CN104717689B - A kind of method for diagnosing faults of wireless sensor network - Google Patents
A kind of method for diagnosing faults of wireless sensor network Download PDFInfo
- Publication number
- CN104717689B CN104717689B CN201510109164.2A CN201510109164A CN104717689B CN 104717689 B CN104717689 B CN 104717689B CN 201510109164 A CN201510109164 A CN 201510109164A CN 104717689 B CN104717689 B CN 104717689B
- Authority
- CN
- China
- Prior art keywords
- network
- failure
- wireless sensor
- observing matrix
- event
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W24/00—Supervisory, monitoring or testing arrangements
- H04W24/04—Arrangements for maintaining operational condition
-
- 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)
- Signal Processing (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention discloses a kind of method for diagnosing faults of wireless sensor network, this method comprises the following steps:S1, observing matrix of the structure based on compressed sensing technology;Observing matrix describes the sequence variation value of the corresponding network event of network failure;S2, using observing matrix build from the network event of network node to the mapping relations of the network failure;S3, the node state sequence for gathering wireless sensor network;Node state sequence is the number that the network event occurs;S4, according to mapping relations, solution corresponds to the corresponding network failure of the network event.The method of the present invention is by building observing matrix, utilize compressed sensing technology and Ao Kanmu razor principles, network failure and network event are modeled, it fast can accurately find network failure, realize and the all-round failure of wireless sensor network is diagnosed, the deep domain knowledge of staff is needed not rely at the same time so that the more convenient and easy implementation of the fault diagnosis of wireless sensor network.
Description
Technical field
The present invention relates to wireless Ad Hoc sensor network field, is more particularly to a kind of failure of wireless sensor network
Diagnostic method.
Background technology
With fast-developing and increasingly mature, the wireless sensing of the communication technology, sensor technology and embedding assembly technology
Device network is widely used in many and states such as environmental monitoring, coal mining accident detection, desperate situation navigation, magnitude of traffic flow monitoring and counts
The people's livelihood has in the scene of significant associations.Due to the limitation of current scientific and technological level and manufacturing process, sensor node (referred to as saves
Point) software be usually unable to ideally agree with hardware and work, generally with error-prone feature.Simultaneously as section
Carried out data transmission between point using wireless signal, the factor such as multipath, interference further increases the shakiness of wireless sensor network
It is qualitative.
In order to strengthen the availability of wireless sensor network, while the reliability of wireless sensor network is lifted, many is ground
The research that mechanism all expands wireless sensor network fault diagnostic techniques is studied carefully, to examine and position the wrong and event in network
Barrier.Traditional wireless sensor network fault diagnostic techniques can be divided into two classes.The first kind is Software correction technology.Typical method
Be node procedure source code layer building be similar to GDB debugging acid, performed by breakpoint, variable observation, heap stack addressing
Deng interface into line code error correction, wherein, GDB is the program debugging tool that GNU increases income under the powerful UNIX that tissue is issued,
UNIX is a kind of title of widely used commercial operating systems.This kind of method may determine that the logic error of program, but not
Can identify communication link be obstructed, the failure in the wireless sensor network such as nodal function disorder.Second class technology is to pass through collection
Relevant information in wireless sensor network carries out profound data analysis, diagnoses wireless sensor network fault.This kind of side
Method can be good at identifying the error condition of wireless sensor network, but usually require deep domain knowledge, and one to nothing
The unfamiliar personnel of line sensor network or a network error not occurred are likely to cause this kind of method to fail.
As the above analysis, traditional wireless sensor network fault diagnostic techniques is largely dependent upon design
Personnel are for the domain knowledge of wireless sensor network and the experience of actual motion so that current diagnostic tool has significant
Limitation.In fact, since wireless sensor network is from the characteristic such as group interconnection, remote deployment, environment complexity, extensive, manage
It is difficult that failure cause, influence being likely to occur on wireless sensor network inside etc. has comprehensive understanding to manage maintenance personnel, especially
Occur from the failure during multi-point interaction and mistake is more difficult to detect.With wireless sensor network continuous development and
Application type is continuously increased, and the scalability of traditional wireless sensor network fault diagnostic techniques is faced with choosing for sternness
War.
The content of the invention
(1) technical problems to be solved
The technical problem to be solved in the present invention is in the fault diagnosis for how overcoming wireless sensor network to domain knowledge
Dependence, realize comprehensively and accurately detect wireless sensor network in failure.
(2) technical solution
In order to solve the above technical problem, the present invention provides a kind of method for diagnosing faults of wireless sensor network, its
It is characterized in that, the described method comprises the following steps:
S1, observing matrix of the structure based on compressed sensing technology;The observing matrix describes the corresponding net of network failure
The sequence variation value of network event (the sequence variation value is the difference of this moment of corresponding states and last moment);
S2, based on Ao Kanmu razors principle and compressed sensing technology, using observing matrix structure from network node
The network event to the network failure mapping relations;
S3, the node state sequence for gathering wireless sensor network;The node state sequence is sent out for the network event
Raw number;
S4, according to the mapping relations, solve the network failure corresponding to the network event.
Preferably, in the step S4 network failure is solved using minimization L1 norms.
Preferably, each of the observing matrix is classified as corresponding to a kind of sequence of each network event of the network failure
Changing value.
Preferably, if the network failure and the network event are not related, respective column pair in the observing matrix
The sequence variation value that should be gone is 0.
Preferably, the mapping relations are:
Y=Φ x
Wherein, y represents the node state sequence of collection, and Φ represents the observing matrix, and x represents the network event
Barrier.
(3) beneficial effect
The present invention provides a kind of method for diagnosing faults of wireless sensor network, method of the invention is observed by building
Matrix, using compressed sensing technology and Ao Kanmu razor principles, network failure and network event are modeled, can be fast accurate
Find network failure, realize to the diagnosis of the all-round failure of wireless sensor network, while need not rely on staff's
Deep domain knowledge so that the more convenient and easy implementation of the fault diagnosis of wireless sensor network.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
There is attached drawing needed in technology description to be briefly described, it should be apparent that, drawings in the following description are only this
Some embodiments of invention, for those of ordinary skill in the art, without creative efforts, can be with
Other attached drawings are obtained according to these attached drawings.
Fig. 1 is a kind of method for diagnosing faults flow chart of wireless sensor network of the present invention;
Fig. 2 is the schematic device for the method for diagnosing faults that a kind of wireless sensor network is completed in the present invention;
Fig. 3 is indoor test bed schematic diagram in the present invention.
Embodiment
The present invention is described in further detail with reference to the accompanying drawings and examples.Following embodiments are used to illustrate this hair
It is bright, but cannot be used for limiting the scope of the invention.
The invention discloses a kind of method for diagnosing faults of wireless sensor network, as shown in Figure 1, the described method includes with
Lower step:
S1, observing matrix of the structure based on compressed sensing technology;The observing matrix describes the corresponding net of network failure
The sequence variation value of network event;
S2, the mapping built using the observing matrix from the network event of network node to the network failure are closed
System;The each of the observing matrix is classified as corresponding to a kind of sequence variation value of each network event of described network failure;If institute
State network failure and the network event is not related, then respective column corresponds to the sequence variation value of row in the observing matrix
For 0;The mapping relations are:
Y=Φ x
Wherein, y represents the node state sequence of collection, and Φ represents the observing matrix, and x represents the network event
Barrier;
S3, the node state sequence for gathering wireless sensor network;The node state sequence is sent out for the network event
Raw number;
S4, according to the mapping relations, solution corresponds to the corresponding network failure of the network event;Wherein using minimum
Change L1 norms and solve the network failure.
Method proposed by the present invention is based on compressed sensing (Compressive Sensing, CS).In simple terms, compression sense
Know that technology provides a kind of acquisition method that can be widely applied to compressible signal.The invention borrow compressed sensing model and
Optimization method, network event, state and fault rootstock are modeled, then further according to Ao Kanmu razor principles, for observation
To Network Abnormal find most possible fault rootstock.
CS theories are pointed out, as long as signal is compressible or is sparse in some transform domain, then can use one
The high dimensional signal for converting gained is projected on a lower dimensional space with the conversion incoherent observing matrix of base, then passes through solution
One optimization problem reconstructs original signal from these a small amount of projections with high probability.It is not direct measurement letter in CS models
Number f in itself, but signal f is projected on observing matrix and obtains observation vector y.Represented with matrix:Y=Φ f, in formula y be M ×
1 observation vector, Φ are M × N (M<<N observing matrix)., can according to the sparse decomposition algorithm in signal sparse resolution theory
F is drawn in the hope of solution.That is, the essence of CS theory reconstructions is exactly the bar in known observation vector y and observing matrix Φ
Under part, how signal f is quickly and accurately reconstructed.
Embodiment:
Among diagnosis application, y is the node state sequence that can be observed in base station, by the journey of intra-node
Sequence is automatically generated according to the interval of setting.Under application environment, interval is set as 10 minutes once.The sequential recording 13
The number that network event occurs, including diversionary toute, channel block, queue overflow, power supply deficiency etc. are abnormal.F is then in network
43 kinds of event of failure (network failure) being likely to occur, including link failure, node damage, network separation etc., the openness of f can
To be ensured by Ao Kanmu razors principle.Observing matrix Φ is artificially demarcated using expertise, for by network failure and
Network event is associated.For example, when queue overflow failure occurs in node, the packet loss of queue record can rapidly increase
It is long;When occurring route circulatory troubles in network, also can substantially it increase because the bag number that packet sequence is identical and loses.Observation
The effect of matrix is exactly to depict what these correspondences quantified, and the data of each row are represented when a certain network failure thing
When part occurs, (if network event and network failure are not in contact with itself, which is the sequence variation value of network event
0).Ao Kanmu razors principle often network management and diagnosis in used, that is to say, that as far as possible with minimum network failure come
Current network condition is explained, it can be said that f possesses openness speciality.Have this model, once obtained in base station
Obtain the state value of some nodes, it is possible to most possible network failure is drawn using trained observing matrix, wherein, solve
Process carries out rapid Optimum using minimization L1 Norm Methods.
During solving f, by signal openness, simple to be interpreted as non-zero element number in signal less.The signal is
For a vector x ∈ RN, use ΣsRepresent s- sparse vector set, i.e.,
Σs:={ x ∈ RN:| | x | | 0≤s },
Here | | x | |0Represent the non-zero element number in x.It is so-called to signal x0∈RNCoding, that is, refer to the matrix with one n × N
Φ and x0∈RNCarry out product, then obtain y=Φ x0.Herein, y ∈ RnAs observed on x0Information.So-called solution
Code is just attempt to by y reverses x0, that is, one is found from RnTo RNMapping, which is denoted as Δ.Represent anti-with Δ (y)
Seek result.If in general, n<N, then have numerous x ∈ RNMeet y=Φ x.Therefore, only by the openness spy of signal
Sign, is possible to the original signal x of reverse0。
So, give a coding, decode to (Φ, Δ), be concerned about its performance, i.e.,:||x0-Δ(Φx0) | | X, X is one herein
Given norm.In this application, it is 1 norm to select X.Work as x0In non-zero element number it is less when, a kind of more natural solution
Code Δ0(y) be following planning problem solution, solving-optimizing process can then use " minimization L1 Norm Methods ".
P0:min x∈RN| | x | | 0, s.t. Φ x=y.
Generally speaking, this method is divided into learning process and on-line checking process under two steps, including line:
1) learning process under line.In this process, the training data in network is truly disposed is have collected, is recorded in node
The data such as various state-events, such as diversionary toute, channel block, queue overflow, power supply deficiency.These numbers are received in base station
According to afterwards, using expertise and deployment experience, can identify causes the root event of failure of each node state to combine,
Observing matrix Φ namely in CS models.
2) on-line checking process.There is observing matrix, once obtaining the state value sequence of node in base station, utilize CS models
With Ao Kanmu razor principles, the network failure combination for most likely resulting in these state values can be found out, solving-optimizing process makes
With minimization L1 Norm Methods.
Realize the algorithm on TinyOS2.1, the load of ROM is 10.6KB, compared to the 48KB of TelosB nodes
ROM Space, the algorithm are feasible.Two kinds of prediction index correctly (the faulty link number of success prediction and actually deposit by deduction rate
Faulty link number ratio) and error inference rate (the faulty link number of error prediction and the failure chain of all predictions
Way purpose ratio) it is used for weighing the performance of this method.
The actual deployment figure of device that Fig. 2 is utilized for the method for the present invention, in figure, white point is sensor node;For
The reliability and accuracy of abundant verification algorithm, have done a series of experiments, as shown in figure 3, can be same based on indoor test bed
When support 100 nodes to carry out parallel programmings, network diameter can reach 8 jumps.It is artificial former toward some networks are filled with network
Barrier, such as node link damage, Routing Loop, network separation etc., have mainly been investigated in heterogeneous networks scale and density (network density
Be defined as the neighbor node number that per node on average possesses) under the algorithm performance.In order to obtain the true representation of network, profit
With USB data line command adapted thereto is sent from base station toward different nodes.By increasing or reducing node, three kinds of network rule have been investigated
Mould, is 20,40 and 60 nodes respectively.Experiment shows that the correct deduction rate of the technology has exceeded 85.3%, and error inference rate is not
More than 9.6%.
Embodiment of above is merely to illustrate the present invention, rather than limitation of the present invention.Although with reference to embodiment to this hair
It is bright to be described in detail, it will be understood by those of ordinary skill in the art that, to technical scheme carry out it is various combination,
Modification or equivalent substitution, without departure from the spirit and scope of technical solution of the present invention, the right that should all cover in the present invention is wanted
Ask among scope.
Claims (4)
1. a kind of method for diagnosing faults of wireless sensor network, it is characterised in that the described method comprises the following steps:
S1, observing matrix of the structure based on compressed sensing technology;The observing matrix describes the corresponding network thing of network failure
The sequence variation value of part;
S2, based on Ao Kanmu razors principle and compressed sensing technology, the institute using observing matrix structure from network node
State network event is to the mapping relations of the network failure, the mapping relations:
Y=Φ x
Wherein, y represents the node state sequence of collection, and Φ represents the observing matrix, and x represents the network failure;
S3, the node state sequence for gathering wireless sensor network;The node state sequence occurs for the network event
Number;
S4, according to the mapping relations, solve the network failure corresponding to the network event.
2. according to the method described in claim 1, it is characterized in that, using described in the solution of minimization L1 norms in the step S4
Network failure.
3. according to the method described in claim 2, it is characterized in that, each of the observing matrix is classified as corresponding to described in one kind
The sequence variation value of each network event of network failure.
4. if according to the method described in claim 3, it is characterized in that, the network failure is not closed with the network event
System, then the sequence variation value that respective column corresponds to row in the observing matrix is 0.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510109164.2A CN104717689B (en) | 2015-03-12 | 2015-03-12 | A kind of method for diagnosing faults of wireless sensor network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510109164.2A CN104717689B (en) | 2015-03-12 | 2015-03-12 | A kind of method for diagnosing faults of wireless sensor network |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104717689A CN104717689A (en) | 2015-06-17 |
CN104717689B true CN104717689B (en) | 2018-05-11 |
Family
ID=53416543
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510109164.2A Active CN104717689B (en) | 2015-03-12 | 2015-03-12 | A kind of method for diagnosing faults of wireless sensor network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104717689B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106959684B (en) * | 2017-03-29 | 2019-02-19 | 清华大学 | Method for diagnosing faults and device based on compressed sensing and joint norm optimization |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101237357A (en) * | 2008-02-04 | 2008-08-06 | 华中科技大学 | Online failure detection method for industrial wireless sensor network |
WO2008137127A1 (en) * | 2007-05-03 | 2008-11-13 | Siemens Building Technologies, Inc. | Diagnostic and trouble-shooting methods in a wireless control and sensor network |
CN101867486A (en) * | 2010-06-08 | 2010-10-20 | 江苏大学 | Wireless sensor network fault diagnosis method |
CN104168599A (en) * | 2014-07-07 | 2014-11-26 | 广东工业大学 | Wireless sensor network fault diagnosis method based on time weight K-neighbor algorithm |
-
2015
- 2015-03-12 CN CN201510109164.2A patent/CN104717689B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2008137127A1 (en) * | 2007-05-03 | 2008-11-13 | Siemens Building Technologies, Inc. | Diagnostic and trouble-shooting methods in a wireless control and sensor network |
CN101237357A (en) * | 2008-02-04 | 2008-08-06 | 华中科技大学 | Online failure detection method for industrial wireless sensor network |
CN101867486A (en) * | 2010-06-08 | 2010-10-20 | 江苏大学 | Wireless sensor network fault diagnosis method |
CN104168599A (en) * | 2014-07-07 | 2014-11-26 | 广东工业大学 | Wireless sensor network fault diagnosis method based on time weight K-neighbor algorithm |
Non-Patent Citations (3)
Title |
---|
Agnostic Diagnosis: Discovering Silent Failures in Wireless Sensor Networks;Xin Miao,et.al.;《IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS》;20131111;第12卷(第12期);第6067-6075页 * |
Link Scanner: Faulty Link Detection for Wireless Sensor Networks;Qiang Ma,et.al.;《2013 Proceedings IEEE INFOCOM》;20130419;第2688-2696页 * |
Self-Diagnosis for Detecting System Failures in Large-Scale Wireless Sensor Networks;Qiang Ma,et.al.;《IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS》;20140708;第13卷(第10期);第5535-5545页 * |
Also Published As
Publication number | Publication date |
---|---|
CN104717689A (en) | 2015-06-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11334407B2 (en) | Abnormality detection system, abnormality detection method, abnormality detection program, and method for generating learned model | |
KR102483025B1 (en) | Operational maintenance systems and methods | |
CN111504676B (en) | Equipment fault diagnosis method, device and system based on multi-source monitoring data fusion | |
CN108764601B (en) | Structural health monitoring abnormal data diagnosis method based on computer vision and deep learning technology | |
JP5278310B2 (en) | Diagnostic system | |
CN105279196B (en) | The generation method and device of test script | |
CN109204389A (en) | A kind of subway equipment fault diagnosis and self-healing method, system | |
CN109471698B (en) | System and method for detecting abnormal behavior of virtual machine in cloud environment | |
CN114970715A (en) | Variable working condition fault diagnosis method and system under small sample and unbalanced data constraint | |
WO2023197617A1 (en) | Method for detecting and diagnosing production abnormality of industrial system on basis of multi-dimensional sensing data | |
CN110515781A (en) | A kind of complication system status monitoring and method for diagnosing faults | |
CN108415810A (en) | Hard disk state monitoring method and device | |
RU2431175C1 (en) | System for integral monitoring of operation of aircraft onboard equipment | |
CN115510950A (en) | Aircraft telemetry data anomaly detection method and system based on time convolution network | |
CN113536658B (en) | Electromechanical equipment lightweight fault diagnosis method based on STM32 embedded processor | |
CN104717689B (en) | A kind of method for diagnosing faults of wireless sensor network | |
CN110262417A (en) | A kind of production equipment failure predication diagnostic method | |
CN108415819A (en) | Hard disk fault tracking method and device | |
CN115859090A (en) | Turnout fault diagnosis method and system based on 1DCNN-LSTM | |
JP7062505B2 (en) | Equipment management support system | |
CN113516179B (en) | Method and system for identifying water leakage performance of underground infrastructure | |
KR20220019560A (en) | Apparatus and method for monitoring network | |
Maroli et al. | A Framework for Automated Collaborative Fault Detection in Large-Scale Vehicle Networks | |
KR102335604B1 (en) | Apparatus and method for intelligent abnormal operation recognition of system using bitmap pattern representation of power-consumption | |
CN116381419B (en) | Transmission line fault processing method, device, computer equipment and storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |