CN109640335A - Wireless sensor fault diagnosis algorithm based on convolutional neural networks - Google Patents
Wireless sensor fault diagnosis algorithm based on convolutional neural networks Download PDFInfo
- Publication number
- CN109640335A CN109640335A CN201910150809.5A CN201910150809A CN109640335A CN 109640335 A CN109640335 A CN 109640335A CN 201910150809 A CN201910150809 A CN 201910150809A CN 109640335 A CN109640335 A CN 109640335A
- Authority
- CN
- China
- Prior art keywords
- layer
- data
- node
- neural networks
- wireless sensor
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/22—Traffic simulation tools or models
- H04W16/225—Traffic simulation tools or models for indoor or short range network
-
- 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
- H04W40/00—Communication routing or communication path finding
- H04W40/02—Communication route or path selection, e.g. power-based or shortest path routing
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W40/00—Communication routing or communication path finding
- H04W40/02—Communication route or path selection, e.g. power-based or shortest path routing
- H04W40/04—Communication route or path selection, e.g. power-based or shortest path routing based on wireless node resources
- H04W40/10—Communication route or path selection, e.g. power-based or shortest path routing based on wireless node resources based on available power or energy
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
Abstract
The present invention relates to a kind of wireless sensor fault diagnosis algorithm based on convolutional neural networks, steps are as follows: 1) constructing the wireless sensor network system model being made of 1 base station, 1 moving trolley and several rechargeable wireless sensor nodes；2) in base station, building is sequentially connected with the convolutional neural networks of input layer, hidden layer, full articulamentum and output layer；3) perception data of all nodes is collected using moving trolley and be transferred to base station stored, and be converted to matrix form；4) perception data of matrix form is input in convolutional neural networks to be trained and is learnt with itself, data characteristics is extracted by convolutional layer convolution kernel, pond layer compresses data characteristics, full articulamentum connects last two layers, output layer output final data classification results；5) every one kind is corresponded to corresponding sensor fault type according to data classification result and carries out fault diagnosis by convolutional neural networks, passes through output layer output node diagnostic state.
Description
Technical field
The present invention relates to technical field of wireless more particularly to a kind of wireless sensors based on convolutional neural networks
Fault diagnosis algorithm.
Background technique
Wireless sensor network is made of by way of self-organizing several wireless sensor nodes, a wireless sensing
Device node is mainly made of sensor module, CPU module, wireless communication module and 4 part of power module, wherein sensor die
Block is mainly used to perception data, and the effect of CPU module is processing, calculates data, wireless communication module guarantee sensor node with
Other sensors node is communicated, and power module carries limited energy and provides energy for sensor node.Due to wirelessly passing
The every aspect that there is sensor node the features such as small in size, easily deployment, cheap to be widely used life, with wireless
The enhancing of sensor computing capability and storage capacity, wireless sensor network be deployed to large area carry out data perception and
The detection of environment is deployed to the environment that the mankind are difficult to enter and is monitored.When wireless sensor node breaks down, meeting
Cause monitoring data abnormal.
The failure of wireless sensor network interior joint can be divided into hard fault and soft fault, and hard fault refers in sensor node
Certain hardware, which go wrong, perception data or can not carry out communication between node, and soft fault refers to due to certain originals
There is exception in the data of cause, node perceived, but node is able to carry out normal communication.Soft fault again may be used according to data exception situation
It is divided into permanent soft fault, intermittent soft fault and transient fault.Accordingly, the Node Diagnostics state in wireless sensor network
Have:
(1) normal node: referring to that the perception data of node is constantly in normal range, and perception data changes with time hair
Changing；
(2) hard fault node: node cannot communicate, therefore not receive the perception data of node；
(3) permanent soft fault node: refer to that the perception data of node deviates always normal sensing range whithin a period of time；
(4) intermittent defect node: refer to that the part perception data of a period of time interior nodes deviates normal range；
(5) transient fault node: referring in a period of time, and the several sensing data of certain of node deviate normal range, but other
Sensing data it is still normal.
Existing wireless sensor fault diagnosis algorithm can be generally divided into centralized and distributed.Centralized fault diagnosis
Algorithm is that the perception data of all the sensors node in network is transferred to base station by way of multi-hop, and base station passes through to node
The analysis of data obtains the fault diagnosis state of node.In the prior art, based on the fault diagnosis algorithm of time series according to biography
The perception data of the existing perception data prediction future time of sensor, if the true value of predicted value and sensor is very nearly the same,
Illustrate that node state is normal, otherwise failure.With the development of neural network, neural network is answered because of its higher classification accuracy
It uses in fault diagnosis, such as the inventors such as Rakesh Ranjan Swain and Pabitra Mohan khilar utilize nerve
Network carries out classification to sensing data to classify to sensor node, achievees the purpose that fault diagnosis.Although centralization
Fault diagnosis algorithm can reach higher fault diagnosis precision, but since sensor node is needed self-sensor number via multi-hop
According to base station is transferred to, then the node energy close to base station can be consumed rapidly, so that whole network be made to paralyse.And distributed event
The it is proposed of barrier diagnosis algorithm solves the problems, such as this.The distributed diagnostics algorithm that the inventors such as Jiang propose passes through comparison
The similitude of sensor itself perception data and neighbor node perception data judges the diagnostic state of sensor node, sensor
Node no longer needs perception data being transferred to base station, only need to carry out data exchange with neighbor node, reduce energy
Loss, but each node requires still to consume excessive energy with neighbor node progress information exchange, for this purpose, there are also documents to mention
Go out the fault diagnosis algorithm based on sub-clustering, wireless sensor is subjected to sub-clustering, by cluster head according to node similitude to cluster internal segment
Point carries out fault diagnosis, so that each node only need to carry out primary information exchange with cluster head, is further reduced the friendship of information
Change (since cluster head needs to receive the information of member in cluster, and need to carry out the status monitoring of cluster interior nodes, cluster head energy consumption compared with
Fastly, it is therefore desirable to which multiple sub-clustering selects suitable cluster head, improves the complexity of algorithm).Although distributed diagnostics algorithm
It alleviates network node energy and consumes too fast problem, but how high requirement of the algorithm to neighbor node be, when neighbor node failures number mistake
Gao Shi, fault diagnosis is error-prone, so distributed diagnostics algorithm can not be in the higher wireless sensor network of failure rate
Node carries out accurate fault detection.
Summary of the invention
In view of the above-mentioned deficiencies in the prior art, it is an object of the present invention to provide, a kind of design is reasonable, and structure is simple, fault diagnosis
Accuracy is high, the wireless sensor fault diagnosis algorithm based on convolutional neural networks of consuming little energy.
To achieve the above object, the invention adopts the following technical scheme:
Wireless sensor fault diagnosis algorithm based on convolutional neural networks comprising following steps:
1) wireless sensor network system model is constructed, wireless sensor network system model is by base station, moving trolley and quiet
Several rechargeable wireless sensor nodes for being only laid in monitoring region are constituted；
2) convolutional neural networks are constructed in base station, the structure of convolutional neural networks is input layer, hidden layer, the Quan Lian sequentially connected
Connect layer and output layer, wherein the convolutional layer and pond layer that hidden layer is sequentially connected by least one set are constituted；
3) perception data of all nodes in wireless sensor network is collected using moving trolley and be transferred to base station, base station stored
The perception data of all nodes, and the form for being converted to matrix is input in convolutional neural networks；
4) perception data of matrix form is input in the convolutional neural networks of building by input layer and is trained, convolution mind
The study for passing through itself through network, extracts data characteristics by the convolution kernel of convolutional layer, is carried out by pond layer to data characteristics
Compression reducing data characteristics, simplifying network query function complexity and extracting main feature, and passes through full articulamentum connection last two
Layer, final data classification result is exported by output layer；
5) convolutional neural networks are according to data classification as a result, every one kind, which is corresponded to corresponding sensor fault type, carries out failure
Diagnosis, and pass through output layer output node diagnostic state.
Preferably, collecting the perception data of all nodes in wireless sensor network using moving trolley in step 2
Steps are as follows:
(1) select anchor node: moving trolley collects the surplus of all nodes when traversing the node in wireless sensor network for the first time
Complementary energy, and it is transferred to base station, base station is selected by the analysis to dump energy and needs the node for carrying out energy supplement as anchor
Node；
(2) planning path: base station plans the movement routine of moving trolley by shortest route problem according to the anchor node selected,
A shortest path is found, reducing the energy consumption of moving trolley, shortening traveling time and slowing down data transfer delay；
(3) determine the transmission path of non-anchor: using linear programming algorithm, by comprehensively consider flow conservation, energy balance and
Link capacity carrys out the transmission path that maximization network effectiveness carrys out non-anchor node in planning network；
(4) charging and data collection: non-anchor node is before moving trolley reaches the anchor node, by self-sensor data and residue
Energy passes through the transmission path planned to anchor node, and moving trolley traverses all anchor sections via the shortest path of planning
Point charges and is collected to anchor node the sensing data and residue energy of node of the non-anchor node of all anchor node storages, is sent to
Base station.
Preferably, hidden layer described in step 2 is by the convolutional layer, pond layer, convolutional layer and the pond layer that sequentially connect
It constitutes.
Preferably, the sorting algorithm of sensor fault type has Decision Tree Algorithm, Bayes's classification in step 4)
Algorithm, artificial neural network sorting algorithm, k- nearest neighbour classification algorithm or support vector cassification algorithm.
The invention adopts the above technical scheme, mainly by based on moving trolley data collection and energy supplement, and
Fault diagnosis algorithm two parts based on convolutional neural networks are constituted, and make it in the use for guaranteeing wireless sensor network interior joint
The perception data of nodes all in network can be all transferred to base station while service life, after receiving data so as to base station, led to
It crosses based on convolutional neural networks (Convolutional Neural Network, CNN) sorting algorithm logarithm in deep learning
It according to being handled and being classified, is acted on according to the classification of convolutional neural networks, is completed by training convolutional neural networks to sensor
The classification of node, to achieve the purpose that carry out fault diagnosis to sensor node.
Detailed description of the invention
Now in conjunction with attached drawing, the present invention is further elaborated:
Fig. 1 is wireless sensor network system model of the present invention.
Specific embodiment
As shown in Figure 1, the wireless sensor fault diagnosis algorithm of the invention based on convolutional neural networks comprising following
Step:
1) wireless sensor network system model is constructed, wireless sensor network system model is by 1 base station, 1 moving trolley
And static several rechargeable wireless sensor nodes for being laid in monitoring region are constituted；
2) convolutional neural networks are constructed in base station, the structure of convolutional neural networks is input layer, hidden layer, the Quan Lian sequentially connected
Connect layer and output layer, wherein the convolutional layer and pond layer that hidden layer is sequentially connected by least one set are constituted；
3) perception data of all nodes in wireless sensor network is collected using moving trolley and be transferred to base station, base station stored
The perception data of all nodes, and the form for being converted to matrix is input in convolutional neural networks；
4) perception data of matrix form is input in the convolutional neural networks of building by input layer and is trained, convolution mind
The study for passing through itself through network, extracts data characteristics by the convolution kernel of convolutional layer, is carried out by pond layer to data characteristics
Compression reducing data characteristics, simplifying network query function complexity and extracting main feature, and passes through full articulamentum connection last two
Layer, final data classification result is exported by output layer；
5) convolutional neural networks are according to data classification as a result, every one kind, which is corresponded to corresponding sensor fault type, carries out failure
Diagnosis, and pass through output layer output node diagnostic state.
Preferably, hidden layer described in step 2 is by the convolutional layer, pond layer, convolutional layer and the pond layer that sequentially connect
It constitutes.
The invention adopts the above technical scheme, and it is all the sensors that wireless sensor network, which can be seen as a non-directed graph,
The set of node is the set on side in figure, if showing to be in communication with each other between node.In wireless sensor network,
If.It can be in communication with each other between node depending on the Euclidean distance between node.Assuming that sensor node is in monitoring region
Position coordinates are respectively, the Euclidean distance between node is.When being less than or equal to the transmission radius of sensor node,
Then have.Sensor node is every to carry out a data acquisition by the time, and stores data into the memory of itself.Assuming that each
Sensor node can store a perception data, and at the moment, the data of sensor node i storage are.Wherein, moving trolley is sensing
It is moved between device node, is mainly used for the collection of nodal information and the supplement of energy；Rechargeable wireless sensor node, it is random
Be deployed to region to be monitored, carry out the acquisition of data, and can be charged to carry out the supplement of energy；Base station is disposed
In area to be tested, base station can analyze the mass data of collection, be handled, to obtain useful information；Convolutional layer
Data characteristics is mainly extracted by convolution kernel, is most important part in convolutional neural networks；Pond layer to data characteristics into
Row compression, both can reduce data characteristics, simplifies network query function complexity, can also carry out Feature Compression to extract main spy
Sign；Full articulamentum is usually used in connecting adjacent two layers, for exporting final classification results.
Due to traditional wireless sensor network data collection be wireless sensor by self-sensor data via multiple biographies
Sensor is transferred to base station in a multi-hop fashion.Sensor node close to base station needs repeatedly to carry out the transmission of data, and being easy will
Depleted of energy is so as to cause node death；And the node close to base station is due to that will serve as the terminal of multiple nodes, to lead
Write breath congestion.Therefore, moving trolley is applied in wireless sensor network by the present invention, is asked with solving data collection and energy consumption
Topic, is walked in a network using moving trolley, is collected to the perception data of sensor node and to the insufficient node of energy
The supplement for carrying out energy, extends the service life of wireless sensor node.In order to extend wireless sensor node in wireless sensing
Service life in device network and the perception data for being collected simultaneously sensor node in network, the present invention use following technical side
Case:
Preferably, the step of collecting the perception data of all nodes in wireless sensor network using moving trolley in step 2
It is as follows:
(1) select anchor node: moving trolley collects the surplus of all nodes when traversing the node in wireless sensor network for the first time
Complementary energy, and it is transferred to base station, base station is selected by the analysis to dump energy and needs the node for carrying out energy supplement as anchor
Node；
(2) planning path: base station plans the movement routine of moving trolley by shortest route problem according to the anchor node selected,
A shortest path is found, reducing the energy consumption of moving trolley, shortening traveling time and slowing down data transfer delay；
(3) determine the transmission path of non-anchor: using linear programming algorithm, by comprehensively consider flow conservation, energy balance and
Link capacity carrys out the transmission path that maximization network effectiveness carrys out non-anchor node in planning network；
(4) charging and data collection: non-anchor node is before moving trolley reaches the anchor node, by self-sensor data and residue
Energy passes through the transmission path planned to anchor node, and moving trolley traverses all anchor sections via the shortest path of planning
Point, charge and collect to anchor node the non-anchor node of all anchor nodes storage sensing data and and residue energy of node, hair
It is sent to base station.
Preferably, the sorting algorithm of sensor fault type has Decision Tree Algorithm, Bayes's classification in step 4)
Algorithm, artificial neural network sorting algorithm, k- nearest neighbour classification algorithm or support vector cassification algorithm.
Above-mentioned algorithm can will be all in network while guaranteeing the service life of wireless sensor network interior joint
The perception data of sensor node is all transferred to base station, so as to next to data analyzed come to node carry out failure examine
It is disconnected.Sensing data in wireless sensor network is transferred to base station via moving trolley, and base station receives data, by data
It is handled and is classified, obtain sensor node diagnostic state.Different failures has different influences to perception data, therefore right
Node progress fault diagnosis in wireless sensor network, which can be converted to, classifies to the perception data of sensor.Therefore,
The present invention is used based on convolutional neural networks (Convolutional Neural Network, the CNN) classification in deep learning
Algorithm to carry out fault diagnosis to the node in wireless sensor network.
Convolutional neural networks are a kind of feedforward neural networks, and architecture has convolutional layer, pond layer and full articulamentum structure
At general basic structure is made of input layer, convolutional layer, pond layer, full articulamentum, output layer, wherein convolutional layer and pond layer
Hidden layer is constituted, it is most important part, pond in convolutional neural networks that convolutional layer, which mainly extracts data characteristics by convolution kernel,
Change layer to compress data characteristics, both can reduce data characteristics, simplifies network query function complexity, feature pressure can also be carried out
To extract main feature, full articulamentum is usually used in connecting last two layers for contracting, and output layer is used to export final classification results.This
Convolutional layer and pond layer can repeatedly occur according to demand in the hidden layer of invention, preferably, convolutional neural networks of the invention
Structure is input layer, convolutional layer, pond layer, convolutional layer, pond layer, full articulamentum and output layer；The present invention is mainly by based on shifting
The data collection of dynamic trolley and energy supplement, and fault diagnosis algorithm two parts based on convolutional neural networks are constituted, and make it
It can be by the perception of sensor node all in network while guaranteeing the service life of wireless sensor network interior joint
Data are all transferred to base station, after receiving data so as to base station, by based on the convolutional neural networks in deep learning
(Convolutional Neural Network, CNN) sorting algorithm is handled and is classified to data, according to convolutional Neural net
The classification of network acts on, and the classification to sensor node is completed by training convolutional neural networks, to reach to sensor node
Carry out the purpose of fault diagnosis.
Emulation experiment
Anaconda is the python version of an open source, wherein containing a large amount of packet of convolutional neural networks needs, is convenient for
The progress of experiment.Keras is a high-rise neural network API, is write by pure python into simultaneously base Tensorflow, theano
And the rear end CNTK, support CNN and RNN, and Keras provides consistent and succinct API, can greatly reduce workload, and
Easily extension, cooperates with python, without individual model configuration file type, provides convenience for extension.Keras is used herein
Computational frame carries out the training built with model of structure.It is the data really about environment such as temperature used in experiment
Collection.During the experiment, a part is therefrom selected as training set, and a part is used as test set, wherein data set has marked
The time of day of good sensor node.The experimental results showed that the accuracy that the present invention classifies is high, the survey of convolutional neural networks model
Rapid and convenient is tried, is able to extend service life of the wireless sensor node in wireless sensor network, and being capable of Precise collection
The perception data of sensor node in network.
Above description should not have any restriction to protection scope of the present invention.
Claims (4)
1. the wireless sensor fault diagnosis algorithm based on convolutional neural networks, it is characterised in that: itself the following steps are included:
1) wireless sensor network system model is constructed, wireless sensor network system model is by base station, moving trolley and quiet
Several rechargeable wireless sensor nodes for being only laid in monitoring region are constituted；
2) convolutional neural networks are constructed in base station, the structure of convolutional neural networks is input layer, hidden layer, the Quan Lian sequentially connected
Connect layer and output layer, wherein the convolutional layer and pond layer that hidden layer is sequentially connected by least one set are constituted；
3) perception data of all nodes in wireless sensor network is collected using moving trolley and be transferred to base station, base station stored
The perception data of all nodes, and the form for being converted to matrix is input in convolutional neural networks；
4) perception data of matrix form is input in the convolutional neural networks of building by input layer and is trained, convolution mind
The study for passing through itself through network, extracts data characteristics by the convolution kernel of convolutional layer, is carried out by pond layer to data characteristics
Compression reducing data characteristics, simplifying network query function complexity and extracting main feature, and passes through full articulamentum connection last two
Layer, final data classification result is exported by output layer；
5) convolutional neural networks are according to data classification as a result, every one kind, which is corresponded to corresponding sensor fault type, carries out failure
Diagnosis, and pass through output layer output node diagnostic state.
2. the wireless sensor fault diagnosis algorithm according to claim 1 based on convolutional neural networks, it is characterised in that:
The step of collecting the perception data of all nodes in wireless sensor network using moving trolley in step 2 is as follows:
(1) select anchor node: moving trolley collects the surplus of all nodes when traversing the node in wireless sensor network for the first time
Complementary energy, and it is transferred to base station, base station is selected by the analysis to dump energy and needs the node for carrying out energy supplement as anchor
Node；
(2) planning path: base station plans the movement routine of moving trolley by shortest route problem according to the anchor node selected,
A shortest path is found, reducing the energy consumption of moving trolley, shortening traveling time and slowing down data transfer delay；
(3) determine the transmission path of non-anchor: using linear programming algorithm, by comprehensively consider flow conservation, energy balance and
Link capacity carrys out the transmission path that maximization network effectiveness carrys out non-anchor node in planning network；
(4) charging and data collection: non-anchor node is before moving trolley reaches the anchor node, by self-sensor data and residue
Energy passes through the transmission path planned to anchor node, and moving trolley traverses all anchor sections via the shortest path of planning
Point charges and is collected to anchor node the sensing data and residue energy of node of the non-anchor node of all anchor node storages, is sent to
Base station.
3. the wireless sensor fault diagnosis algorithm according to claim 1 based on convolutional neural networks, it is characterised in that:
Hidden layer described in step 2 is made of convolutional layer, pond layer, convolutional layer and the pond layer sequentially connected.
4. the wireless sensor fault diagnosis algorithm according to claim 1 based on convolutional neural networks, it is characterised in that:
The sorting algorithm of sensor fault type has Decision Tree Algorithm, Bayesian Classification Arithmetic, artificial neural network in step 4)
Sorting algorithm, k- nearest neighbour classification algorithm or support vector cassification algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910150809.5A CN109640335B (en) | 2019-02-28 | 2019-02-28 | Wireless sensor fault diagnosis method based on convolutional neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910150809.5A CN109640335B (en) | 2019-02-28 | 2019-02-28 | Wireless sensor fault diagnosis method based on convolutional neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109640335A true CN109640335A (en) | 2019-04-16 |
| CN109640335B CN109640335B (en) | 2022-02-08 |
Family
ID=66066151
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910150809.5A Active CN109640335B (en) | 2019-02-28 | 2019-02-28 | Wireless sensor fault diagnosis method based on convolutional neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109640335B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110213788A (en) * | 2019-06-15 | 2019-09-06 | 福州大学 | WSN abnormality detection and kind identification method based on data flow space-time characteristic |
| CN111356108A (en) * | 2020-03-06 | 2020-06-30 | 山东交通学院 | Neural network-based underwater wireless sensor network anomaly diagnosis method |
| US20220353952A1 (en) * | 2018-08-02 | 2022-11-03 | Huawei Technologies Co., Ltd. | Network Switching Method, Electronic Device, and System On Chip |
| CN115460647A (en) * | 2022-10-21 | 2022-12-09 | 北京中电飞华通信有限公司 | Internet of things fault positioning method and system based on eSIM card and 5G base station |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105841961A (en) * | 2016-03-29 | 2016-08-10 | 中国石油大学(华东) | Bearing fault diagnosis method based on Morlet wavelet transformation and convolutional neural network |
| CN106168539A (en) * | 2016-10-21 | 2016-11-30 | 常州信息职业技术学院 | Fault Diagnosis of Roller Bearings based on convolutional neural networks and system |
| CN106226074A (en) * | 2016-09-22 | 2016-12-14 | 华中科技大学 | Based on convolutional neural networks and the rotary machinery fault diagnosis method of small echo gray-scale map |
| CN107277840A (en) * | 2017-06-09 | 2017-10-20 | 浙江工业大学 | A kind of rechargeable wireless sensor network data acquisition method |
| CN206658064U (en) * | 2017-03-27 | 2017-11-21 | 河南师范大学 | A kind of wireless charging system of wireless sensor network node |
| CN108419255A (en) * | 2018-02-02 | 2018-08-17 | 浙江工业大学 | A kind of wireless sensor network mobile charging and method of data capture |
| CN108647716A (en) * | 2018-05-09 | 2018-10-12 | 北京理工大学 | A kind of diagnosing failure of photovoltaic array method based on composite information |
-
2019
- 2019-02-28 CN CN201910150809.5A patent/CN109640335B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105841961A (en) * | 2016-03-29 | 2016-08-10 | 中国石油大学(华东) | Bearing fault diagnosis method based on Morlet wavelet transformation and convolutional neural network |
| CN106226074A (en) * | 2016-09-22 | 2016-12-14 | 华中科技大学 | Based on convolutional neural networks and the rotary machinery fault diagnosis method of small echo gray-scale map |
| CN106168539A (en) * | 2016-10-21 | 2016-11-30 | 常州信息职业技术学院 | Fault Diagnosis of Roller Bearings based on convolutional neural networks and system |
| CN206658064U (en) * | 2017-03-27 | 2017-11-21 | 河南师范大学 | A kind of wireless charging system of wireless sensor network node |
| CN107277840A (en) * | 2017-06-09 | 2017-10-20 | 浙江工业大学 | A kind of rechargeable wireless sensor network data acquisition method |
| CN108419255A (en) * | 2018-02-02 | 2018-08-17 | 浙江工业大学 | A kind of wireless sensor network mobile charging and method of data capture |
| CN108647716A (en) * | 2018-05-09 | 2018-10-12 | 北京理工大学 | A kind of diagnosing failure of photovoltaic array method based on composite information |
Non-Patent Citations (1)
| Title |
|---|
| 马梦莹等: "基于邻居节点预状态的无线传感器网络故障诊断算法", 《计算机应用》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20220353952A1 (en) * | 2018-08-02 | 2022-11-03 | Huawei Technologies Co., Ltd. | Network Switching Method, Electronic Device, and System On Chip |
| US11910489B2 (en) * | 2018-08-02 | 2024-02-20 | Huawei Technologies Co., Ltd. | Network switching method, electronic device, and system on chip |
| CN110213788A (en) * | 2019-06-15 | 2019-09-06 | 福州大学 | WSN abnormality detection and kind identification method based on data flow space-time characteristic |
| CN110213788B (en) * | 2019-06-15 | 2021-07-13 | 福州大学 | WSN (Wireless sensor network) anomaly detection and type identification method based on data flow space-time characteristics |
| CN111356108A (en) * | 2020-03-06 | 2020-06-30 | 山东交通学院 | Neural network-based underwater wireless sensor network anomaly diagnosis method |
| CN111356108B (en) * | 2020-03-06 | 2022-04-19 | 山东交通学院 | Neural network-based underwater wireless sensor network anomaly diagnosis method |
| CN115460647A (en) * | 2022-10-21 | 2022-12-09 | 北京中电飞华通信有限公司 | Internet of things fault positioning method and system based on eSIM card and 5G base station |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109640335B (en) | 2022-02-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109640335A (en) | Wireless sensor fault diagnosis algorithm based on convolutional neural networks | |
| Messaoud et al. | A survey on machine learning in Internet of Things: Algorithms, strategies, and applications | |
| Kumar et al. | Machine learning algorithms for wireless sensor networks: A survey | |
| Guo et al. | Multi-source temporal data aggregation in wireless sensor networks | |
| CN106095099A (en) | A kind of user behavior motion detection recognition methods | |
| CN101415256A (en) | Method of diagnosing wireless sensor network fault based on artificial immunity system | |
| CN106056269A (en) | NanoSat satellite house-keeping health management system based on Bayes network model | |
| Rida et al. | EK-means: A new clustering approach for datasets classification in sensor networks | |
| CN101516099A (en) | Test method for sensor network anomaly | |
| CN109688602A (en) | Wireless sensor fault detection algorithm based on node pre-state | |
| CN108289285B (en) | Method for recovering and reconstructing lost data of ocean wireless sensor network | |
| Harb et al. | En-route data filtering technique for maximizing wireless sensor network lifetime | |
| CN107424416A (en) | Road section traffic volume flow monitoring system based on unmanned plane | |
| CN114265913A (en) | Space-time prediction algorithm based on federal learning on industrial Internet of things edge equipment | |
| Lata et al. | Machine learning based energy efficient wireless sensor network | |
| CN109638826A (en) | A kind of topology identification method and system based on CNN | |
| Laidi et al. | On predicting sensor readings with sequence modeling and reinforcement learning for energy-efficient iot applications | |
| Bhatnagar et al. | Machine learning techniques to reduce error in the internet of things | |
| CN102118884B (en) | Data transmission method for opportunistic mobile sensor network based on closeness centrality | |
| Yang et al. | SA1D-CNN: A separable and attention based lightweight sensor fault diagnosis method for solar insecticidal lamp Internet of Things | |
| Anagnostopoulos et al. | Path prediction through data mining | |
| Nguyen et al. | Extracting typical users' moving patterns using deep learning | |
| Ali et al. | Energy optimization of wireless sensor network using neuro-fuzzy algorithms | |
| Bhowmik et al. | An improved PSO based fuzzy clustering algorithm in WSNs | |
| CN114781598A (en) | Fault prediction method based on hierarchical neural network distributed training |
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 |