AND TELECOMMUNICATIONS Country or region before: China |
|
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20240408 Address after: 984, 34th Floor, Building 3, Area A4, Hanyu Jingu, Jingshi Road, Jinan Area, China (Shandong) Pilot Free Trade Zone, Jinan City, Shandong Province, 250000 Applicant after: Jinan Jerui Information Technology Co.,Ltd. Country or region after: China Address before: 1003, Building A, Zhiyun Industrial Park, No. 13 Huaxing Road, Henglang Community, Dalang Street, Longhua District, Shenzhen City, Guangdong Province, 518000 Applicant before: Shenzhen Wanzhida Technology Transfer Center Co.,Ltd. Country or region before: China |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |