CN108682140A - A kind of enhanced method for detecting abnormality based on compressed sensing and autoregression model - Google Patents
A kind of enhanced method for detecting abnormality based on compressed sensing and autoregression model Download PDFInfo
- Publication number
- CN108682140A CN108682140A CN201810367791.XA CN201810367791A CN108682140A CN 108682140 A CN108682140 A CN 108682140A CN 201810367791 A CN201810367791 A CN 201810367791A CN 108682140 A CN108682140 A CN 108682140A
- Authority
- CN
- China
- Prior art keywords
- node
- cluster
- monitoring data
- data
- moment
- 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
- 238000000034 method Methods 0.000 title claims abstract description 19
- 230000005856 abnormality Effects 0.000 title claims abstract description 7
- 238000012544 monitoring process Methods 0.000 claims abstract description 101
- 230000002776 aggregation Effects 0.000 claims abstract description 22
- 238000004220 aggregation Methods 0.000 claims abstract description 22
- 238000001514 detection method Methods 0.000 claims abstract description 22
- 230000002547 anomalous effect Effects 0.000 claims abstract description 12
- 238000005259 measurement Methods 0.000 claims abstract description 7
- 230000002123 temporal effect Effects 0.000 claims abstract description 6
- 241000854291 Dianthus carthusianorum Species 0.000 claims description 43
- 230000002159 abnormal effect Effects 0.000 claims description 23
- 230000008447 perception Effects 0.000 claims description 11
- 238000012360 testing method Methods 0.000 claims description 10
- 239000011159 matrix material Substances 0.000 claims description 9
- 238000005070 sampling Methods 0.000 claims description 8
- 230000005540 biological transmission Effects 0.000 claims description 7
- 238000013480 data collection Methods 0.000 claims description 4
- 230000004807 localization Effects 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims 1
- 241000255789 Bombyx mori Species 0.000 abstract 2
- 239000010902 straw Substances 0.000 abstract 2
- 230000006835 compression Effects 0.000 description 3
- 238000007906 compression Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08C—TRANSMISSION SYSTEMS FOR MEASURED VALUES, CONTROL OR SIMILAR SIGNALS
- G08C25/00—Arrangements for preventing or correcting errors; Monitoring arrangements
-
- G—PHYSICS
- G08—SIGNALLING
- G08C—TRANSMISSION SYSTEMS FOR MEASURED VALUES, CONTROL OR SIMILAR SIGNALS
- G08C17/00—Arrangements for transmitting signals characterised by the use of a wireless electrical link
- G08C17/02—Arrangements for transmitting signals characterised by the use of a wireless electrical link using a radio link
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/30—Services specially adapted for particular environments, situations or purposes
- H04W4/38—Services specially adapted for particular environments, situations or purposes for collecting sensor information
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W84/00—Network topologies
- H04W84/18—Self-organising networks, e.g. ad-hoc networks or sensor networks
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Signal Processing (AREA)
- Testing And Monitoring For Control Systems (AREA)
- Testing Or Calibration Of Command Recording Devices (AREA)
Abstract
The present invention proposes a kind of enhanced method for detecting abnormality based on compressed sensing and autoregression model.A small bundle of straw, etc. for silkworms to spin cocoons on head carries out preliminary abnormality detection using compressed sensing to the monitoring data at a small bundle of straw, etc. for silkworms to spin cocoons on inner sensor node moment first;Then the temporal correlation for utilizing sensor node monitoring data, accurately detects the thundering observed data at sensor node moment in conjunction with autoregression model;Last aggregation node utilizes the spatial coherence of each cluster interior nodes monitoring data, to judge that thundering observed data is and to position the generation area of anomalous event caused by anomalous event or measurement error.The present invention can improve the accuracy of accident detection in wireless sensor network, reduce false alarm rate and judge the incidence of event by accident, have wide applicability.
Description
Technical field
Present invention relates generally to the accident detection fields in wireless sensor network.
Background technology
Accident detection is a kind of important application of wireless sensor network.When anomalous event (such as chemical spills,
Fire etc.) occur after, sensor node should be able to detect the region that outgoing event may occur as early as possible, and in time to aggregation node
Report.
Compressed sensing is that the data collection of distributed sensing network opens new thinking.Sparse data is adopted first
Sample, the data then obtained to sampling carry out compression measurement, obtain measured value, finally carry out weight to initial data by measured value
Structure.The theory reduces the sample frequency of signal, also reduces data space and network transmission volume.Since anomalous event is
Event occurs for small probability, therefore compressed sensing provides thinking for the detection of anomalous event in wireless sensor network.However, single
Node data is reconstructed by the restructing algorithm in compressed sensing purely, and is directly sent out come detection node by reconstruction result
The shortcomings that whether data sent are abnormal, and there are testing results inaccurately, false alarm rate rises.In addition, wireless sensor network is in thing
The main problem that is faced, which is accuracy of detection, in part monitoring is influenced by ambient noise and equipment are instable, if only according to one
The perception data at a some time point of separate nodes judges event, then easily causes in wireless sensor network and judge by accident
The generation of event, and the monitoring data at each node each moment are detected using the method for time series, it needs to count
The amount of calculation can be very big.
In conclusion carrying out Preliminary detection to the anomalous event in wireless sensor network node using compressed sensing
On the basis of, how in conjunction with node autoregression model to improve accuracy of detection, and how to utilize the spatial coherence of cluster interior nodes
It measurement error and anomalous event is identified, positions, at present the still not no solution of science.
Invention content
In view of the above-mentioned problems, proposing a kind of in wireless sensor network based on compressed sensing and autoregression model
Enhanced method for detecting abnormality, is as follows:
Step 1: the arrangement of network scenarios and the initialization process of network:
1, the sensor node for being N in monitoring region random distribution quantity;All the sensors node is having the same initial
Energy and transmission rate;All the sensors node can obtain itself geographical location information by localization methods such as GPS;
2, according to the distribution of anomalous event, the monitoring region of entire wireless sensor network is divided into C cluster, and choose
Cluster head and aggregation node form Cluster Networks, and the sensor node number of j-th of cluster head monitoring is Nj, j=1,2 ..., C.
Step 2: being based on history normal data, cluster head monitors each sensor node in region for it and establishes one certainly
Regression model:
1, within certain time, the monitoring data of each sensor node are stable, i.e. i-th of node t moments
Monitoring dataWith the monitoring data at t+1 momentThere is identical distribution;
2, the monitoring data of each sensor node meet the law of large numbers, i.e., for arbitrary T moment, each sensor
The monitoring data value of node converges on the desired value of all monitoring data, and desired value is:
3, the normal historical data monitored using each node is i-th of joint structure, one p as prior informationiRank
Autoregression model:
Wherein,For the residual error of model t moment, obedience mean value is μi, variance isNormal distribution,Point
The p before i-th of node t moment is not indicatediInfluence intensity of the monitoring data at a moment to current t moment monitoring data;
Step 3: in t moment, node perceived monitoring data simultaneously carry out binaryzation to it, according to preset threshold θ,
IfThenOtherwiseThe monitoring data after binaryzation are obtained, at this time the node data vector in i-th of cluster
ForWhereinValue be 0 or 1;
Step 4: cluster head carries out compression sampling to the node data vector in t moment cluster, it then will sampling the data obtained road
By to aggregation node, the node data vector in each cluster is reconstructed in aggregation node, Preliminary detection is obtained as a result, last cluster
Head is collected abnormal monitoring data:
1, random observation matrix is generated according to j-th of cluster headWherein M is the vector dimension that compressed sensing obtains, Nj
For j-th of cluster head monitoring sensor node number, by the monitoring data vector projection of its cluster interior nodes to observing matrix
On, perception data sequence is obtained, and perception data sequence is routed to aggregation node, the perception data sequence of j-th of cluster head routing
It is classified as
2, aggregation node is according to the perception data sequence received, using OMP algorithms to the node data vector in each cluster
It is reconstructed;
3, the monitoring data that reconstruct obtains in each cluster are that its primary monitoring data is sent to by the sensor node of non-zero
Cluster head, the primary monitoring data that the sensor node of wherein serial number i is sent are
Step 5: what each cluster head obtained reconstruct in each cluster of t moment according to the temporal correlation of monitoring data in cluster
Monitoring data are that the monitoring result of the sensor node (hereinafter referred to as objective sensor node) of non-zero is identified:
1, the autoregression model established for each sensor node before:
By the monitoring data and its p before this of its t momentiThe historical data at a moment can calculate its t moment according to model
Residual values
2, judgeValue whether fall in section [μi-3σi,μi+3σi] within.IfValue fall in [μi-3σi,μi+3σi]
It is interior, then illustrate that the monitoring data of the objective sensor node t moment of serial number i are normal, otherwise illustrates the target of serial number i
The monitoring data of sensor node t moment are abnormal;
3, each cluster head of t moment receives the thundering observed data that the objective sensor node of serial number i is sentAfterwards, to
Other sensors node sends transmission data request in cluster, its current monitoring data is sent to by other sensors node in cluster
Cluster head;The monitoring data collection of all the sensors node is combined into
4, the monitoring data that all nodes are sent in cluster are ranked up by cluster head, obtain intermediate value Me (if there is even number monitoring
Data then take the mean value of two number of centre);The monitoring data of each node and the collection of intermediate value Me differences are combined into
Mean valueStandard deviationThe monitoring data of i-th of nodeStandardized value
5, whenWhen more than pre-determined threshold α, i-th of node send abnormal data be caused by measurement error, it is no
Then, the region that i-th of node is monitored is abnormal situation, and cluster head of the node into cluster where it sends pre-warning signal;Respectively
A cluster head sends respective testing result to aggregation node, that is, sends the node serial number extremely convergence that monitoring region is abnormal situation
Node finally obtains a global testing result;
In conclusion advantages of the present invention is as follows:
1, on the basis of carrying out Preliminary detection to abnormal data using compressive sensing theory, by establishing autoregression mould
Type makes full use of the temporal correlation of sensor node monitoring data to carry out more accurate detection to abnormal data, reduces
False alarm rate in network keeps testing result more reliable.
2, the spatial coherence between each cluster inner sensor node monitoring data is made full use of, certain sensor node is supervised
The data that the abnormal data of survey is monitored with other sensors node in cluster where it are compared, so as to identify abnormal data
It is caused by anomalous event has occurred in the region monitored by sensor node or due to caused by measurement error.
Description of the drawings
Fig. 1 is the flow chart of the present invention;
Fig. 2 is the procedure chart for being further detected to thundering observed data and thundering observed data being identified;
Fig. 3 is inventive network model sub-clustering schematic diagram.
Specific implementation method
The present invention devises a kind of enhanced based on compressed sensing and autoregression model in wireless sensor network
Method for detecting abnormality, in conjunction with Fig. 1, the specific implementation method of accident detection is as follows:
Step 1: as shown in Fig. 2, the arrangement of network scenarios and the initialization process of network:
1, the sensor node for being N in monitoring region random distribution quantity;All the sensors node is having the same initial
Energy and transmission rate;All the sensors node can obtain itself geographical location information by localization methods such as GPS;
2, according to the distribution of anomalous event, the monitoring region of entire wireless sensor network is divided into C cluster, and choose
Cluster head and aggregation node form Cluster Networks, and the sensor node number of j-th of cluster head monitoring is Nj, j=1,2 ..., C.
Step 2: being based on history normal data, cluster head monitors each sensor node in region for it and establishes one certainly
Regression model:
1, assume within certain time, the monitoring data of each sensor node are stable, i.e. i-th of node t
The monitoring data at momentWith the monitoring data at t+1 momentThere is identical distribution;
2, the monitoring data of each sensor node meet the law of large numbers, i.e., for arbitrary T moment, each sensor
The monitoring data value at node moment all converges on the desired value of all monitoring data values, and desired value is:
3, the normal historical data monitored using each node is i-th of joint structure, one p as prior informationiRank
Autoregression model:
Wherein,For the residual error of model t moment, obedience mean value is μ i, and variance isNormal distribution,
The p before i-th of node t moment is indicated respectivelyiInfluence intensity of the monitoring data at a moment to current t moment monitoring data;
Step 3: in t moment, node perceived monitoring data simultaneously carry out binaryzation to it, according to preset threshold θ,
IfThenOtherwiseObtain the monitoring data after binaryzation, at this time the node data in i-th of cluster to
Amount isWhereinValue be 0 or 1;
Step 4: cluster head carries out compression sampling to the node data vector in t moment cluster, it then will sampling the data obtained road
By to aggregation node, operation is reconstructed to the node data vector in each cluster in aggregation node, obtains Preliminary detection as a result, most
Cluster head is collected abnormal monitoring data afterwards:
1, random observation matrix is generated according to j-th of cluster headWherein M is the vector dimension that compressed sensing obtains, Nj
For j-th of cluster head monitoring sensor node number, by the monitoring data vector projection of its cluster interior nodes to observing matrix
On, perception data sequence is obtained, and perception data sequence is routed to aggregation node, the perception data sequence of j-th of cluster head routing
It is classified as
2, aggregation node is according to the perception data sequence received, using OMP algorithms to the node data vector in each cluster
It is reconstructed;
The OMP pseudo-code of the algorithm used in the present invention is as follows:
Input:Dictionary matrix Φ, original signal y, degree of rarefication K identify the index of nonzero element position in signal to be reconstructed
Collect T
Output:Reconstruction signal x
Initialization:X=0, r0=y, cycle mark k=0, indexed set T0For empty set
When not meeting termination condition, cycle execute step 1.~6.
1. k=k+1
2. finding out the index λ of residual error r and most matched atoms in sampling matrixk, i.e.,
λk=argmaxJ=1,2 ..., N{|<rk-1,φk>|}
3. updating indexed set Tk=Tk-1∪{λk, and update the reconstruction atom set in corresponding sampling matrix
4. being obtained by least square method
5. updating residual error:
6. judging whether to meet k > K, if satisfied, then stopping iteration, if not satisfied, thening follow the steps 1.
3, the monitoring data that reconstruct obtains in each cluster are that its primary monitoring data is sent to by the sensor node of non-zero
Cluster head, the primary monitoring data that the sensor node of wherein serial number i is sent are
Step 5: cluster head according to the temporal correlations of monitoring data to the monitoring data that are reconstructed in each cluster of t moment
Monitoring result for the sensor node (hereinafter referred to as objective sensor node) of non-zero is identified:
1, the autoregression model (2) established for each sensor node before:
By the monitoring data and its p before this of its t momentiThe historical data at a moment can calculate its t moment according to model
Residual values
2, judgeValue whether fall in section [μi-3σi,μi+3σi] within.IfValue fall in [μi-3σi,μi+3σi]
It is interior, then illustrate that the monitoring data of the objective sensor node t moment of serial number i are normal, otherwise illustrates the target of serial number i
The monitoring data of sensor node t moment are abnormal;
3, each cluster head of t moment receives the thundering observed data that the objective sensor node of serial number i is sentAfterwards, to
Other sensors node sends transmission data request in cluster, its current monitoring data is sent to by other sensors node in cluster
Cluster head;The monitoring data collection of all the sensors node is combined into
4, the monitoring data that all nodes are sent in cluster are ranked up by cluster head, obtain intermediate value Me (if there is even number monitoring
Data then take the mean value of two number of centre);The monitoring data of each node are with intermediate value Me difference setsMean valueStandard deviation Standardized valueCluster head sends all nodes in cluster
Monitoring data be ranked up, obtain intermediate value Me (if there is even number monitoring data, take centre two number mean values);The prison of each node
The collection of measured data and intermediate value Me differences is combined intoMean valueStandard deviation Standardized value
5, whenWhen more than pre-determined threshold α, i-th of node send abnormal data be caused by measurement error, it is no
Then, the region that i-th of node is monitored is abnormal situation, and cluster head of the node into cluster where it sends pre-warning signal;Respectively
A cluster head sends respective testing result to aggregation node, that is, sends the node serial number extremely convergence that monitoring region is abnormal situation
Node finally obtains a global testing result.
Claims (5)
1. a kind of enhanced method for detecting abnormality based on compressed sensing and autoregression model, the method includes at least following step
Suddenly:
Step 1: the arrangement of network scenarios and the initialization process of network;
Step 2: the normal historical data monitored using each node is prior information, each cluster head is each sensing in cluster
Device node establishes an autoregression model:
Step 3: in t moment, i-th of node perceived monitoring dataAnd binaryzation is carried out to it, i.e., according to preset threshold
Value θ, ifThenOtherwise
Step 4: cluster head carries out compressed sensing to the vector of the binaryzation monitoring data composition of each node in t moment cluster, then
Sampling the data obtained is routed into aggregation node, aggregation node carries out the vector of the binaryzation monitoring data composition in each cluster
Reconstruct obtains Preliminary detection as a result, last cluster head is collected abnormal monitoring data;
Step 5: according to send the autoregression model that the node of abnormal data is established, by the currently monitored data of nodeWith
Historical Monitoring dataCalculate the residual values of "current" modelIfIt falls and then illustrates in target interval
Monitoring data are normal, are modified to the Preliminary detection result for using compressed sensing to obtain before;
Step 6: detection of each cluster head of t moment according to the temporal correlation of monitoring data in cluster to monitoring data abnormal nodes
As a result it is identified, last each cluster head sends the testing result of anomalous event in cluster, that is, monitors region and be abnormal event
Node serial number obtains a global testing result to aggregation node.
2. the accident detection method according to claim 1 based on compressed sensing and autoregression model, feature exist
It is at least further comprising the steps of in the network scenarios arrangement and the process of netinit processing:
1) sensor node for being N in monitoring region random distribution quantity;All the sensors node primary power having the same
And transmission rate;All the sensors node can obtain itself geographical location information by localization methods such as GPS;
2) according to the distribution of anomalous event, the monitoring region of entire wireless sensor network is divided into C cluster, and choose cluster head
And aggregation node, Cluster Networks are formed, the sensor node number of j-th of cluster head monitoring is Nj, j=1,2 ..., C.
3. the accident detection method according to claim 1 based on compressed sensing and autoregression model, feature exist
In the normal historical data monitored using each node as prior information, built for i-th of sensor node in each cluster
Found a piThe autoregression model of rank:
Wherein,For the residual error of model t moment, obedience mean value is μi, variance isNormal distribution,Table respectively
Show the p before i-th of node t momentiInfluence intensity of the monitoring data at a moment to current t moment monitoring data.
4. the accident detection method according to claim 1 based on compressed sensing and autoregression model, feature exist
Compressed sensing is carried out in the node data vector in each cluster of t moment, then it is reconstructed and is carried out preliminary
Detection is at least further comprising the steps of:
1) j-th of cluster head generates random observation matrixWherein M is the vector dimension that compressed sensing obtains, NjFor j-th of cluster
The sensor node number of head monitoring, by the monitoring data vector projection of its cluster interior nodes to observing matrixOn, felt
Primary data sequence, and perception data sequence is routed to aggregation node, the perception data sequence of j-th of cluster head routing is
WhereinFor the node data vector in j-th of cluster, j=1,2 ..., C;
2) aggregation node is according to the compressed sensing sequence receivedUsing OMP algorithms to the node data vector in each cluster into
Row reconstruct;
3) monitoring data that reconstruct obtains in each cluster are that its primary monitoring data is sent to cluster head by the sensor node of non-zero,
The primary monitoring data for the t moment that the sensor node of wherein serial number i is sent isI=1,2 ..., N.
5. the accident detection method according to claim 4 based on compressed sensing and autoregression model, feature exist
In the temporal correlation according to monitoring data to the sensing that the monitoring data reconstructed in each cluster of t moment are non-zero
The monitoring result of device node is identified, at least further comprising the steps of:
1) autoregression model established for each sensor node before:
Thus the residual values of its t moment are calculated
2) judgeValue whether fall in section [μi-3σi,μi+3σi] within, ifValue fall in [μi-3σi,μi+3σi] in, then
Illustrate that the monitoring data of the objective sensor node t moment of serial number i are normal, otherwise illustrates the sensor of interest of serial number i
The monitoring data of node t moment are abnormal;
3) each cluster head of t moment receives the thundering observed data that the objective sensor node of serial number i is sentAfterwards, into cluster
Other sensors node sends transmission data request, its monitoring data in t moment is sent to by other sensors node in cluster
Cluster head;The monitoring data collection of all the sensors node t moment is combined into
4) monitoring data that all nodes in cluster are sent are ranked up by cluster head, obtain intermediate value Me (if there is even number monitoring data,
Then take the mean value of two number of centre);The monitoring data of each node and the collection of intermediate value Me differences are combined intoMean valueStandard deviation Standardized value
5) whenWhen more than pre-determined threshold α, abnormal data that i-th of node is sent be caused by measurement error, otherwise, the
The region that i node is monitored is abnormal situation, and cluster head of the node into cluster where it sends pre-warning signal;Each cluster head
Respective testing result is sent to aggregation node, that is, sends monitoring region and is abnormal the node serial number of situation to aggregation node,
Finally obtain a global testing result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810367791.XA CN108682140B (en) | 2018-04-23 | 2018-04-23 | Enhanced anomaly detection method based on compressed sensing and autoregressive model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810367791.XA CN108682140B (en) | 2018-04-23 | 2018-04-23 | Enhanced anomaly detection method based on compressed sensing and autoregressive model |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108682140A true CN108682140A (en) | 2018-10-19 |
CN108682140B CN108682140B (en) | 2020-07-28 |
Family
ID=63802144
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810367791.XA Expired - Fee Related CN108682140B (en) | 2018-04-23 | 2018-04-23 | Enhanced anomaly detection method based on compressed sensing and autoregressive model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108682140B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109348450A (en) * | 2018-11-01 | 2019-02-15 | 北京京航计算通讯研究所 | High reliability data transmission method applied to emergency disposal |
CN109492193A (en) * | 2018-12-28 | 2019-03-19 | 同济大学 | Abnormal network data based on depth machine learning model generate and prediction technique |
CN109921952A (en) * | 2019-04-01 | 2019-06-21 | 安徽农业大学 | A kind of method of data capture based on compressed sensing and model-driven |
CN111189488A (en) * | 2019-12-13 | 2020-05-22 | 精英数智科技股份有限公司 | Sensor value abnormity identification method, device, equipment and storage medium |
CN112672302A (en) * | 2020-12-21 | 2021-04-16 | 国网甘肃省电力公司电力科学研究院 | Clustering and data sensing method applied to photovoltaic power station wireless sensor |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101511099A (en) * | 2009-04-01 | 2009-08-19 | 南京邮电大学 | Collection method for wireless sensor network data based on time series prediction model |
US20100091834A1 (en) * | 2007-05-11 | 2010-04-15 | The Trustees Of Columbia University In The City Of New York | Systems and methods for telescopic data compression in sensor networks |
CN102594904A (en) * | 2012-03-04 | 2012-07-18 | 浙江大学 | Method for detecting abnormal events of wireless sensor network in distributed way |
EP3016352A1 (en) * | 2014-11-03 | 2016-05-04 | Fujitsu Limited | Method of managing sensor network |
CN105636094A (en) * | 2016-03-16 | 2016-06-01 | 中国地质大学(武汉) | Wireless sensor network early warning method and system based on clustering compressed sensing |
TW201619582A (en) * | 2014-11-25 | 2016-06-01 | 國立臺灣大學 | Method for vibration monitoring and alarming using autoregressive models |
CN105682171A (en) * | 2016-03-09 | 2016-06-15 | 北京航空航天大学 | Spatio-temporal clustering method for compressive data gathering |
CN106792435A (en) * | 2016-11-23 | 2017-05-31 | 西安电子科技大学 | Data aggregation method based on compressed sensing in a kind of wireless sensor network |
-
2018
- 2018-04-23 CN CN201810367791.XA patent/CN108682140B/en not_active Expired - Fee Related
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100091834A1 (en) * | 2007-05-11 | 2010-04-15 | The Trustees Of Columbia University In The City Of New York | Systems and methods for telescopic data compression in sensor networks |
CN101511099A (en) * | 2009-04-01 | 2009-08-19 | 南京邮电大学 | Collection method for wireless sensor network data based on time series prediction model |
CN102594904A (en) * | 2012-03-04 | 2012-07-18 | 浙江大学 | Method for detecting abnormal events of wireless sensor network in distributed way |
EP3016352A1 (en) * | 2014-11-03 | 2016-05-04 | Fujitsu Limited | Method of managing sensor network |
TW201619582A (en) * | 2014-11-25 | 2016-06-01 | 國立臺灣大學 | Method for vibration monitoring and alarming using autoregressive models |
CN105682171A (en) * | 2016-03-09 | 2016-06-15 | 北京航空航天大学 | Spatio-temporal clustering method for compressive data gathering |
CN105636094A (en) * | 2016-03-16 | 2016-06-01 | 中国地质大学(武汉) | Wireless sensor network early warning method and system based on clustering compressed sensing |
CN106792435A (en) * | 2016-11-23 | 2017-05-31 | 西安电子科技大学 | Data aggregation method based on compressed sensing in a kind of wireless sensor network |
Non-Patent Citations (3)
Title |
---|
CARLO CAIONE 等: "Distributed Compressive Sampling for Lifetime Optimization in Dense Wireless Sensor Networks", 《IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS》 * |
刘洲洲 等: "采用压缩感知和GM(1,1)的无线传感器网络异常检测方法", 《西安交通大学学报》 * |
朱更明 等: "基于压缩感知的无线传感网数据收集研究进展", 《西北大学学报(自然科学版)》 * |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109348450A (en) * | 2018-11-01 | 2019-02-15 | 北京京航计算通讯研究所 | High reliability data transmission method applied to emergency disposal |
CN109348450B (en) * | 2018-11-01 | 2021-05-25 | 北京京航计算通讯研究所 | High-reliability data transmission method applied to emergency disposal |
CN109492193A (en) * | 2018-12-28 | 2019-03-19 | 同济大学 | Abnormal network data based on depth machine learning model generate and prediction technique |
CN109492193B (en) * | 2018-12-28 | 2020-11-27 | 同济大学 | Abnormal network data generation and prediction method based on deep machine learning model |
CN109921952A (en) * | 2019-04-01 | 2019-06-21 | 安徽农业大学 | A kind of method of data capture based on compressed sensing and model-driven |
CN109921952B (en) * | 2019-04-01 | 2021-11-23 | 安徽农业大学 | Data collection method based on compressed sensing and model driving |
CN111189488A (en) * | 2019-12-13 | 2020-05-22 | 精英数智科技股份有限公司 | Sensor value abnormity identification method, device, equipment and storage medium |
CN111189488B (en) * | 2019-12-13 | 2020-12-04 | 精英数智科技股份有限公司 | Sensor value abnormity identification method, device, equipment and storage medium |
CN112672302A (en) * | 2020-12-21 | 2021-04-16 | 国网甘肃省电力公司电力科学研究院 | Clustering and data sensing method applied to photovoltaic power station wireless sensor |
CN112672302B (en) * | 2020-12-21 | 2022-07-26 | 国网甘肃省电力公司电力科学研究院 | Clustering and data sensing method applied to photovoltaic power station wireless sensor |
Also Published As
Publication number | Publication date |
---|---|
CN108682140B (en) | 2020-07-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108682140A (en) | A kind of enhanced method for detecting abnormality based on compressed sensing and autoregression model | |
Muhammed et al. | An analysis of fault detection strategies in wireless sensor networks | |
CN104205063B (en) | Operation administration device, operation administration method, and program | |
US8638680B2 (en) | Applying policies to a sensor network | |
US20080222068A1 (en) | Inferring Candidates that are Potentially Responsible for User-Perceptible Network Problems | |
Yuan et al. | A distributed Bayesian algorithm for data fault detection in wireless sensor networks | |
US20120026938A1 (en) | Applying Policies to a Sensor Network | |
CN106017729B (en) | A kind of motor temperature monitoring method based on statistical Process Control | |
CN106464506A (en) | Method and apparatus for registering devices in gateway | |
Chitradevi et al. | Outlier aware data aggregation in distributed wireless sensor network using robust principal component analysis | |
CN105915388A (en) | Sensor fault detection method and system based on distributed network | |
CN106452877B (en) | Fault positioning method for power information network | |
WO2012086824A1 (en) | Operation management device, operation management method, and program | |
Abid et al. | Anomaly detection through outlier and neighborhood data in Wireless Sensor Networks | |
CN109714402A (en) | A kind of redundant data acquisition system and its operation application method | |
Karthik et al. | Data trust model for event detection in wireless sensor networks using data correlation techniques | |
CN111884859B (en) | Network fault diagnosis method and device and readable storage medium | |
RU2013130664A (en) | METHOD FOR PERFORMING DIAGNOSTIC OF STRUCTURE SUBJECT TO LOADS AND SYSTEM FOR IMPLEMENTATION OF THE MENTIONED METHOD | |
CN108650649A (en) | Abnormal deviation data examination method suitable for wireless sensor network | |
CN104486113A (en) | Fault link positioning method based on active greed and passive greed in sensor network | |
Liu et al. | Fault tolerant complex event detection in WSNs: A case study in structural health monitoring | |
CN110266527B (en) | Sensor node fault classification alarm method and device based on spatial correlation | |
CN105407496B (en) | A kind of method of erroneous measurements in identification wireless sensor network | |
CN112734977B (en) | Equipment risk early warning system and algorithm based on Internet of things | |
Abid et al. | Centralized KNN anomaly detector for WSN |
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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200728 |
|
CF01 | Termination of patent right due to non-payment of annual fee |