CN110247975A - Based on the more equipment collaboration service construction methods of Internet of Things for improving D-S evidence - Google Patents
Based on the more equipment collaboration service construction methods of Internet of Things for improving D-S evidence Download PDFInfo
- Publication number
- CN110247975A CN110247975A CN201910529870.0A CN201910529870A CN110247975A CN 110247975 A CN110247975 A CN 110247975A CN 201910529870 A CN201910529870 A CN 201910529870A CN 110247975 A CN110247975 A CN 110247975A
- Authority
- CN
- China
- Prior art keywords
- node
- awareness apparatus
- data
- value
- service
- 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
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer And Data Communications (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of based on the more equipment collaboration service construction methods of Internet of Things for improving D-S evidence, the method of the present invention is by extracting the semantic information in perception data, and then it designs new atomic service and substitutes existing atomic service, and the input and output of new atomic service are portrayed using semantic information, the loose coupling of upper-layer service logical AND bottom perception data may be implemented, multiple workflows can share identical atomic service when to construction work stream, the reusability of the final adaptability for improving workflow and resource；The resource of multiple similar or inhomogeneous awareness apparatus node can fully be utilized by Data fusion technique, be combined according to certain Optimality Criteria, be subject to complementation, reduce uncertainty, to provide the relatively complete consistent description to perceptive object.Compared with single awareness apparatus node, the reliability and robustness of acquisition data, the confidence level of reinforcer the Internet services can be improved with the Data fusion technique of more device nodes.
Description
Technical field
The present invention relates to internet of things field, and in particular to a kind of based on the more equipment associations of Internet of Things for improving D-S evidence
With service construction method.
Background technique
Internet of Things (Internet of Things, IoT) is the important component of generation information technology.Internet of Things is general
It reads and receives the extensive concern of countries in the world government and research institution once proposition, started third time information industry revolution.Object
Networking is that a kind of there is mark, perception, the equipment of processing capacity and internet to be attached magnanimity by various access technologies
It is formed by a kind of large-scale network.Interpersonal information exchange is only realized different from internet, Internet of Things connects
Physical world and information world have been connect, the interaction of object Yu object, object and people, person to person is realized, has greatly enriched people's information
The ability of acquisition, promote people deeper into perception and understanding physical world.However traditional Internet of Things is applied using one
Kind closing private network form, due to standard disunity, using verticalization so that be between different sensor networks it is isolated,
Various sensings and access device can not achieve shared.And the development of Internet of Things, it is necessary to by internet that these are originally isolated
Sensor network connect, realize the shared of resource.Services Oriented Achitecture (Service-oriented
Architecture, SOA) for the building and application integration of such complication system provide ideal solution.SOA is will to answer
Business logic modules turn to different functional units (referred to as service), by defined between these services good interface and
Contract connects.For interface independently of hardware platform, operating system and the programming language for realizing service, this makes building various
Service in system can be interacted with a kind of unification and general mode.The core concept of SOA is the reuse of resource, i.e., one
It can be used for multiple applications and operation flow after a service construction.
(1) under normal conditions, the service of Internet of Things is designed to lesser unit, that is, realizes more single function, such as
Temperature obtains service etc..It is not only because lesser service unit and is easier to realize reuse, more importantly these take
Business is generally directly or indirectly deployed in resource-constrained equipment, the excessively complicated efficiency to these equipment of service function design
It will be very big test.
(2) simultaneously, Internet of Things possesses different types of equipment of substantial amounts, and collected data resource is magnanimity and more
Seed type, these cause the quantity of service under environment of internet of things to be also magnanimity.The service of magnanimity necessarily give services selection and
Services Composition brings very big burden.
(3) internet of things equipment is resource-constrained, and awareness apparatus node leads to the letter of acquisition due to own hardware or chain environment
Breath has certain unreliability.
Therefore, effective service construction just becomes very necessary.By effective service construction, resource-constrained devices are reduced
Resource consumption, reduce the complexity of services selection and Services Composition.
Internet of Things application is the data by obtaining awareness apparatus acquisition and carries out certain working process to it to realize
's.Therefore, the acquisition of Internet of Things resource is the precondition that Internet of Things application is achieved.The acquisition modes of Internet of Things resource are big
Body can be divided into four kinds, i.e., the method based on API, the method based on database, the method based on script and based on service
Method.In method based on API, awareness apparatus is open in a manner of API by the ability of itself, calls accordingly using as needed
API obtain resource, such as nesC；In method based on database, thing network sensing layer is considered as a distributed data base, answers
The resource needed for being obtained by gateway node using class SQL query language, such as TinyDB and IrisNet；Method based on script
In, using obtaining required resource to sensing layer related perceiving device node injection script program by gateway node or to perception
Equipment, which carries out corresponding configuration, makes it obtain perception data active upload resource simultaneously, such as Agilla and SensorWare；It is based on
In the method for service, resource is encapsulated as servicing, and obtains resource by the method for service discovery.
Since the method based on service preferably realizes the loose coupling of upper layer application and resource, and the thought screen of modularization
Bottom layer realization details has been covered, the exploitation of application has been simplified, gets the attention recently.Flavia Coimbra Delicato
Et al. the concept of service is introduced into sensor network in 2003.They point out traditional sensor network deployment when just with spy
Fixed upper layer application binding, and should be able to merge various heterogeneous devices and network richer to support for following sensor network
Rich more complicated application.Therefore the separation in order to realize upper layer application and underlying infrastructure, they propose with service
Method designs sensor network.Sensor network nodes are divided into two classes: ordinary node and aggregation node by them.Ordinary node tool
Have an ability of perception and routing, and aggregation node does not have perceptional function, protocol conversion is only provided realize sensor network with
The communication of external network.Upper layer application is equivalent to service requester to sensor network request data, and aggregation node maintenance is entire
All service describings of sensor network simultaneously provide the call method of these services, are equivalent to ISP.And it converges simultaneously
Node, to ordinary node request data, is equivalent to service requester, ordinary node response convergence section according to the demand of upper layer application
The request of data of point is equivalent to ISP.The tool that Flavia Coimbra Delicato et al. is realized without providing service
Body details, but the service-oriented blank of sensor network is depicted, it is specified for the realization of sensor network Service-Oriented Architecture Based
Direction.
N.Y.Othman et al. proposes the sensor network service model of no aggregation node in the literature.Ordinary node is first
Node location and the data type that can be provided, data format etc. are first encapsulated as service description information, and are registered to sensing
The service register center of the network overlapped layer of device.Next, overlapping layer middleware is responsible for believing the service describing of service register center
The corresponding service of breath is mapped to corresponding ordinary node.Compared to the service model for having aggregation node, ordinary node directly as
ISP provides data service for upper layer application, has lesser energy consumption and network load.
It is worth noting that, the service model without aggregation node that N.Y.Othman et al. is proposed is built upon node energy
Power is powerful enough, and on the basis of embedded service platform (TinyWS) and IP protocol stack can be supported to load, there is its office
It is sex-limited.Awareness apparatus is divided into wholly-owned source device node, limited resources device node (such as by Jeremie Leguay et al.
Personal Digital Assistant, PDA) and low resource device node (such as Crossbow sensing node).For complete
Resource apparatus node uses the service agreement of standard, such as WSDL, UDDI and SOAP；DPWS is used for limited resources device node
(Devices Profile for Web Services)；For low resource device node, WSN-SOA is proposed in text, one
Simple agreement and software architecture.By the SOA protocol stack of lightweight, enable the service of building of low resource device node, and
It can be found that, call other equipment node service.On this basis, Jeremie Leguay et al. gives SOA and realizes frame
Structure supports the dynamic of network, self-configuring, service discovery and heterogeneous device interaction.
Current techniques use rule-based reasoning technology, extract the semantic information in perception data, realize upper layer industry
The loose coupling of business logical AND bottom sensing data.Wherein, Rule Engine is the core of existing model, by predetermined
Rule extraction goes out the semantic information of current sensor data；Service provider establishes atomic service, and uses the semantic information extracted
The input and output of atomic service are portrayed, realize the loose coupling of upper-layer service logic and bottom sensing data, the original of foundation
Sub-services are stored in Atomic Services Repository；Workflow Engine is used for the foundation of workflow, monitoring
And execution.
It include three phases when regulation engine is judged: matching, selection and execution.Wherein, the prior art uses RETE
Algorithm efficiently realizes matching stage, and the conflict resolving algorithms selection optimal rules based on context-aware priority.
This method needs to edit rule, extracts semantic information using regulation engine, is easy to produce when regular quantity is larger
The problems such as raw rule conflict, and the fusion of perception data is not considered.
Bibliography:
[1]Gay.David,Levis.Philip,Von Behren.Robert et.al.The nesC language:A
holistic approach to networked embedded systems.Acm Sigplan Notices.38(5),pp:
1-11.2003.
[2]Madden.Samuel R,Franklin.Michael J,Hellerstein.Joseph M,
et.al.TinyDB:An acquisitional query processing system for sensor networks[J]
.ACM Transactions on Database Systems.30(1),pp:122-173.2005.
[3]Campbell.Jason,Gibbons.Phillip B,Nath.Suman,et.al.IrisNet:an
internet-scale architecture for multimedia sensors[C].In Proceedings of the
13th annual ACM international conference on Multimedia.pp:81-88.2005.
[4]Fok.Chien-Liang,Roman.Gruia-Catalin,Lu.Chenyang,et.al.Agilla:A
mobile agent middleware for self-adaptive wireless sensor networks[J].ACM
Transactions on Autonomous and Adaptive Systems.4(3),pp:16.2009.
[5]Boulis.Athanassios,Han.Chih-Chieh,Shea.Roy,et.al.SensorWare:
Programming sensor networks beyond code update and querying[J].Pervasive and
Mobile Computing.3(4),pp:386-412.2007.
[6]Delicato.F.C,Pires.P.F,Pinnez.L,et.al.A flexible web service based
architecture for wireless sensor networks[C].In Proceedings of the 23rd
International Conference on Distributed Computing Systems(ICDCSW 2003),pp:
730-735.2003.
[7]Othman,N.Y.；Chebbine,S.；Khendek,F；Glitho,R.A Web Services-Based
Architecture for the Interactions between End-User Applications and Sink-less
Wireless Sensor Networks[C].In Proceedings of the 4th International
Conference on Consumer Communications and Networking(CCNC2007),pp:865-
869.2007.
[8]Endong Tong,Wenjia Niu,Hui Tang,Gang Li,and Zhi jun Zhao."
Reasoning-based Context-aware Workflow Management in Wireless Sensor
Network".The 9th International Conference on Service Oriented Computing(ICSOC
2011),pp:270-282.2011.
Summary of the invention
In view of the deficiencies of the prior art, the present invention is intended to provide it is a kind of based on the more equipment associations of Internet of Things for improving D-S evidence
With service construction method, the loose coupling of upper-layer service logical AND bottom perception data is realized, and solve to set up slave node due to certainly
The problem of perception data poor reliability caused by body hardware or chain environment problem, provides to the relatively complete consistent of perceptive object
Description, it is credible to improve service.
To achieve the goals above, the present invention adopts the following technical scheme:
Based on the more equipment collaboration service construction methods of Internet of Things for improving D-S evidence, include the following steps:
S1, semantic information in perception data is extracted, designs new semantic-based atomic service and use extracts
Semantic information portray it and output and input, substitute existing atomic service with new atomic service；
S2, the basic probability assignment letter that each hypothesis of identification framework in D-S evidence theory is generated using subordinating degree function
Number；
S3, according to the significance level of each awareness apparatus node, assign different weights for different awareness apparatus nodes,
And weight redistributes awareness apparatus node to the Basic probability assignment function of each hypothesis, as shown in formula (1) accordingly:
Wherein, mi(Aj) it is the basic probability assignment value that i-th of awareness apparatus node before redistributing assumes j-th；
mi'(Aj) it is the basic probability assignment value that i-th of awareness apparatus node after redistributing assumes j-th；WiFeel for i-th
Know the weight of device node；For the maximum value of weight in all n awareness apparatus nodes；Be shown to be meet basic probability assignment value and for 1 constraint condition, extra basic probability assignment value is assigned
Θ；
S4, the Basic probability assignment function finally redistributed as obtained in step S3, it is public according to D-S evidence fusion
Formula obtains final conclusion.
Further, in step S3, there is the weight of each awareness apparatus node dynamic to repair in Basic probability assignment function
Positive process, detailed process are as follows:
(1) the history weight based on trust management is generated:
When the data of certain awareness apparatus node are consistent with the conclusion finally merged, it is believed that the awareness apparatus node is just
Data really are acquired, SR is usediRecord awareness apparatus node i correctly acquires the ratio of the total times of collection of number Zhan of data；
When the data of certain awareness apparatus node are runed counter to the conclusion finally merged, then it is assumed that the awareness apparatus node is mistakenly adopted
Collect data, uses URiRecord the ratio of the total times of collection of number Zhan of awareness apparatus node i mistake acquisition data；
It carries out calculating time attribute t as the following formulasiAnd tui:
Wherein, TcurrentIt is current time, STiFor the time of the last correct acquisition data of awareness apparatus node i, UTi
For the time of the last mistake acquisition data of awareness apparatus node i, Δ T is the threshold value of setting；
According to time attribute tsi, tuiValue, be arranged SRiAnd URiWeight wsiAnd wui:
X, y, z are artificial setting value, x > y > z；
The history weight HW of awareness apparatus node i is calculated as followsiAre as follows:
HWiValue range is [0,1]；
(2) the instant weight based on similarity is generated:
Note service has n awareness apparatus node, is denoted as N1, N2..., Nn, acquired data values are respectively D1, D2...,
Dn；Awareness apparatus node NiAnd NjDistance definition be its acquire data normalization after difference, calculated as the following formula:
I, j=1,2 ..., n and i ≠ j (5)
Wherein, dist (Ni,Nj)=Di-Dj, dist_min indicates in the dist () between all awareness apparatus nodes most
Small value, dist_max indicate the maximum value in the dist () between all awareness apparatus nodes；
Awareness apparatus node NiAnd NjBetween distance it is bigger, show that its similarity is smaller；Awareness apparatus node is defined as a result,
NiAnd NjSimilarity is as follows:
Sim(Ni,Nj)=1-Dist (Ni,Nj), i, j=1,2 ..., n and i ≠ j (6)
The summation of each awareness apparatus node and the similarity of other n-1 awareness apparatus node is calculated by formula (7):
It is different senses according to the summation of each awareness apparatus node and the similarity of other n-1 awareness apparatus node
Know that device node distributes instant weight CW:
Wherein
(3) it is calculated as follows to obtain the amendment weight of each awareness apparatus node:
Wi=λ HWi+(1-λ)·CWi (9)
Wherein, λ is the distribution factor being manually set, and value interval is [0,1], and λ numerical value is bigger, and history weight is to final power
The value influence of value is bigger, on the contrary then smaller.
Further, x=0.2, y=0.3, z=0.5.
The beneficial effects of the present invention are:
1, the method for the present invention is by extracting the semantic information in perception data, and then designs new atomic service substitution
There is atomic service, and portray the input and output of new atomic service using semantic information, upper-layer service logical AND bottom may be implemented
The loose coupling of layer perception data, so that multiple workflows can share identical atomic service when construction work stream, it is final to improve
The adaptability of workflow and the reusability of resource；
2, the method for the present invention can be by the money of multiple similar or inhomogeneous awareness apparatus node by Data fusion technique
Source fully utilizes, and combines according to certain Optimality Criteria, is subject to complementation, uncertainty is reduced, to provide to sense
Know the relatively complete consistent description of object.Compared with single awareness apparatus node, with the data fusion skill of more device nodes
Art can be improved the reliability and robustness of acquisition data, the confidence level of reinforcer the Internet services.
Detailed description of the invention
Fig. 1 is the overall procedure schematic diagram of the embodiment of the present invention；
Fig. 2 is the membership function exemplary diagram in the embodiment of the present invention.
Specific embodiment
Below with reference to attached drawing, the invention will be further described, it should be noted that the present embodiment is with this technology side
Premised on case, the detailed implementation method and specific operation process are given, but protection scope of the present invention is not limited to this reality
Apply example.
The present embodiment provides a kind of based on the more equipment collaboration service construction methods of Internet of Things for improving D-S evidence, and process is such as
Shown in Fig. 1.
Internet of Things application be it is data-centered, Service-Oriented Architecture Based realize in, atomic service is tight with perception data
Coupling, therefore there are the atomic services of bulk redundancy, cause the wasting of resources of Internet of Things.With room temperature control service in smart home
For (obtain the temperature value in room and judge whether to need to open/close air-conditioning).Under different application scenarios, temperature is high
The condition of low judgement is different (T > 58 DEG C, T > 100 DEG C etc.).Therefore it needs correspondingly to construct original for different application scenarios
Sub-services.But these services include same semantic information in fact, i.e. Current Temperatures are high.
The present embodiment method introduces inference technology, extracts the semantic information in perception data, and then service provider sets
Count new atomic service and substitute existing atomic service, wherein the input and output of new atomic service by the semantic information that extracts into
Row is portrayed, and the loose coupling of upper-layer service logical AND bottom perception data is realized, so that multiple workflows can in construction work stream
To share identical atomic service, the reusability of the final adaptability for improving workflow and resource.Or it is controlled and is applied with room temperature
For, new atomic service (judging that Current Temperatures are high or low) can be constructed and use semantic information (temperature is high or low) as defeated
Enter output to substitute existing atomic service (judging whether temperature value is greater than 58 DEG C or 100 DEG C etc.).In this way, although locating application
Scene is different, and temperature control application can share created atomic service.
In addition, internet of things equipment is resource-constrained, awareness apparatus node causes to adopt due to own hardware or chain environment problem
There are certain unreliabilities for the information of collection.Rely on the problem of poor reliability also will be present in the service of the perception data.However, object
There are a variety of different types of awareness apparatus nodes under networked environment, and there is also redundant deployments for same type of device node.
The collected information type of these device nodes may be also different: fuzzy perhaps determining is accurate or incomplete
, the resource of multiple similar or inhomogeneous awareness apparatus node can fully be utilized by Data fusion technique, foundation
Certain Optimality Criteria combines, and is subject to complementation, uncertainty is reduced, to provide relatively complete one to perceptive object
The description of cause.Compared with single awareness apparatus node, it can be improved acquisition data with the Data fusion technique of more device nodes
Reliability and robustness, the confidence level of reinforcer the Internet services.
The premise of D-S evidence theory is the Basic probability assignment function of each event in identification framework to be obtained.However, true
In the real world, all there is multi-meaning and probabilistic bloomings for the thinking concept of people and feeling judgement etc..
By taking the temperature judgement service in smart home as an example, identification framework is { high, low, moderate }.Awareness apparatus node acquisition be
One specific temperature value, but how much numerical value can be considered as specifically temperature height, everyone definition is not quite similar, also
It is the state for saying these things, mostly and the judgement of the subjective sensation of people has substantial connection.
Fuzzy theory is taught by bundle moral (L.A.Zadeh) and is proposed in nineteen sixty-five.Its " fuzzy set theory " text delivered is opened
Invasive proposes use " degree of membership " (Membership Grade) and " subordinating degree function " (Membership Function)
Concept come correctly describe with the blooming in Coping with Reality, thus breach classical sets opinion in logic true value { 0,1 }
Based on determination mathematical logic.
Fuzzy theory uses subordinating degree function, is more in line with the understanding mode of people.The concept of degree of membership and D-S are demonstrate,proved simultaneously
There is very big similitude according to the Basic probability assignment function in theory.Therefore, the present embodiment method using subordinating degree function come
Generate the Basic probability assignment function of identification framework in D-S evidence theory.Or by taking room temperature control service as an example, using such as Fig. 2
Shown subordinating degree function.
Room temperature controls identification framework Θ={ high and low, moderate } of service, only gives in the present embodiment and assumes "high" and vacation
The subordinating degree function of " if low ".As shown in Figure 2, when temperature value is 25 DEG C, the Basic probability assignment function respectively assumed is respectively m
(height)=0.33, m (low)=0.16, probability assignments and not be 1.For the precondition for meeting Basic probability assignment function, no
The part of foot 1 can be assigned to hypothesis " moderate ", i.e. m (moderate)=1-m (height)-m (low)=0.51.By being subordinate in fuzzy theory
The mode of category degree obtains the mode of thinking that basic allocation probability function is more in line with people, while using fuzzy obtained semantic letter
Breath portrays the input/output of service, realizes the loose coupling of bottom specific perception data and upper layer application service logic, reduces
The adaptability of service is also improved while quantity of service.
By the available each evidence of subordinating degree function to the Basic probability assignment function of hypothesis each in identification framework.?
In the service construction of more equipment collaboration perception, the data (i.e. a plurality of evidence) of multiple equipment node perceived can be obtained, then just
It needs to merge this plurality of evidence, to obtain respectively assuming reliable trust value distribution in identification framework.However, traditional
Evidence fusion theoretical method has an inevitable defect.When between a plurality of evidence merged there are larger conflict when
It waits, fusion results are often disagreed with intuitive judgment, and here it is famous " Zadeh antinomys ".
It is merged using the information that Dempster combining evidences rule is included to two evidences, as shown in table 1.From table
In it can be seen that, evidence one with 90% probability support assume A, with 10% probability support C, and completely negate assume B, evidence
Two support to assume B with 90% probability, support C with 10% probability, and completely negate to assume A.
Table 1 " Zadeh " is contrary to table
| m(A) | m(B) | m(C) | |
| Evidence one | 0.9 | 0 | 0.1 |
| Evidence two | 0 | 0.9 | 0.1 |
From the point of view of basic probability assignment, there are biggish conflicts between evidence one and evidence two, if using traditional card
It is carried out according to blending algorithm, as a result: m (A)=0, m (B)=0, m (C)=1, final conclusion are that C is assumed in complete support, this is aobvious
So runed counter to convention.Cause such the result is that because evidence one is being merged with evidence two in traditional evidence fusion algorithm
There is identical significance level in the process, and two evidences have obliterated remaining evidence to the vacation to the complete negative of hypothesis A and B respectively
If high supporting rate.
In fact, different awareness apparatus has different reliabilities.The technological level of different awareness apparatus and current institute
The factors such as place's network environment may all influence the accuracy of its perception data.Additionally, it is possible to which there are part awareness apparatus node evils
Meaning sends the data of inaccuracy.Therefore, different evidences not necessarily has identical significance level, cannot will be each in practical application
Evidence is directly merged using D-S fusion formula.It is different cards by considering to participate in the significance level of each evidence of fusion
According to the different weight of imparting, and weight redistributes evidence to the Basic probability assignment function of each hypothesis accordingly, and then overcomes
Zadeh antinomy.
According to the weight of different evidences, the Basic probability assignment function of evidence is corrected, as shown in formula (1).
Wherein, mi(Aj) it is the basic probability assignment value that i-th of awareness apparatus node before redistributing assumes j-th；
mi'(Aj) it is the basic probability assignment value that i-th of awareness apparatus node after redistributing assumes j-th；WiFeel for i-th
Know the weight of device node；For the maximum value of weight in all n awareness apparatus nodes；Be shown to be meet basic probability assignment value and for 1 constraint condition, extra basic probability assignment value is assigned
Θ；
Finally by corrected Basic probability assignment function, final determine is obtained according to traditional D-S evidence fusion formula
Plan.By distributing corresponding weight for different awareness apparatus nodes (i.e. evidence), to overcome Zadeh antinomy to provide possibility,
But environment of internet of things has stronger dynamic.The normal awareness apparatus node of work at present may go out in the next time
There is biggish deviation in the case where existing environmental disturbances enhancing, the data for causing it to perceive.Therefore, the weight of awareness apparatus node is not
It should be unalterable, but dynamic corrections.
The history lists of awareness apparatus has stronger reference value referring now to its weight is measured.But it is based only on history lists
It is existing, there can be the case where response lag.Therefore, instant weight is introduced, i.e., the similarity of each awareness apparatus node comprehensively considers sense
The history weight and instant weight for knowing device node obtain the correction value of weight.
(1) the history weight based on trust management is generated
When the data of certain awareness apparatus node are consistent with the conclusion finally merged, it is believed that the node correctly acquires
Data.Use SRi(Successful Collection Ratio) record awareness apparatus node i correctly acquires time of data
The ratio of the number total times of collection of Zhan, SR value is bigger, and trust value is accordingly bigger；Similarly, when the data of certain awareness apparatus node and most
When merging obtained conclusion eventually and runing counter to, then it is assumed that the node mistakenly acquires data.Use URi(Unsuccessful
Collection Ratio) ratio that node i mistake acquires the total times of collection of number Zhan of data is recorded, UR value is bigger, trusts
Value is corresponding smaller.Node correctly acquires the number of data and the number of mistake acquisition data for node trust value in history
Calculating has important influence.SR and UR is measured therefore, it is necessary to comprehensive, assigns different weights for awareness apparatus node.
It is worth noting that, data continuously correctly ought be acquired a period of time interior nodes recently, then the trust value of the node is answered
This increases sharply；Equally, when the continuous mistake acquisition data of nearest a period of time interior nodes, then the trust value of the node should be rapid
Reduce.T is used in the present embodimentsiAnd tuiIndicate the attribute.
Wherein, TcurrentIt is current time, STiFor the time of the last correct acquisition data of awareness apparatus node i, UTi
For the time of the last mistake acquisition data of awareness apparatus node i, Δ T is the threshold value of setting；
According to time attribute tsi, tuiValue, be arranged SRiAnd URiWeight wsiAnd wui:
Wherein, x, y, z is obtained according to real data training fitting, usually default x=0.2, y=0.3, z=0.5.
Formula (3) makes system according to different time attributes, dynamically distributes different weights for SR and UR, to obtain more
Add reasonable trust value.Compare in terms of numerical values recited: x > y > z.Value gap between x, y, z is bigger, then for the last time just
Influence of true or mistake acquisition data the times to history weight is bigger.
The history weight HW of awareness apparatus node i is calculated as followsiAre as follows:
HWiValue range is [0,1].By formula (4) as it can be seen that working as SRi>>URi, i.e., the node correctly acquires data in history
When number acquires the number of data much larger than mistake, the historical reliability of the node is higher, HWiLevel off to 1.Work as SRi≤ URi,
When the number that i.e. node correctly acquires data in history is less than or equal to the number of mistake acquisition data, the history of the node is reliable
Minimum, the HW of propertyiIt is 0.
(2) the instant weight based on similarity generates
Method based on trust value is based on historical data, and the higher node of reliability is due to environmental factor in history
It influences to may cause the numerical value currently perceived and physical presence large error.It needs exist for considering each awareness apparatus node and its
The consistency of his awareness apparatus node.Therefore, the present embodiment method calculate the similarity between certain node and every other node it
With, and portray with this weight of node.The weight of node is bigger, shows that the node is supported by other more nodes, then should
The credibility that node correctly acquires data is bigger.
The specific data that similarity calculation between each awareness apparatus node can be acquired based on each device node, can also be with base
In the basic probability assignment value that it is obtained by subordinating degree function.But existed centainly based on the calculation method of basic probability assignment value
Problem.By taking temperature control clothes are engaged in as an example, it is assumed that current temperature value is 28 DEG C, subordinating degree function as shown in Figure 2, obtain m (height)=
0.75, m (low)=0, m (moderate)=0.25.So the semantic information of Current Temperatures is " temperature is high ".A device node if it exists
The temperature value of acquisition be 40 DEG C, seriously departing from actual temperature value, but its basic probability assignment value be m (height)=1, m (low)=
0, m (moderate)=0.Similarity calculation based on basic probability assignment value will lead to it and the device node of normal acquisition data has
There is higher similarity, but the device node does not have reference value in fact.To avoid such case, it is based in the present embodiment
The specific data of each device node acquisition calculate its similarity.
Assuming that service has n awareness apparatus node, it is denoted as N1, N2..., Nn, acquired data values are respectively D1, D2...,
Dn；Device node N1, N2Distance definition be its acquire data normalization after difference, calculated as the following formula:
I, j=1,2 ..., n and i ≠ j (5)
Wherein, dist (Ni,Nj)=Di-Dj, dist_min indicates in the dist () between all awareness apparatus nodes most
Small value, dist_max indicate the maximum value in the dist () between all awareness apparatus nodes；
Device node N1, N2Between distance it is bigger, show that its similarity is smaller.Device node N is defined as a result,1, N2It is similar
It spends as follows:
Sim(N1,N2)=1-Dist (N1,N2) (6)
Device node N can further be acquired by formula (6)1With the similarity between other n-1 node, and by formula (7) count
Calculate summation.
Sum () value of n sensing node is ranked up, it can be seen that the credibility of each awareness apparatus node.Sum
() is higher, shows that the credibility of the awareness apparatus node is higher.It is worth noting that, when the equipment section for participating in similarity calculation
When there are two points, even if the reliability of one of device node is very low, the result of calculating is still Sim (Ni, Nj)=Sim
(Ni, Nj).The confidence level that system calculates two awareness apparatus device nodes is identical.To avoid such case, the present embodiment
Middle awareness apparatus interstitial content is all larger than two.Actually in emerging system, it will also tend to provide many device nodes.
According to Sum () value of each node, instant weight (Current can be distributed for different awareness apparatus nodes
Weight, CW).
WhereinSum_MAX is the maximum value of Sum () in n evidence.Sum
(Ni) be i-th awareness apparatus node and remaining n-1 awareness apparatus node the sum of similarity.
(3) comprehensive weight
What the history weight based on trust value was measured is the degree of reliability of the node in all previous data acquisition, and is based on phase
Like degree instant weight measure be the node and other nodes in Current data acquisition the degree of consistency.Only consider history power
Value, will lead to cannot timely and effectively reflect that node mistake caused by environmental factor acquires information；Only consider instant weight, can put
The accidental sexual factor of overall situation etc., and the reliability attributes of node itself are had ignored, it is same undesirable.By comprehensively considering node
History weight and instant weight, so that the fusion accuracy of higher information is obtained, shown in fusion method such as formula (9).
Wi=λ HWi+(1-λ)·CWi (9)
Wherein, λ is distribution factor, and value interval is [0,1], and λ numerical value is bigger, value shadow of the history weight to final weight
Sound is bigger, on the contrary then smaller.Under different application scenarios, λ value is different.Environment relative quiescent, node stable working state
In the case of, the relatively large influence to improve history weight of λ value；Environment relative dynamic, the fluctuation of node working condition are biggish
In the case of, λ value should the relatively small influence to improve instant weight.
Or by taking the control service of smart home room temperature as an example, it is assumed that the service relies on seven awareness apparatus nodes, gives every
The number of one node successful acquisition data in history, the number of mistake acquisition data and the data acquired recently, such as table 2
It is shown.
2 awareness apparatus node data acquisition state of table
For all these awareness apparatus nodes, x=0.3, y=0.2, c=0.1 are set.Acquire the time interval of data
For t, Δ T=3t.By formula (9), after 20 data acquisitions in history, the history weight of device node is respectively HW1=
0.64, HW2=0.81, HW3=0.75, HW4=0.71, HW5=0.81, HW6=0.78, HW7=0.64.
According to the data of each device node acquisition for the first time, degree of membership as shown in Figure 2 obtains a series of evidences, such as table 3
It is shown.
The initial basic probability assignment of 3 sensing node of table
| Sensing node mark | It is high | It is moderate | It is low |
| 1 | 0.5 | 0.5 | 0 |
| 2 | 0.33 | 0.51 | 0.16 |
| 3 | 0.75 | 0.25 | 0 |
| 4 | 0.25 | 0.5 | 0.25 |
| 5 | 0.5 | 0.5 | 0 |
| 6 | 0.5 | 0.5 | 0 |
| 7 | 0.58 | 0.42 | 0 |
First time data are acquired, the sum of the similarity with remaining six node of each device node is calculated by formula (7),
Respectively Sum (N1)=4.67, Sum (N2)=4.17, Sum (N3)=3.33, Sum (N4)=1.67, Sum (N5)=4.67, Sum
(N6)=4.67, Sum (N7)=4.17.Further calculate the instant weight CW of each node1=1, CW2=0.83, CW3=0.55,
CW4=0, CW5=1, CW6=1, CW7=0.83.Distribution factor λ=0.5 is set here, then the synthesis weight obtained is respectively,
W1=0.82, W2=0.82, W3=0.65, W4=0.36, W5=0.91, W6=0.89, W7=0.74.
Formula is reassigned according to the basic probability function of formula (1), obtains new Basic probability assignment function, as shown in table 4.
Basic probability assignment after 4 awareness apparatus node regulation of table
| Device node mark | It is high | It is moderate | It is low | Θ |
| Know 1 | 041 | 041 | 0 | 018 |
| 2 | 0.27 | 0.42 | 0.13 | 0.18 |
| 3 | 0.49 | 0.16 | 0 | 0.35 |
| 4 | 0.09 | 0.18 | 0.09 | 0.64 |
| 5 | 0.455 | 0.455 | 0 | 0.09 |
| 6 | 0.445 | 0.445 | 0 | 0.11 |
| 7 | 0.43 | 0.31 | 0 | 0.26 |
The result of more device node collaborative perceptions is obtained according to D-S evidence fusion formula.M (height)=0.532, m (moderate)
=0.462, m (low)=0, m (Θ)=0.006.
By fusion results it is found that temperature at this time is height.After the acquisition of this time data, according to adopting for each device node
Collection result updates the value of its SR and UR.According to newest SR, UR and second of data acquired of each device node by formula (4)
And formula (8) recalculates its respective history weight and instant weight, further redistributes elementary probability letter by formula (1)
Number, the result of more device node collaborative perceptions is obtained through evidence fusion.It is successively carried out with this.
For those skilled in the art, it can be provided various corresponding according to above technical solution and design
Change and modification, and all these change and modification, should be construed as being included within the scope of protection of the claims of the present invention.
Claims (3)
1. based on the more equipment collaboration service construction methods of Internet of Things for improving D-S evidence, which comprises the steps of:
S1, semantic information in perception data is extracted, designs new semantic-based atomic service and using the language extracted
Adopted information is portrayed it and is output and input, and substitutes existing atomic service with new atomic service；
S2, the Basic probability assignment function that each hypothesis of identification framework in D-S evidence theory is generated using subordinating degree function；
S3, according to the significance level of each awareness apparatus node, assign different weights for different awareness apparatus nodes, and according to
This weight redistributes awareness apparatus node to the Basic probability assignment function of each hypothesis, as shown in formula (1):
Wherein, mi(Aj) it is the basic probability assignment value that i-th of awareness apparatus node before redistributing assumes j-th；m′i
(Aj) it is the basic probability assignment value that i-th of awareness apparatus node after redistributing assumes j-th；WiIt is set for i-th of perception
The weight of slave node；For the maximum value of weight in all n awareness apparatus nodes；Show
For the constraint condition for meeting basic probability assignment value and being 1, extra basic probability assignment value is assigned to Θ；
S4, the Basic probability assignment function finally redistributed as obtained in step S3 are obtained according to D-S evidence fusion formula
To final conclusion.
2. the method according to claim 1, wherein in step S3, each perception in Basic probability assignment function
The weight of device node has the process of dynamic corrections, detailed process are as follows:
(1) the history weight based on trust management is generated:
When the data of certain awareness apparatus node are consistent with the conclusion finally merged, it is believed that the awareness apparatus node is correctly
Data are acquired, SR is usediRecord awareness apparatus node i correctly acquires the ratio of the total times of collection of number Zhan of data；When certain
When the data of awareness apparatus node are runed counter to the conclusion finally merged, then it is assumed that the awareness apparatus node mistakenly acquires
Data use URiRecord the ratio of the total times of collection of number Zhan of awareness apparatus node i mistake acquisition data；
It carries out calculating time attribute t as the following formulasiAnd tui:
Wherein, TcurrentIt is current time, STiFor the time of the last correct acquisition data of awareness apparatus node i, UTiFor sense
Know the time of the last mistake acquisition data of device node i, Δ T is the threshold value of setting；
According to time attribute tsi, tuiValue, be arranged SRiAnd URiWeight wsiAnd wui:
X, y, z are artificial setting value, x > y > z；
The history weight HW of awareness apparatus node i is calculated as followsiAre as follows:
HWiValue range is [0,1]；
(2) the instant weight based on similarity is generated:
Note service has n awareness apparatus node, is denoted as N1, N2..., Nn, acquired data values are respectively D1, D2..., Dn；Sense
Know device node NiAnd NjDistance definition be its acquire data normalization after difference, calculated as the following formula:
Wherein, dist (Ni,Nj)=Di-Dj, dist_min indicates the minimum in the dist () between all awareness apparatus nodes
Value, dist_max indicate the maximum value in the dist () between all awareness apparatus nodes；
Awareness apparatus node NiAnd NjBetween distance it is bigger, show that its similarity is smaller；Awareness apparatus node N is defined as a result,iAnd Nj
Similarity is as follows:
Sim(Ni,Nj)=1-Dist (Ni,Nj), i, j=1,2 ..., n and i ≠ j (6)
The summation of each awareness apparatus node and the similarity of other n-1 awareness apparatus node is calculated by formula (7):
According to the summation of each awareness apparatus node and the similarity of other n-1 awareness apparatus node, set for different perception
Slave node distributes instant weight CW:
Wherein
(3) it is calculated as follows to obtain the amendment weight of each awareness apparatus node:
Wi=λ HWi+(1-λ)·CWi (9)
Wherein, λ is the distribution factor being manually set, and value interval is [0,1], and λ numerical value is bigger, and history weight is to final weight
Value influence is bigger, on the contrary then smaller.
3. the method according to claim 1, wherein x=0.2, y=0.3, z=0.5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910529870.0A CN110247975B (en) | 2019-06-19 | 2019-06-19 | Internet of things multi-device cooperative service construction method based on improved D-S evidence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910529870.0A CN110247975B (en) | 2019-06-19 | 2019-06-19 | Internet of things multi-device cooperative service construction method based on improved D-S evidence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110247975A true CN110247975A (en) | 2019-09-17 |
| CN110247975B CN110247975B (en) | 2022-01-07 |
Family
ID=67887967
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910529870.0A Active CN110247975B (en) | 2019-06-19 | 2019-06-19 | Internet of things multi-device cooperative service construction method based on improved D-S evidence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110247975B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112286507A (en) * | 2020-11-09 | 2021-01-29 | 无锡新思联信息技术有限公司 | RETE algorithm rule engine design method for autonomy of internet of things resources |
| CN116455800A (en) * | 2023-03-10 | 2023-07-18 | 华中科技大学 | Internet of things credibility coverage reliability assessment method based on D-S evidence theory |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105761001A (en) * | 2016-02-19 | 2016-07-13 | 国网浙江省电力公司金华供电公司 | Method for assessing states of distribution network equipment incorporated with multi-source information |
| CN105975984A (en) * | 2016-04-29 | 2016-09-28 | 吉林大学 | Evidence theory-based network quality evaluation method |
| CN106250935A (en) * | 2016-08-15 | 2016-12-21 | 广东石油化工学院 | The rotary machinery fault diagnosis method that genetic programming and weighted evidence theory merge |
| CN109241146A (en) * | 2018-09-21 | 2019-01-18 | 太原太工天宇教育科技有限公司 | Student's intelligence aid method and system under cluster environment |
| CN109447279A (en) * | 2018-09-18 | 2019-03-08 | 中国人民解放军海军航空大学 | Based on the multi-parameter fusion guided missile state evaluating method for improving evidence theory |
-
2019
- 2019-06-19 CN CN201910529870.0A patent/CN110247975B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105761001A (en) * | 2016-02-19 | 2016-07-13 | 国网浙江省电力公司金华供电公司 | Method for assessing states of distribution network equipment incorporated with multi-source information |
| CN105975984A (en) * | 2016-04-29 | 2016-09-28 | 吉林大学 | Evidence theory-based network quality evaluation method |
| CN106250935A (en) * | 2016-08-15 | 2016-12-21 | 广东石油化工学院 | The rotary machinery fault diagnosis method that genetic programming and weighted evidence theory merge |
| CN109447279A (en) * | 2018-09-18 | 2019-03-08 | 中国人民解放军海军航空大学 | Based on the multi-parameter fusion guided missile state evaluating method for improving evidence theory |
| CN109241146A (en) * | 2018-09-21 | 2019-01-18 | 太原太工天宇教育科技有限公司 | Student's intelligence aid method and system under cluster environment |
Non-Patent Citations (1)
| Title |
|---|
| ENDONG TONG等: "《Reasoning-Based Context-Aware Workflow Management in Wireless Sensor Network》", 《INTERNATIONAL CONFERENCE ON SERVICE-ORIENTED COMPUTING，ICSOC 2011》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112286507A (en) * | 2020-11-09 | 2021-01-29 | 无锡新思联信息技术有限公司 | RETE algorithm rule engine design method for autonomy of internet of things resources |
| CN116455800A (en) * | 2023-03-10 | 2023-07-18 | 华中科技大学 | Internet of things credibility coverage reliability assessment method based on D-S evidence theory |
| CN116455800B (en) * | 2023-03-10 | 2024-05-07 | 华中科技大学 | Internet of things credibility coverage reliability assessment method based on D-S evidence theory |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110247975B (en) | 2022-01-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| De Maio et al. | A knowledge-based framework for emergency DSS | |
| Evans et al. | Opinion formation on dynamic networks: identifying conditions for the emergence of partisan echo chambers | |
| CN104135381B (en) | Hierarchical service management and system | |
| Doval | Dynamically stable matching | |
| CN110247975A (en) | Based on the more equipment collaboration service construction methods of Internet of Things for improving D-S evidence | |
| CN109493052A (en) | Across catenary system contract and its transfer process based on the more subchains of main chain adduction row | |
| Trojahn et al. | Argumentation for reconciling agent ontologies | |
| Marzal et al. | A novel locality algorithm and peer-to-peer communication infrastructure for optimizing network performance in smart microgrids | |
| Valero et al. | AIoTES: Setting the principles for semantic interoperable and modern IoT-enabled reference architecture for Active and Healthy Ageing ecosystems | |
| Zhao et al. | A consensus-reaching method for large-scale group decision-making based on integrated trust–opinion similarity relationships | |
| CN104166581B (en) | A kind of virtual method towards increment manufacturing equipment | |
| Bui et al. | Beyond rationality: information design for supporting emergent groups in emergency response | |
| Bhat et al. | Intelligent scheduling in health care domain | |
| CN102769671A (en) | Competition and cooperation-based Web service relation network system | |
| Neyem et al. | Designing mobile shared workspaces for loosely coupled workgroups | |
| Camarinha-Matos et al. | TeleCARE: Collaborative virtual elderly support communities. | |
| CN109862049A (en) | A kind of resource share method, intelligent terminal and storage medium | |
| CN102929605A (en) | Cloud-computing-based open interface of data mining system | |
| De Florio et al. | How resilient are our societies? analyses, models, and preliminary results | |
| CN113380389A (en) | Intelligent regional medical first-aid method, system architecture, equipment and storage medium | |
| Rodríguez-Valenzuela et al. | Data fusion mechanism based on a service composition model for the internet of things | |
| CN117217856B (en) | Platform unified large market cooperation method based on block chain | |
| Swain et al. | FESC: functionally equivalent service composition | |
| Swain et al. | An ontology based approach for satisfying user requests in context aware settings | |
| Kadouche et al. | Disability centered approach in smart space management |
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 |