CN108289285A - A kind of ocean wireless sensor network is lost data and is restored and reconstructing method - Google Patents
A kind of ocean wireless sensor network is lost data and is restored and reconstructing method Download PDFInfo
- Publication number
- CN108289285A CN108289285A CN201810030395.8A CN201810030395A CN108289285A CN 108289285 A CN108289285 A CN 108289285A CN 201810030395 A CN201810030395 A CN 201810030395A CN 108289285 A CN108289285 A CN 108289285A
- Authority
- CN
- China
- Prior art keywords
- data
- node
- cluster
- ocean
- 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
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/02—Services making use of location information
- H04W4/023—Services making use of location information using mutual or relative location information between multiple location based services [LBS] targets or of distance thresholds
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W28/00—Network traffic management; Network resource management
- H04W28/02—Traffic management, e.g. flow control or congestion control
- H04W28/04—Error control
-
- 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
- H04W40/20—Communication route or path selection, e.g. power-based or shortest path routing based on geographic position or location
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Abstract
The invention discloses a kind of ocean wireless sensor networks to lose data recovery and reconstructing method, including：Node clustering cluster and loss data are restored；The wherein described node clustering cluster includes：In sea area to be monitored, arrangement meets the node of ocean network communication of wireless sensor requirement and real-time detection requirement first, completes the topology constructing of ocean wireless sensor network, and effectively cluster cluster is carried out to deployment node according to K means algorithms are improved；The wherein described loss data are restored：When node data is lost, carry out the temporal correlation of mining data lost clusters interior joint data using the RBF neural of PSO algorithm optimizations, and then restore to lose data value according to history wheel in lost clusters and when front-wheel data.The present invention can adapt to the high dynamic of ocean Wireless Sensor Network Topology, and can reduce the transmission energy consumption of data between node, to achieve the purpose that extend the network life period.
Description
Technical field
The present invention relates to ocean wireless sense network data prediction and transmission technologys, and in particular to a kind of ocean wireless sensor
Network (OWSNs) loses data recovery and reconstructing method.
Background technology
The real-time acquisition of oceanographic data is 21 century full appreciation ocean, exploitation marine resources and the premise for protecting ocean.
Wireless sensor network (WSN, wireless sensor network) is with its low-power consumption, low cost, distribution and self-organizing
Feature is widely used in environmental monitoring.In marine eco-environment monitoring, shed in a large amount of of target marine site
Sensor node can self-organizing be a preferable monitoring network of adaptability in a manner of wireless multi-hop rapidly, so as to reach
The requirement of acquisition oceanographic data in real time.
In recent years, ocean wireless sense network has been widely used in marine environmental monitoring, such as oil pollution at sea diffusion prison
The monitoring in real time of survey, ocean water quality and marine information acquisition etc..Wireless sensor network main employing wireless in ocean senses network technology
To realize to the real-time monitoring of the marine eco-environment and oceanographic data collecting work.
However, (hardware fault, data packet collisions, signal decaying, energy deficiency, time irreversibility, evil due to various reasons
Meaning attack) data in the wireless sensor network of ocean are easy to occur to lose on a large scale.This just need to lose data into
Row restores to obtain complete environmental data collection.In data acquisition, the recovery to losing data is a basic operation.
In the wireless sense network of ocean, the major issue to be solved is exactly the lifetime of the extend as far as possible network,
And at the same time ensureing that each nodal information can be transferred to aggregation node, to ensure that network monitor information can be transmitted timely and effectively
It is used for relevant departments and user to control centre.Due to the extensive dense deployment of ocean wireless sensing net node, therefore network
The collected data spatio-temporal redundancies of institute are larger.If the whole oceanographic datas acquired are all sent to convergence terminal by node, no
A large amount of energy consumption is only generated, and the congestion of data transmission channel can be caused.Restored using data and reconfiguration technique reduces OWSNs
Middle volume of transmitted data becomes the effective means for reducing network energy consumption.
Existing data are restored and prediction technique is mainly based upon temporal correlation model and probabilistic model progress data are extensive
It is multiple, but in the case where monitoring data change over time larger, data recovery errors are larger for both models, and consider
The shadow of characteristic and wave the shadowing effect data transmission between node of the real-time dynamic change of ocean Wireless Sensor Network Topology
It rings, therefore the temporal correlation of marine environment data and spatial coherence is combined and propose that a kind of new ocean wirelessly passes
The recovery of sensor network loss data is necessary with reconstructing method.
Invention content
The object of the present invention is to provide a kind of ocean wireless sensor networks to lose data recovery and reconstructing method, solves sea
Foreign wireless sensing net node loses data and restores problem, and this method adapts to the high dynamic of ocean Wireless Sensor Network Topology,
And the transmission energy consumption that data between node can be reduced, to achieve the purpose that extend the network life period.
In order to achieve the goal above, the present invention is achieved by the following technical solutions：
A kind of ocean wireless sensor network is lost data and is restored and reconstructing method, its main feature is that, including：Node clustering at
Cluster and loss data are restored；
The wherein described node clustering cluster includes：In sea area to be monitored, arrangement meets ocean wireless sensor network first
The node of network communicating requirement and real-time detection requirement completes the topology constructing of ocean wireless sensor network, according to improvement K-
Means algorithms carry out effectively cluster cluster to deployment node；
The wherein described loss data are restored：When node data is lost, the RBF god of PSO algorithm optimizations is used
Carry out the temporal correlation of mining data lost clusters interior joint data through network, and then according to history wheel in lost clusters and works as front-wheel number
According to come restore lose data value.
The ocean network communication of wireless sensor requires to be sensor node self-organizing in the marine monitoring region
It forms wireless sensor network and completes topology structure initialization, and the network covers monitored sea area comprehensively.
The node clustering cluster detailed process is：
The selection of cluster centre：
Consider sensor node data collection to be clusteredFor corresponding index collection,
dij=dist (xi,xj) indicate node xiWith xjThe distance between, for any sensor node xi, define local density ρiWith
Distance δiTwo amounts portray cluster centre；
X is calculated by Gaussian kernel algorithmsiLocal density ρi：
Wherein dcSelect be so that per node on average neighbours' number for all number of nodes 1%-2%；
Distance δi
WhereinIt isDescending arrangement；
So far, for the arbitrary node x in set of node Si, (ρ can be calculatedi,δi),i∈IS, while there is larger ρ values
Point with δ values is cluster centre, while can calculate γi=ρiδi, γ values are bigger, are more likely to be cluster centre.
The improvement K-means algorithms input：
Node data collectionWith k determining initial cluster center；
The improvement K-means algorithms export:K cluster, meets criterion functionConvergence；
Wherein：K is cluster number, kiIt is the interstitial content in the i-th cluster, wijIt is j-th of node in the i-th cluster,It is i-th
The cluster center of cluster,It is defined as foloows
The improvement K-means algorithm performs following steps：
A) according to distance by node division to apart from nearest cluster；
B the average value for) recalculating object in each cluster updates the cluster centre of cluster；
C A~B) is repeated, until criterion function E no longer changes.
The RBF neural using PSO algorithm optimizations carrys out the temporal and spatial correlations of mining data lost clusters interior joint data
Detailed process is：
Determine that RBF neural parameter obtains optimum network structure using APSO algorithm is improved, it is then sharp
The temporal correlation that ocean sensing data is excavated with APSO algorithm and obtained optimum network structure, finally in cluster
Realize the accurate recovery and reconstruct for losing data.
It is described in losing node data recovery process, if individual node cluster, utilize the node history wheel data
Data recovery is carried out, if multiple node clusters, works as front-wheel data using other nodes in node history wheel data and cluster are lost
Carry out data recovery.
Described determines that RBF neural parameter obtains optimum network structure tool using APSO algorithm is improved
Body is：
If RBF neural has k center, each center to be tieed up for m, then the position of particle is tieed up for k × (m+1), accordingly
The speed of particle is also that k × (m+1) is tieed up, and adds the fitness of particle, the coding structure of particle is as follows：
X11X12…X1mσ1…Xi1Xi2…Ximσi…Xk1Xk2…Xkmσk
V1V2…Vk×(m+1)
f(x)
Wherein：Xi1Xi2…XimThe position at a neural network center (i=1 ..., k) for i-th；σiFor the width of basic function；
V1V2…Vk× (m+1) is the speed of particle；F (x) is the fitness function of particle；
The purpose of RBF neural training is that the optimal value for finding parameter makes its node loss data restore mean square error
And ERRiMinimum, therefore select ERRiInverse be fitness function, i.e., as fitness value fiWhen being maximized, RBF neural
Structure is optimal, the fitness function f of i-th of individualiIt is as follows
Wherein, ERRiFor the Mean Square Error of node loss data recovery value, ykFor practical marine environmental monitoring data
Value,For node loss data recovery value.
Compared with prior art, the present invention haing the following advantages：
Compared to the prior art wireless sense network loss of data in ocean of the present invention, the advantage is that, this hair with reconstructing method
Bright to carry out cluster cluster in real time to OWSNs nodes using improvement K-means algorithms, this will be well positioned to meet ocean and wirelessly pass
Feel the characteristic of the real-time dynamic change of net topology structure, while being excavated using the good non-linear mapping capabilities of APSO-RBFNNs
With the temporal correlation of cluster interior nodes data, data value is lost so as to accurately recovery nodes.
If all-network node is all sent to Sink node by environmental data is perceived, a large amount of energy is not only consumed, but also
Limited OWSNs bandwidth resources are wasted, data packet collisions are caused, communication efficiency is reduced, increases the energy consumption and time delay of data transmission.
The loss data recovering algorithms of this paper are used at node data transmission both ends, if restoring data and actual monitoring data in user
In given error threshold, then monitoring data is replaced with recovery data, reduce the transmission quantity of data in network, to reduce net
The energy consumption of network and the collecting efficiency for improving ocean big data.
Description of the drawings
Fig. 1 is the flow chart that a kind of ocean wireless sensor network of the present invention loses data recovery and reconstructing method.
Fig. 2 is that line sensing net node clusters cluster schematic diagram using K-means algorithms are improved.
Fig. 3 is the position topology of 54, the laboratories Intel node and is illustrated using K-means algorithms cluster cluster is improved
Figure.
Fig. 4 A, Fig. 4 B are data recovery of the present invention and the prediction result of reconstructing method and the comparison diagram of truthful data.
Fig. 4 C are the data recovery errors figures of data recovery and reconstructing method of the present invention.
Fig. 5 is the method for the present invention figure compared with the restoration errors of NCGP and MASTER.
Fig. 6 is the method for the present invention, NGCP and the comparison diagram without using volume of transmitted data when data recovering algorithms.
Specific implementation mode
The present invention is further elaborated by the way that a preferable specific embodiment is described in detail below in conjunction with attached drawing.
Restore with reconstructing method to include node as shown in Figure 1, a kind of ocean wireless sensor network (OWSNs) loses data
It clusters cluster and loses data and restore.
The node clustering cluster includes:
In sea area to be monitored, arrangement meets the node of OWSNs communicating requirements and real-time detection requirement first, completes sea
The topology and route construction of foreign wireless sensor network, it is then effectively poly- to deployment node progress according to K-means algorithms are improved
Class cluster.Specifically, disposing n sensor node for marine site to be monitored, it is ensured that realize that marine site to be monitored is wirelessly passed by ocean
Sensor network signal all standing.Wireless sensor network P={ the p being made of n node1,p2,…,pn, it is generated by network
Data G={ g1,g2,…,gn, wherein giIt is node piThe time series data collection H={ h acquired as the period using time Ti
(t1),hi(t2),hi(t3)…}.N node restores data R={ R in network1,R2,…,Rn, interior joint piRecovery data
Value is Ri={ Ri(t1),Ri(t2),Ri(t3)…}.The OWSN communicating requirements are：The sensor node is monitoring marine site certainly
It organizes the formation of wireless sense network and completes topology structure initialization, and the comprehensive coverage goal of the network monitors marine site.
Secondly, the local density's ρ values and relative distance δ values of each sensor node are calculated using formula (1)-(4).Such as
Shown in Fig. 2, disposes sensor node for a monitoring marine site and cluster cluster schematic diagram, ultimately form CH1, CH2, CH3 and CH4 tetra-
Cluster.Wherein collected environmental information is sent to cluster head node by ordinary node, and then cluster head node will receive ordinary node
Information and the collected information of own node are sent to base station.
Item nodes cluster cluster process in the rooms Intel：The project is opened by Intel Berkeley Research Lab
Exhibition deploys 54 nodes that TinyOS operating systems are housed in test room, and it includes temperature that node acquires for every 31 seconds, wet
The data such as degree, illumination and voltage, and be collected by TinyDB processing networks.Collecting work was by 2 28th, 2004
It is continued until April 5.The position of 54, the laboratories Intel node is topological and clusters cluster using K-means algorithms are improved
As a result as shown in Fig. 3.It is as follows that it clusters cluster process：1) the local density ρ of each node is calculatediValue：Due to cut-off
Kernel is centrifugal pump, and Gaussian kernel are successive value, so calculating different sensors using Gaussian kernel
Node has the probability of identical local density values smaller, therefore the present invention calculates each sensor section using Gaussian kernel
The local density values of point, formula are as follows：2) the relative distance δ of each node is calculatediValue：
WhereinIt isDescending arrangement.
3) so far, for the arbitrary node x in set of node Si, (ρ can be calculatedi,δi),i∈IS.There is larger ρ simultaneously
The point of value and δ values is cluster centre.γ can be calculated simultaneouslyi=ρiδi, γ values are bigger, are more likely to be cluster centre.4) basis
Other non-leader cluster nodes are divided into apart from nearest cluster by distance；5) average value of object in each cluster is recalculated, cluster is updated
Cluster centre；6) (4)~(5) are repeated, until criterion functionConvergence；
Wherein：K is cluster number, kiIt is the interstitial content in the i-th cluster, wijIt is the jth node in the i-th cluster,It is
The cluster center of i clusters,It is defined as foloows
Improving the input of K-means algorithms is：
Node data collectionWith k determining initial cluster center；
Improving the output of K-means algorithms is:K cluster meets criterion function convergence.
It is calculated by above-mentioned, 54 nodes of the laboratories Intel deployment are divided into 7 clusters.Here (the packet of cluster 1 is selected
Containing node 1-5) and two groups of temperature data collection of cluster 3 (include node 20-27) restore and reconstruct side to verify the loss data of proposition
Method, node 4 and 22 are respectively respective cluster head.
The loss data are restored：When node data is lost, the RBF nerve nets of PSO algorithm optimizations are used
Network carrys out the temporal correlation of mining data lost clusters interior joint data, so according to history wheel in lost clusters and when front-wheel data come
Recovery nodes lose data value.
The above-mentioned RBF neural using PSO algorithm optimizations carrys out the temporal and spatial correlations of mining data lost clusters interior joint data
Detailed process is：
Determine that RBF neural parameter obtains optimum network structure using APSO algorithm is improved, it is then sharp
The temporal correlation that ocean sensing data is excavated with APSO algorithm and obtained optimum network structure, finally in cluster
Realize the accurate recovery and reconstruct for losing data.
It is described in losing node data recovery process, if individual node cluster, utilize the node history wheel data
Data recovery is carried out, if multiple node clusters, works as front-wheel data using other nodes in node history wheel data and cluster are lost
Carry out data recovery.
Described determines that RBF neural parameter obtains optimum network structure tool using APSO algorithm is improved
Body is：
If RBF neural has k center, each center to be tieed up for m, then the position of particle is tieed up for k × (m+1), accordingly
The speed of particle is also that k × (m+1) is tieed up, and adds the fitness of particle, the coding structure of particle is as follows：
X11X12…X1mσ1…X21X22…X2mσ2…Xk1Xk2…Xkmσk
V1V2…Vk×(m+1)
f(x)
Wherein：Xi1Xi2…XimThe position at a neural network center (i=1 ..., k) for i-th；σiFor the width of basic function；
V1V2…Vk× (m+1) is the speed of particle；F (x) is the fitness function of particle
The purpose of neural metwork training be find parameter optimal value make its node loss data restore mean square error and
ERRiMinimum, therefore select ERRiInverse be fitness function, i.e., as fitness value fiWhen being maximized, RBF neural knot
Structure is optimal, the fitness function f of i-th of individualiIt is as follows：
Wherein, ERRiFor the Mean Square Error of node loss data recovery value, ykFor practical marine environmental monitoring data
Value,For node loss data recovery value.
In a particular embodiment, node loss data recovery comprises the steps of：
Step 1, initialization of population.The population that a population is N is randomly generated, the position and speed of particle is initialized.
The initial parameter (weights of Basis Function Center value and width, hidden layer to output layer) of RBF neural takes in (- 8,8) section
Arbitrary value.
Step 2, sequence.The fitness value of each particle is calculated using formula (12) and descending arranges.
Step 3 calculates speed and the position for updating each particle using formula (5) and formula (6), then update Pbest values and
Gbest values.
Wherein w, c1And c2It is adaptively determined by formula (1)-(3)：
It is all in order to determine the optimum combination of tri- parameter values of α, β and γ (α, beta, gamma ∈ { 0.5,1,1.5,2,2.5 })
Combination must all be tested.α, β and γ are possible to combine a total of 53=125 kinds of combinations.If executing all combinations to survey
Examination, this will expend a large amount of computing resource.Here we are in order to sample out one group small of representative α, β and γ group zygote
Collect and takes a kind of cross design technique.L25(56) it is an orthogonal array, it can contain up to 5 with 25 emulation experiment processing
Six variables of a value.Therefore, only need to execute 25 emulation experiments can determine the optimum combination of α, β and γ value.Finally, pass through
The optimum combination value that emulation obtains α, β and γ is α=2, β=0.5, γ=1.5.
Step 4 generates new population.
If the fitness value of particle cannot meet algorithm end condition after step 5, update, repeatedly step 2-4.It is no
Then, RBF neural optimized parameter scheme is exported.
Step 6 loses data prediction using the RBFNNS optimum structures obtained in step 5 to execute ocean wireless sense network
With recovery.
Available by observing data set, 54 nodes of project are collected in 84600 time slice institutes in the rooms Intel
Data have 23% initial data to lose.Our target is by other nodes in N4 nodes history wheel data and cluster
(N1, N2, N3, N5) estimates the temperature data missing value of leader cluster node N4 when front-wheel data.We set historical sensor reading
Number is predicted to lose information as terminal.We have chosen 1200 temperature datas of N4 nodes and do simulation and prediction to it.Fig. 4 A, figure
4B, Fig. 4 C are N4 node the simulation experiment results.From Fig. 4 A, Fig. 4 B, Fig. 4 C we can see that the present invention flutter grasped except
All temperature data information other than crest value, and restoration errors are within 0.1.
We compare the performance of the method for the present invention and different data recovery algorithms under event of data loss below.Data are lost
Mistake rate is set between 10% to 90%.Fig. 5 is demonstrated by the data recovery performance of several algorithms.X-axis represents losing probability, Y
Axis indicates restoration errors RE (Recovery Error).In general, RE is improved with the increase of Loss Rate.The method of the present invention exists
Best performance is achieved under ocean temperature data set.Even if 80% data are lost, the restoration errors of the method for the present invention
Again smaller than 25%, and the error that the error for corresponding to MASTER algorithms is more than 60%, NGCP algorithms is 40%.
Invention additionally discloses a kind of ocean wireless sense network power-economizing methods, if all nodes all send monitoring of environmental data
To aggregation node, a large amount of communication energy consumption is not only generated, but also waste limited network bandwidth resources, cause channel confliction, dropped
Low communication efficiency increases the energy consumption and time delay of data transmission.Data reconstruction technology proposed in this paper can become to the variation of data
Gesture is prejudged, and the loss data recovering algorithms of this paper is used at node data transmission both ends, if restoring data and reality
Border monitoring data then replace actual node perception data with recovery data, reduce network in the error threshold that user gives
The transmission quantity of middle data, to reduce the energy consumption of network and improve the collecting efficiency of ocean big data.
It is the method for the present invention, NGCP and the comparison without using volume of transmitted data when data recovering algorithms shown in Fig. 6.From Fig. 6
It can be seen that the method for the present invention can reduce by 50% volume of transmitted data compared to NCGP algorithms.In data recovery phase, the present invention
Method is more acurrate compared to NCGP algorithms, and the parameter adaptive calculating time is shorter.Therefore, we are by receiving low volume data just
It can obtain all OWSNs monitoring data.Therefore restore greatly improve ocean big data with reconstructing method using data of the present invention
Collecting efficiency.
Although present disclosure is discussed in detail by above preferred embodiment, but it should be appreciated that above-mentioned
Description is not considered as limitation of the present invention.After those skilled in the art have read the above, for the present invention's
A variety of modifications and substitutions all will be apparent.Therefore, protection scope of the present invention should be limited to the appended claims.
Claims (7)
1. a kind of ocean wireless sensor network loses data recovery and reconstructing method, which is characterized in that including：Node clustering at
Cluster and loss data are restored；
The wherein described node clustering cluster includes：Arrange that meeting ocean wireless sensor network leads in sea area to be monitored first
Letter requires and the node of real-time detection requirement, completes the topology constructing of ocean wireless sensor network, is calculated according to K-means is improved
Method carries out effectively cluster cluster to deployment node；
The wherein described loss data are restored：When node data is lost, the RBF nerve nets of PSO algorithm optimizations are used
Network carrys out the temporal correlation of mining data lost clusters interior joint data, so according to history wheel in lost clusters and when front-wheel data come
Restore to lose data value.
2. ocean wireless sensor network as described in claim 1 loses data recovery and reconstructing method, which is characterized in that institute
It is that sensor node self-organizing in the marine monitoring region forms wireless biography to state ocean network communication of wireless sensor requirement
Sensor network simultaneously completes topology structure initialization, and the network covers monitored sea area comprehensively.
3. ocean wireless sensor network as described in claim 1 loses data recovery and reconstructing method, which is characterized in that institute
The node clustering cluster detailed process stated is：
The selection of cluster centre：
Consider sensor node data collection to be clusteredIS={ 1,2 ..., N } is corresponding index collection, dij=
dist(xi,xj) indicate node xiWith xjThe distance between, for any sensor node xi, define local density ρiWith distance δi
Two amounts portray cluster centre；
X is calculated by Gaussian kernel algorithmsiLocal density ρi：
Wherein dcSelect be so that per node on average neighbours' number for all number of nodes 1%-2%；
Distance δi
WhereinIt isDescending arrangement；
So far, for the arbitrary node x in set of node Si, (ρ can be calculatedi,δi),i∈IS, while there is larger ρ values and δ values
Point be cluster centre, while γ can be calculatedi=ρiδi, γ values are bigger, are more likely to be cluster centre.
4. ocean wireless sensor network as described in claim 1 loses data recovery and reconstructing method, which is characterized in that institute
The improvement K-means algorithms stated input：
Node data collectionWith k determining initial cluster center；
The improvement K-means algorithms export:K cluster, meets criterion function
Convergence；
Wherein：K is cluster number, kiIt is the interstitial content in the i-th cluster, wijIt is j-th of node in the i-th cluster,It is the i-th cluster
Cluster center,It is defined as foloows
The improvement K-means algorithm performs following steps：
A) according to distance by node division to apart from nearest cluster；
B the average value for) recalculating object in each cluster updates the cluster centre of cluster；
C A~B) is repeated, until criterion function E no longer changes.
5. ocean wireless sensor network as described in claim 1 loses data recovery and reconstructing method, which is characterized in that institute
It states and carrys out the temporal and spatial correlations detailed processes of mining data lost clusters interior joint data using the RBF neural of PSO algorithm optimizations and be：
It determines that RBF neural parameter obtains optimum network structure using APSO algorithm is improved, then utilizes certainly
It adapts to particle cluster algorithm and obtained optimum network structure excavates the temporal correlation of ocean sensing data, finally realized in cluster
Lose the accurate recovery and reconstruct of data.
6. ocean wireless sensor network as described in claim 1 loses data recovery and reconstructing method, which is characterized in that institute
State in losing node data recovery process, if individual node cluster, using the node history wheel data carry out data it is extensive
Again, if multiple node clusters, using other nodes in loss node history wheel data and cluster when front-wheel data progress data are extensive
It is multiple.
7. ocean wireless sensor network as claimed in claim 5 loses data recovery and reconstructing method, which is characterized in that institute
State using improve APSO algorithm be specially to determine that RBF neural parameter obtains optimum network structure：
If RBF neural has k center, each center to be tieed up for m, then the position of particle is tieed up for k × (m+1), corresponding particle
Speed be also that k × (m+1) is tieed up, add the fitness of particle, the coding structure of particle is as follows：
X11X12…X1mσ1…Xi1Xi2…Ximσi…Xk1Xk2…Xkmσk
V1V2…Vk×(m+1)
f(x)
Wherein：Xi1Xi2…XimThe position at a neural network center (i=1 ..., k) for i-th；σiFor the width of basic function；
V1V2…Vk× (m+1) is the speed of particle；F (x) is the fitness function of particle；
RBF neural training purpose be find parameter optimal value make its node loss data restore mean square error and
ERRiMinimum, therefore select ERRiInverse be fitness function, i.e., as fitness value fiWhen being maximized, RBF neural knot
Structure is optimal, the fitness function f of i-th of individualiIt is as follows
Wherein, ERRiFor the Mean Square Error of node loss data recovery value, ykFor practical marine environmental monitoring data value,
For node loss data recovery value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810030395.8A CN108289285B (en) | 2018-01-12 | 2018-01-12 | Method for recovering and reconstructing lost data of ocean wireless sensor network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810030395.8A CN108289285B (en) | 2018-01-12 | 2018-01-12 | Method for recovering and reconstructing lost data of ocean wireless sensor network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108289285A true CN108289285A (en) | 2018-07-17 |
| CN108289285B CN108289285B (en) | 2020-11-13 |
Family
ID=62835181
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810030395.8A Active CN108289285B (en) | 2018-01-12 | 2018-01-12 | Method for recovering and reconstructing lost data of ocean wireless sensor network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108289285B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109886825A (en) * | 2018-12-25 | 2019-06-14 | 北京农业信息技术研究中心 | A kind of agricultural internet of things data multi-angle of view projective clustering reconstructing method and system |
| CN110704549A (en) * | 2019-10-09 | 2020-01-17 | 中国石油大学(华东) | Method, system, medium and device for selecting and constructing marine environment data service granularity |
| CN110874645A (en) * | 2019-11-14 | 2020-03-10 | 北京首汽智行科技有限公司 | Data reduction method |
| CN113408644A (en) * | 2021-07-02 | 2021-09-17 | 南京信息工程大学 | Satellite data reconstruction method and method for detecting response of upper ocean to typhoon |
| CN113766359A (en) * | 2021-09-13 | 2021-12-07 | 杭州英迈科技有限公司 | Power monitoring method and system based on sensor network technology |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102014344A (en) * | 2010-09-03 | 2011-04-13 | 浙江工业大学 | Clustering control method of intelligent wireless sensor network based on DPSO (Discrete Particle Swarm Optimization) |
| CN104299035A (en) * | 2014-09-29 | 2015-01-21 | 国家电网公司 | Method for diagnosing fault of transformer on basis of clustering algorithm and neural network |
| CN106528705A (en) * | 2016-10-26 | 2017-03-22 | 桂林电子科技大学 | Repeated record detection method and system based on RBF neural network |
| CN106658577A (en) * | 2016-11-17 | 2017-05-10 | 厦门理工学院 | Sensor network data recovery method combining space-time sparsity |
-
2018
- 2018-01-12 CN CN201810030395.8A patent/CN108289285B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102014344A (en) * | 2010-09-03 | 2011-04-13 | 浙江工业大学 | Clustering control method of intelligent wireless sensor network based on DPSO (Discrete Particle Swarm Optimization) |
| CN104299035A (en) * | 2014-09-29 | 2015-01-21 | 国家电网公司 | Method for diagnosing fault of transformer on basis of clustering algorithm and neural network |
| CN106528705A (en) * | 2016-10-26 | 2017-03-22 | 桂林电子科技大学 | Repeated record detection method and system based on RBF neural network |
| CN106658577A (en) * | 2016-11-17 | 2017-05-10 | 厦门理工学院 | Sensor network data recovery method combining space-time sparsity |
Non-Patent Citations (5)
| Title |
|---|
| J.K. SING: "Improved k-means algorithm in the design of RBF neural networks", 《IEEE》 * |
| J.M. PE~NA: "An empirical comparison of four initialization methods for the k-means algorithm", 《PATTERN RECOGNITION LETTERS》 * |
| SUNG-KWUN OH: "Design of K-means clustering-based polynomial radial basis function neural networks (pRBF NNs) realized with the aid of particle swarm optimization and differential evolution", 《NEUROCOMPUTING》 * |
| WANG SHULIANG: "Clustering by Fast Search and Find of Density Peaks with Data Field", 《CHINESE JOURNAL OF ELECTRONICS》 * |
| 鲜江峰: "基于神经网络的WSN丢失数据恢复研究", 《物联网技术》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109886825A (en) * | 2018-12-25 | 2019-06-14 | 北京农业信息技术研究中心 | A kind of agricultural internet of things data multi-angle of view projective clustering reconstructing method and system |
| CN109886825B (en) * | 2018-12-25 | 2021-06-29 | 北京农业信息技术研究中心 | Agricultural Internet of things data multi-view projection clustering reconstruction method and system |
| CN110704549A (en) * | 2019-10-09 | 2020-01-17 | 中国石油大学(华东) | Method, system, medium and device for selecting and constructing marine environment data service granularity |
| CN110874645A (en) * | 2019-11-14 | 2020-03-10 | 北京首汽智行科技有限公司 | Data reduction method |
| CN113408644A (en) * | 2021-07-02 | 2021-09-17 | 南京信息工程大学 | Satellite data reconstruction method and method for detecting response of upper ocean to typhoon |
| CN113408644B (en) * | 2021-07-02 | 2023-07-14 | 南京信息工程大学 | Satellite data reconstruction method and method for detecting response of upper ocean to typhoon |
| CN113766359A (en) * | 2021-09-13 | 2021-12-07 | 杭州英迈科技有限公司 | Power monitoring method and system based on sensor network technology |
| CN113766359B (en) * | 2021-09-13 | 2024-02-06 | 杭州英迈科技有限公司 | Electric power monitoring method and system based on sensor network technology |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108289285B (en) | 2020-11-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108289285A (en) | A kind of ocean wireless sensor network is lost data and is restored and reconstructing method | |
| Harb et al. | A distance-based data aggregation technique for periodic sensor networks | |
| Huang et al. | A power-efficient routing protocol for underwater wireless sensor networks | |
| Rida et al. | EK-means: A new clustering approach for datasets classification in sensor networks | |
| Shi et al. | Coverage in battery-free wireless sensor networks | |
| CN108055685A (en) | Leader cluster node and sensor node in wireless sensor network | |
| Cao et al. | A novel data fusion strategy based on extreme learning machine optimized by bat algorithm for mobile heterogeneous wireless sensor networks | |
| CN101848529A (en) | Method for compressing multiple principle component analysis data of wireless sensor network | |
| Raj Priyadarshini et al. | Enhancing coverage and connectivity using energy prediction method in underwater acoustic WSN | |
| Wu et al. | Missing data recovery using reconstruction in ocean wireless sensor networks | |
| Cao et al. | Data fusion algorithm for heterogeneous wireless sensor networks based on extreme learning machine optimized by particle swarm optimization | |
| Wang et al. | A survey of compressive data gathering in WSNs for IoTs | |
| Singh et al. | Designing an energy efficient network using integration of KSOM, ANN and data fusion techniques | |
| CN111405605B (en) | Wireless network interruption detection method based on self-organizing mapping | |
| Alsulami et al. | A reliable underwater computing system | |
| Yue et al. | Large-scale mobile wireless sensor network data fusion algorithm | |
| CN111800201B (en) | Method for identifying key nodes of Sink node underwater acoustic sensor network | |
| Molose et al. | Data aggregation schemes for maximal network lifetime | |
| CN111542011B (en) | Layered wireless sensor network clustering routing method based on particle swarm optimization | |
| Talib | Optimization approach to optimal power efficient based on cluster top option in wireless sensor networks | |
| Jellali et al. | Principal component analysis based clustering approach for wsn with locally uniformly correlated data | |
| CN102882727B (en) | Monitoring area partition method for hierarchical monitoring network | |
| Kumar et al. | Data fusion for the Internet of Things | |
| Vatan | Soft Computing based Clustering Protocols in IoT for Precision and Smart Agriculture: A Survey | |
| Zahwe et al. | A fast clustering algorithm for analyzing big data generated in ubiquitous sensor networks |
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 |