CN103716824B - Radio sensing network message digging system and method for digging based on correlation rule - Google Patents
Radio sensing network message digging system and method for digging based on correlation rule Download PDFInfo
- Publication number
- CN103716824B CN103716824B CN201310743725.5A CN201310743725A CN103716824B CN 103716824 B CN103716824 B CN 103716824B CN 201310743725 A CN201310743725 A CN 201310743725A CN 103716824 B CN103716824 B CN 103716824B
- Authority
- CN
- China
- Prior art keywords
- sequence
- message
- network
- radio sensing
- candidate
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Mobile Radio Communication Systems (AREA)
Abstract
The invention discloses a kind of radio sensing network message digging system and method for digging based on correlation rule, the digging system of the present invention includes industrial computer, radio sensing network audiomonitor and wireless sensor node network, the radio sensing network audiomonitor and industrial computer network connection, the radio sensing network audiomonitor are located in wireless sensor node network；Message method for digging in the present invention is solved in distributed network, the complex interaction between each node and the mistake of non-frequent generation that produces, and these mistakes are difficult to position.The present invention finds sequence of question present in sequence of message using the improved data digging method based on correlation rule, helps find the mistake being difficult to find that in wireless sensor network debugging.Agreement for radio sensing network of the invention is improved, debugging has good reference function, is with a wide range of applications.
Description
Technical field
The invention belongs to technology of wireless sensing network field, more particularly to a kind of radio sensing network based on correlation rule
Message digging system and method for digging.
Background technology
With developing rapidly for embedding assembly technology, wireless communication technology, microprocessor technology and micro-sensor technologies
With increasingly mature, the micro sensing node with perception, computing capability and wireless communication ability starts to occur.By miniature biography
The radio sensing network that sense node is constituted in an ad-hoc fashion is aroused great concern, and this new network combines biography
Sensor technology, embedding assembly technology, distributed information processing and wireless communication technology, can collaboratively real-time monitoring,
Perceive and gather the environmental information in network arrangement region or monitor the information of foreign target, and these information are processed,
Detailed and accurate result is obtained, being ultimately transmitted needs the user of these information.
It is well known that due to the restriction of resource, exploitation radio sensing network application program is unlike developing PC application programs one
Sample prescription is just；Likewise, debugging radio sensing network application program also has many difficulties.In order to realize radio sensing network application journey
The debugging function of sequence, code level debugging is commonplace method used in present sensor network application program, code level
Debugging is referred to and can allow developer with sentence or code block configuration processor and while the tune of observable running state of programs
Examination.Also occur in that now the code level debugging acid for sensor network application program of many, these instruments lead to both at home and abroad
Cross, this provides phase in troubleshooting process
When big convenient effect.
For unfortunately for radio sensing network, the generation of application error is frequently not due to the coding of a single point
Mistake is produced, but caused by the incorrect interaction between different sensors node.Cause such interaction possible
The reason for include：1. the design defect of radio sensing network agreement (lost the process to some border incidents so that agreement is not
These events can be properly processed) 2. unpredictalbe artificial node alternative events.Meanwhile, the mistake of these interactivity is
Nonrepeatability, this is because repeating identical experimental procedure can not necessarily enter identical boundary condition.Simultaneously because
The work structuring of the multi-hop of Ad-hoc in radio sensing network, outside be difficult in observation radio sensing network interaction and knot
Structure, for such mistake, for wrong positioning and exclusion are relatively difficult with regard to shown under code level debugging.
Content of the invention
It is an object of the invention to overcoming the shortcoming and deficiency of prior art, there is provided a kind of based on the wireless of correlation rule
Sensing network message digging system.
Another object of the present invention is to, there is provided a kind of radio sensing network message excavation side based on correlation rule
Method, by this method, solve radio sensing network in look like on the surface between each independent sensor node due to
Reciprocal action and the Problem-Error that produces.
In order to reach above-mentioned first purpose, the present invention is employed the following technical solutions：
Based on the radio sensing network message digging system of correlation rule, supervise including industrial computer, radio sensing network
Listen device and wireless sensor node network, the radio sensing network audiomonitor and industrial computer network connection, the nothing
Line sensing network audiomonitor is located in wireless sensor node network；
The wireless sensor node network includes multiple wireless sensor nodes, wireless sensor node Web vector graphic without
Being communicated, wireless sensor node receives the letter for carrying out other sensor nodes in automatic network to line sensing network communications protocol
Breath, and other sensor nodes send return information and the network information in network；
Radio sensing network audiomonitor, for the order in monitoring wireless sensor node network, and by the life for listening to
Information transmission is made to industrial computer；
Industrial computer, for receiving return information that radio sensing network audiomonitor listens to and the return information being entered
Row analyzing and processing.
Preferably, the networking types of the wireless sensor node Web vector graphic include Star Network, zonal networks, tree-like
Network, loop network and mesh network.
Preferably, wireless sensor node includes connecing completing internal storage information for receiving the energy of read write line transmission
Receive and send transceiver module, for corresponding informance data write and storage memory module and antenna, transceiver module and
Memory module signal connects.
Preferably, the radio sensing network audiomonitor includes that the monitoring module and data of being capable of monitoring wireless signal are uploaded
Module, its upload data by ethernet network, and the monitoring scope of radio sensing network audiomonitor covers whole wireless sense network
Network.
In order to reach above-mentioned second purpose, the present invention is employed the following technical solutions：
Based on the method for digging of the radio sensing network message digging system of correlation rule, comprise the steps：
(1) data acquisition：Radio sensing network is made up of any wireless sensor node, by being located at wireless senser
Radio sensing network audiomonitor in meshed network is collected the communication information of each sensor node and sends industrial computer to；
(2) data prediction：Before message excavation is carried out, the communication message data first to collecting carry out pretreatment；
(3) message analysis：Pretreated message data is produced by candidate sequence generation, data sectional, Frequent episodes
Raw, sequence of question produces the generation that step obtains sequence of question.
Preferably, step (2) is specially：
2.1) unwanted response message and broadcasting packet are deleted, broadcasting packet is periodically packet, is a certain node pair
The transmission of the information of other nodes in network, other nodes are not given a response to which when receiving, response message and broadcasting packet
It is not the constituent of sequence of message；
2.2) message of consolidated network ID is chosen at, due to the message more than one network ID for listening to, but same
The communication message of network ID, the communication being only between complete each node；
2.3) redundant message is deleted, in communication process, it is possible to there is the phenomenon that node repeats to send same message,
Redundant message will be deleted for this, will synchronization frame sequence number value repeat message delete；
2.4) sequence of message, each message is represented with alphanumeric symbol, 2.2) and 2.3) is processed in above step
The data of gained are the data type message of each inter-node communication afterwards.
Preferably, step (3) is specially：
3.1) candidate sequence is produced, and generates candidate sequence using the candidate sequence generating mode based on sequence pattern；
3.2) data sectional, in order to accurately find out occurred sequence of message, each time during candidate's item count, with each time
Message of first element of sequence to having serialized is selected to carry out segmentation；
3.3) generation of Frequent episodes, in candidate sequence counting process is carried out, in order to ensure finding correct sequence, needs
Is deleted by interference sequence, error count is prevented for the sequence recognition that finds, in each segment if there is the message that finds
Sequence then search to quantity add 1, you can obtain candidate sequence c generation number of times σ (c), when σ (c) more than set value m
When, which is added in Frequent episodes；
3.4) sequence of question is produced, and after to each sequence count, will be counted income value and is compared, to find packet loss
Phenomenon, for a k-1 sequence, may be included by multiple k sequences, and for k-1 subsequences, frequent episode is counted as σ (ck-1),
The counting of the Frequent episodes comprising this subsequence, counts and is respectively σ1(ck)、σ2(ck)...σn(ck), by all comprising the sub- sequence
The Frequent episodes of row are counted and are added, and are obtainedAnd division is done, the situation of obtaining is classified as follows：
The all values for obtaining are not equal to 1 subsequence, and subsequence is collected in a matrix, are designated as PC (c, σ)={ ck-1|
p(ck-1) ≠ 1 }, wherein k values be more than 1, less than maximal term set value, to being not less than 2, but the sequence less than efficient items,
All it is compared, has checked whether packet loss phenomenon；
In the sequence of the different dimensions of the various presence that finds out, by arranging appropriate minimum m values, to less than m values
All item collections, are not repeat sequence, illustrate wireless sensor node network when residing for audiomonitor receives the sequence
Between section there may be communication issue, such sequence is collected in a matrix, SC (c, σ)={ c | S=σ (c) ＜ m } is designated as,
Its original message is returned to again by these sequences, analysing content determines whether sequence of question.
Preferably, the message format that radio sensing network is listened to includes：Message length, synchronization frame control domain, synchronization frame
Serial number, Personal Area Network identifier, destination address, source address, link layer message explanation, payload, check bit；The message of whole piece
It is to be represented by alphanumeric symbol, its final structure is：
(g1,g2,g3,g4,e1,e1,g1,g2,g3,g4,e1,g2,g3,g4,g1,e1,g2,g3,g4,g1,e1,g1,
G2, g3, g4 ...), wherein g, e represent different types of data message respectively, and 1,2,3,4 represents the communication between different nodes, and
Learn that message has 5 kinds of species g1, g2, g3, g4, e1, various item collections to count, find out possible sequence of question.
Preferably, candidate sequence production method is as follows：Merged by a pair frequent (k-1)-sequences, produce candidate's k- sequences
Row, for (k-1)-sequence S(1)And S(2), only when from S(1)In remove subsequence that first event obtain with from S(2)In remove
The subsequence that last event is obtained is identical, and as a result candidate is sequence S(1)With S(2)Last event connection, obtain
Result candidate sequence in, if (k-1)-sequence at least one right and wrong of candidate's k- sequences are frequently, then this candidate
K- sequences will be removed.
Preferably, during Frequent episodes are produced, the method for deleting interference sequence is as follows：
In the sequence of message S for needing to be processed, after message to be divided into each fragment with certain individual data unit,
To certain fragment in search certain candidate sequence when, certain data cell in the candidate sequence for reading individually is reported for next
Element in literary sequence, to this phenomenon, judges length of the acquired sequence in fragment, arranges parameter h, arbitrary excavation institute
The sequence length for obtaining must not exceed h, once more than h, then the sequence is deleted.
The present invention is had the following advantages relative to prior art and effect：
1st, sequence of question present in sequence of message is found using the message method for digging based on correlation rule in the present invention,
In due to distributed network, complex interaction between each node and the non-frequency mistake that produces, these mistakes are to be difficult to position
's.Using the improved message method for digging based on correlation rule in the present invention, problem present in sequence of message can be found
Sequence, helps find in wireless sensor network debugging and is difficult to the mistake found using code debugging, is having for code debugging
Power is supplemented.
2nd, present invention type not by sensor, the home control network communication protocol of sensor are affected, and be can be widely applied for respectively
Class radio sensing network, with wide applicability.
3rd, the present invention have simple and reliable, for the positioning precision of sequence of question high the characteristics of, can be quickly and accurately
Pinpoint the problems sequence.The invention for radio sensing network agreement improve, debugging have good reference function, with extensive
Using value.
Description of the drawings
Fig. 1 is the composition structural representation of message digging system of the present invention；
Fig. 2 is message data structural representation of the present invention；
Fig. 3 is message method for digging main flow chart of the present invention；
Fig. 4 is message analysis method flow diagram of the present invention；
Fig. 5 is that candidate sequence of the present invention produces schematic diagram；
Fig. 6 is faulty sequence analysis process figure of the present invention.
Specific embodiment
With reference to embodiment and accompanying drawing, the present invention is described in further detail, but embodiments of the present invention are not limited
In this.
Embodiment
As shown in figure 1, present embodiment discloses a kind of radio sensing network message digging system based on correlation rule, bag
Include industrial computer 1, radio sensing network audiomonitor 2 and wireless sensor node network 3；Radio sensing network audiomonitor 2
With 1 network connection of industrial computer, the radio sensing network audiomonitor 2 is in wireless sensor node network 3.
Wireless sensor node network：Including multiple wireless sensor nodes, wireless sensor node Web vector graphic is wireless
Being communicated, wireless sensor node receives the information for carrying out other sensor nodes in automatic network to sensing network communications protocol,
And other sensor nodes send return information and the network information in network；
Radio sensing network audiomonitor：Order in monitoring wireless sensor node network, and the order for listening to is believed
Breath passes to industrial computer；
Industrial computer：For receiving return information that radio sensing network audiomonitor listens to and the return information being entered
Row analyzing and processing.
As shown in Fig. 2 the message format that the radio sensing network is listened to includes：Message length, synchronization frame control domain, with
Step-frame serial number, Personal Area Network identifier, destination address, source address, link layer message explanation, payload, check bit.
In message format, message length occupies two bytes, for representing the total bytes shared by wall scroll message；Synchronization frame
Control domain accounts for two bytes, and first character is saved to represent whether the message needs to carry out ACK response, and second byte is used
To define the address format of link layer；Synchronous number of frames indicates the serial number of the message, accounts for a byte, and each node is sent out
The message for going out all is sorted in the way of serial number；Personal Area Network identifier is used for identifying a network, is answered for one or more
Network, for distinguishing different ZigBee-networks, the PAN-ID of all nodes is unique, network only one of which PAN-
ID, the message for listening to are likely to be from different networks, the communication between the same network node that we study, in this regard,
We filter out the message of same network according to PAN-ID values；Destination address represents the radio sensing network section for receiving message
Dot address, is the point of destination of message transmission；Source address represents the radio sensing network node address for sending message, is that message sends
Source point；Source address and destination address account for two bytes under short address, account for 8 bytes under long address；Link layer message is said
Bright account for a byte, the description of contents of the byte type of message, according to the binary system that its hexadecimal value is converted into, is sentenced
Break its type of message；Payload is used for the greatest length for representing message, accounts for a byte；Check bit uses CRC check, accounts for two
Individual byte, for checking the correctness of message data.
As shown in figure 3, present embodiment discloses a kind of message method for digging, comprises the following steps：
(1) data acquisition：Radio sensing network is made up of any wireless sensor node, by being located at wireless senser
Radio sensing network audiomonitor in meshed network is collected the communication information of each sensor node and sends industrial computer to, makes
The data for collecting in this way are exactly the communication message of node, and the form of message is as shown in Figure 2.
(2) data prediction：Before message excavation is carried out, the communication message data first to collecting carry out pretreatment
2.1) unwanted response message and broadcasting packet are deleted, broadcasting packet is periodically packet, is a certain node pair
The transmission of the information of other nodes in network, other nodes are not given a response to which when receiving, and are not therefore sequence of message
Constituent.If the response message in data is the response that makes for a upper message, when the control frame of a upper message is
When ' 61 ', the destination address of a upper message is responded to source address.Such response message is represented when receiving to a upper message
The response that makes, has no actual analysis meaning, and such message is deleted.
2.2) message of consolidated network ID is chosen at.Due to the message more than one network ID for listening to, but same
The communication message of network ID, the communication being only between complete each node.
2.3) redundant message is deleted, in communication process, it is possible to there is the phenomenon that node repeats to send same message,
Redundant message to be deleted for this, the message that synchronization frame sequence number value is repeated by we is deleted.
2.4) sequence of message.Each message is represented with alphanumeric symbol.Gained after above two-step pretreatment
Data are the data type message of each inter-node communication, and the message of whole piece is represented with alphanumeric symbol.Its structure final
Be expressed as (g1, g2, g3, g4, e1, e1, g1, g2, g3, g4, e1, g2, g3, g4, g1, e1, g2, g3, g4, g1, e1, g1, g2,
G3, g4 ...), wherein g, e represent different types of data message respectively.1,2,3,4 represents the communication between specific difference node,
And learning that message has 5 kinds of species g1, g2, g3, g4, e1, using the data digging method based on correlation rule to various item collection meters
Number, finds out possible sequence of question.
(3) as shown in figure 4, the message analysis of the present embodiment：To pretreated data by candidate sequence generation, data
The steps such as segmentation, Frequent episodes are produced, sequence of question generation obtain the generation of sequence of question.
3.1) candidate sequence is produced.Candidate using the candidate sequence generating mode next life cost method based on sequence pattern
Sequence.Embodiment as shown in figure 5, by be merged into frequent 3- sequences obtain candidate's 4- sequences.First candidate<(g1)
(g2)(g3)(g4)>By merging<(g1)(g2)(g3)>With<(g2)(g3)(g4)>Obtain；And<(g1)(g2)(g3)>With<
(g2)(g5)(g3)>Need not merge because remove first event of first sequence with remove second sequence last
Individual event does not produce identical subsequence.Any feasible candidate will not be so lost, is also avoided that and is produced the candidate for repeating
Sequence.
During candidate sequence is produced, in addition it is also necessary to carry out candidate's beta pruning to candidate sequence, for candidate<(g1)(g2)
(g3)(g4)>It is candidate's 4- sequence, by checking<(g1)(g2)(g4)>With<(g1)(g3)(g4)>Whether it is that 3- is frequent
Sequence, as they are not frequently, therefore can delete<(g1)(g2)(g3)(g4)>.After candidate's beta pruning, stay only
One candidate's 4- sequences are candidates<(g1)(g2)(g5)(g3)>.
3.2) data sectional.In order to accurately find out occurred sequence of message, each time during candidate's item count, with each time
Message of first element of sequence to having serialized is selected to carry out segmentation.For sequence of message (g1, g2, g3, g4, e1, e1,
G1, g2, g3, g4, e1, g2, g3, g4, g1, e1, g2, g3, g4, g1, e1, g1, g2, g3, g4), when candidate's sequence to be searched
It is classified as<(g1)(g2)(g3)(g4)>When, need to use<(g1)>Come first to carry out segmentation to message data, the message after segmentation
For (g1, g2, g3, g4, e1, e1), (g1, g2, g3, g4, e1, g2, g3, g4), (g1, e1, g2, g3, g4), (g1, e1), (g1,
g2,g3,g4).
3.3) generation of Frequent episodes.In candidate sequence counting process is carried out, in order to ensure finding correct sequence, I
The sequence recognition that finds to be deleted interference sequence, prevent error count.For the report in each segment if there is searching
Literary sequence then search to quantity add 1, you can obtain candidate sequence c generation number of times σ (c).When σ (c) is more than value m for setting
When, which is added in Frequent episodes.The method for deleting interference sequence is as follows：
In sequence (g1, g2, g3, g4, e1, g1, e1, g2, g3, g4, e1, g2, g3, g4, g1, e1), will reported with g1
After text is divided into each fragment, in the fragment for obtaining<g1,e1,g2,g3,g4,e1,g2,g3,g4,>In, may read<g1,
g2,g4,g3>In g4 be second g4 in fragment, and second g4 is the element in next sequence g2g3g4g1 actually,
To this phenomenon, we judge length of the acquired sequence in fragment, and we arrange parameter h, and (m combines sequence to be counted
Depending on the length of row), arbitrary sequence length for excavating gained must not exceed h, once more than h, then the sequence is deleted.Above-mentioned
In example, it is assumed that h is 8, it is 16 in fragment length to take the g1g2g3g4 that second g4 obtain, and more than 10, then deletes.But reading
<g1,g2,g3,g4>When, in order to accurately be counted, the result that different h are obtained to be analyzed, interference sequence is deleted.
3.4) as shown in fig. 6, the basic process of sequence of question generation is as follows：
After to each sequence count, we are compared income value is counted, to find packet loss phenomenon.For one
K-1 sequences, may be included by multiple k sequences.For k-1 subsequences, frequent episode is counted as σ (ck-1).Comprising this subsequence
The counting of Frequent episodes, counts and is respectively σ1(ck)、σ2(ck)...σn(ck), by all Frequent episodes meters comprising the subsequence
Number is added, and obtainsAnd division is done, the situation of obtaining is classified as follows：
The all values for obtaining are not equal to 1 subsequence, and subsequence is collected in a matrix, are designated as PC (c, σ)={ ck-1|
p(ck-1) ≠ 1 }, wherein k values are more than 1, less than maximal term set value.To being not less than 2, but the sequence less than efficient items,
All it is compared, has checked whether packet loss phenomenon.
In the sequence of the different dimensions of the various presence that finds out, we are arranged by test of many times, and setting is appropriate most
Little m values, to all item collections less than m values, are not repeat sequence, illustrate that wireless sensor node network is received in audiomonitor
Communication issue be there may be to the time period residing for the sequence, such sequence is collected in a matrix, be designated as SC (c, σ)=
C | S=σ (c) ＜ m }, then its original message is returned to by these sequences, analysing content determines whether sequence of question.
Above-described embodiment is the present invention preferably embodiment, but embodiments of the present invention are not by above-described embodiment
Limit, other any spirit and the changes, modification, replacement that is made under principle without departing from the present invention, combine, simplification,
Equivalent substitute mode is should be, is included within protection scope of the present invention.
Claims (9)
1. the radio sensing network message digging system based on correlation rule, it is characterised in that including industrial computer, wireless pass
Sense sniffer and wireless sensor node network, the radio sensing network audiomonitor are connected with industrial computer network
Connect, the radio sensing network audiomonitor is located in wireless sensor node network；
The wireless sensor node network includes that multiple wireless sensor nodes, wireless sensor node Web vector graphic are wirelessly passed
Feel network communication protocol to be communicated, wireless sensor node receives the information for carrying out other sensor nodes in automatic network, and
To in network, other sensor nodes send return information and the network information；
Radio sensing network audiomonitor, for the order in monitoring wireless sensor node network, and the order for listening to is believed
Breath passes to industrial computer；
Industrial computer, for receiving return information that radio sensing network audiomonitor listens to and the return information being carried out point
Analysis is processed；
The digging system carries out the method for data mining：
(1) data acquisition：Radio sensing network is made up of any wireless sensor node, by being located at wireless sensor node
Radio sensing network audiomonitor in network is collected the communication information of each sensor node and sends industrial computer to；
(2) data prediction：Before message excavation is carried out, the communication message data first to collecting carry out pretreatment；
(3) message analysis：Pretreated message data is produced, is asked by candidate sequence generation, data sectional, Frequent episodes
Topic sequence generation step obtains the generation of sequence of question；
Step (3) is specially：
3.1) candidate sequence is produced, and generates candidate sequence using the candidate sequence generating mode based on sequence pattern；
3.2) data sectional, in order to accurately find out occurred sequence of message, each time during candidate's item count, with each candidate's sequence
First element of row carries out segmentation to the message for having serialized；
3.3) generation of Frequent episodes, in candidate sequence counting process is carried out, in order to ensure finding correct sequence, it is right to need
The sequence recognition for finding, deletes interference sequence, prevents error count, for the sequence of message in each segment if there is searching
Then search to quantity add 1, you can obtain candidate sequence c generation number of times σ (c), when σ (c) more than set value m when,
Which is added in Frequent episodes；
3.4) sequence of question is produced, and after to each sequence count, will be counted income value and is compared, to find that packet loss is existing
As, for a k-1 sequence, may be included by multiple k sequences, for k-1 subsequences, frequent episode is counted as σ (ck-1), bag
The counting of the Frequent episodes containing this subsequence, counts and is respectively σ1(ck)、σ2(ck)...σn(ck), by all comprising the subsequence
Frequent episodes count be added, obtainAnd division is done, the situation of obtaining is classified as follows：
The all values for obtaining are not equal to 1 subsequence, and subsequence is collected in a matrix, are designated as PC (c, σ)={ ck-1|p
(ck-1) ≠ 1 }, wherein k values are more than 1, less than maximal term set value, to being not less than 2, but the sequence less than efficient items, all
It is compared, has checked whether packet loss phenomenon；
In the sequence of the different dimensions of the various presence that finds out, by arranging appropriate minimum m values, to owning less than m values
Item collection, is not repeat sequence, illustrates that wireless sensor node network receives the time period residing for the sequence in audiomonitor
Communication issue is there may be, such sequence is collected in a matrix, be designated as SC (c, σ)={ c | S=σ (c) ＜ m }, then by
These sequences return to its original message, and analysing content determines whether sequence of question.
2. radio sensing network message digging system according to claim 1 based on correlation rule, it is characterised in that described
The networking types of wireless sensor node Web vector graphic include Star Network, zonal networks, tree network, loop network and netted
Network.
3. radio sensing network message digging system according to claim 1 based on correlation rule, it is characterised in that wireless
Sensor node include for receive read write line transmission energy come complete internal storage information receive and send transceiver module,
Memory module and antenna for the write and storage of corresponding informance data, transceiver module and the connection of memory module signal.
4. radio sensing network message digging system according to claim 1 based on correlation rule, it is characterised in that described
Radio sensing network audiomonitor includes that the monitoring module for being capable of monitoring wireless signal and data upload module, and which passes through Ethernet net
Network uploads data, and the monitoring scope of radio sensing network audiomonitor covers whole radio sensing network.
5. according to any one of claim 1-4 based on correlation rule radio sensing network message digging system excavation
Method, it is characterised in that comprise the steps：
(1) data acquisition：Radio sensing network is made up of any wireless sensor node, by being located at wireless sensor node
Radio sensing network audiomonitor in network is collected the communication information of each sensor node and sends industrial computer to；
(2) data prediction：Before message excavation is carried out, the communication message data first to collecting carry out pretreatment；
(3) message analysis：Pretreated message data is produced, is asked by candidate sequence generation, data sectional, Frequent episodes
Topic sequence generation step obtains the generation of sequence of question；
Step (3) is specially：
3.1) candidate sequence is produced, and generates candidate sequence using the candidate sequence generating mode based on sequence pattern；
3.2) data sectional, in order to accurately find out occurred sequence of message, each time during candidate's item count, with each candidate's sequence
First element of row carries out segmentation to the message for having serialized；
3.3) generation of Frequent episodes, in candidate sequence counting process is carried out, in order to ensure finding correct sequence, it is right to need
The sequence recognition for finding, deletes interference sequence, prevents error count, for the sequence of message in each segment if there is searching
Then search to quantity add 1, you can obtain candidate sequence c generation number of times σ (c), when σ (c) more than set value m when,
Which is added in Frequent episodes；
3.4) sequence of question is produced, and after to each sequence count, will be counted income value and is compared, to find that packet loss is existing
As, for a k-1 sequence, may be included by multiple k sequences, for k-1 subsequences, frequent episode is counted as σ (ck-1), bag
The counting of the Frequent episodes containing this subsequence, counts and is respectively σ1(ck)、σ2(ck)...σn(ck), by all comprising the subsequence
Frequent episodes count be added, obtainAnd division is done, the situation of obtaining is classified as follows：
The all values for obtaining are not equal to 1 subsequence, and subsequence is collected in a matrix, are designated as PC (c, σ)={ ck-1|p
(ck-1) ≠ 1 }, wherein k values are more than 1, less than maximal term set value, to being not less than 2, but the sequence less than efficient items, all
It is compared, has checked whether packet loss phenomenon；
In the sequence of the different dimensions of the various presence that finds out, by arranging appropriate minimum m values, to owning less than m values
Item collection, is not repeat sequence, illustrates that wireless sensor node network receives the time period residing for the sequence in audiomonitor
Communication issue is there may be, such sequence is collected in a matrix, be designated as SC (c, σ)={ c | S=σ (c) ＜ m }, then by
These sequences return to its original message, and analysing content determines whether sequence of question.
6. method for digging according to claim 5 based on the radio sensing network message digging system of correlation rule, its are special
Levy and be, step (2) is specially：
2.1) unwanted response message and broadcasting packet is deleted, broadcasting packet is periodically packet, is a certain node to network
In other nodes information transmission, other nodes are not given a response to which when receiving, and response message and broadcasting packet are not
The constituent of sequence of message；
2.2) message of consolidated network ID is chosen at, due to the message more than one network ID for listening to, but in same network
The communication message of ID, the communication being only between complete each node；
2.3) redundant message is deleted, in communication process, it is possible to there is the phenomenon that node repeats to send same message, be this
Redundant message to be deleted, will synchronization frame sequence number value repeat message delete；
2.4) sequence of message, each message is represented with alphanumeric symbol, in step 2.2) and 2.3) process after gained
Data be each inter-node communication data type message.
7. method for digging according to claim 5 based on the radio sensing network message digging system of correlation rule, its are special
Levy and be, the message format that radio sensing network is listened to includes：Message length, synchronization frame control domain, synchronous number of frames, individual
Domain net identifier, destination address, source address, link layer message explanation, payload, check bit；The message of whole piece is by word
Representing, its final structure is alphameric character：(g1,g2,g3,g4,e1,e1,g1,g2,g3,g4,e1,g2,g3,g4,g1,
E1, g2, g3, g4, g1, e1, g1, g2, g3, g4 ...), wherein g, e represent different types of data message, 1,2,3,4 generation respectively
Communication between the specific difference node of table, and learn that message has 5 kinds of species g1, g2, g3, g4, e1, various item collections to count, find out
Possible sequence of question.
8. method for digging according to claim 5 based on the radio sensing network message digging system of correlation rule, its are special
Levy and be, candidate sequence production method is as follows：Merged by a pair frequent (k-1)-sequences, produce candidate's k- sequences, for (k-
1)-sequence S(1)And S(2), only when from S(1)In remove subsequence that first event obtain with from S(2)In remove last thing
The subsequence that part is obtained is identical, and as a result candidate is sequence S(1)With S(2)Last event connection, the result candidate for obtaining
In sequence, if (k-1)-sequence at least one right and wrong of candidate's k- sequences are frequently, then this candidate k- sequence will be by
Remove.
9. method for digging according to claim 5 based on the radio sensing network message digging system of correlation rule, its are special
Levy and be, during Frequent episodes are produced, the method for deleting interference sequence is as follows：
In the sequence of message S for needing to be processed, after message to be divided into each fragment with certain individual data unit, obtained
When searching certain candidate sequence in certain fragment, certain data cell in the candidate sequence for reading is next individually message sequence
Element in row, to this phenomenon, judges length of the acquired sequence in fragment, arranges parameter h, arbitrary excavation gained
Sequence length must not exceed h, once more than h, then the sequence is deleted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310743725.5A CN103716824B (en) | 2013-12-27 | 2013-12-27 | Radio sensing network message digging system and method for digging based on correlation rule |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310743725.5A CN103716824B (en) | 2013-12-27 | 2013-12-27 | Radio sensing network message digging system and method for digging based on correlation rule |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103716824A CN103716824A (en) | 2014-04-09 |
| CN103716824B true CN103716824B (en) | 2017-03-15 |
Family
ID=50409289
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310743725.5A Active CN103716824B (en) | 2013-12-27 | 2013-12-27 | Radio sensing network message digging system and method for digging based on correlation rule |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103716824B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113259265B (en) * | 2021-04-30 | 2023-07-25 | 阿里云计算有限公司 | Message processing method and device, electronic equipment and storage medium |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102006198A (en) * | 2010-12-16 | 2011-04-06 | 中国电子科技集团公司第三十研究所 | Method and device for acquiring network failure association rule |
-
2013
- 2013-12-27 CN CN201310743725.5A patent/CN103716824B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102006198A (en) * | 2010-12-16 | 2011-04-06 | 中国电子科技集团公司第三十研究所 | Method and device for acquiring network failure association rule |
Non-Patent Citations (3)
| Title |
|---|
| A Novel Association Rule Mining Mechanism in Wireless Sensor Networks;Anjan Das;《IEEE，NCETACS,2012 3rd National Conference on》;20120331;正文第274-276页 * |
| 关联规则数据挖掘方法的研究;徐辉增;《科学技术与工程》;20120131;第12卷(第1期);正文第60-63页 * |
| 基于实时数据关联分析的煤矿安全预警系统设计;侯燕，乔蕊;《计算机测量与控制》;20131025;第21卷(第10期);正文第2826-2828页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103716824A (en) | 2014-04-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Sahoo | Faulty Node Detection in Wireless Sensor Networks Using Cluster | |
| Khatib et al. | Self-healing in mobile networks with big data | |
| Miao et al. | Agnostic diagnosis: Discovering silent failures in wireless sensor networks | |
| CN106250424B (en) | A kind of searching method, the apparatus and system of log context | |
| Keller et al. | How was your journey? Uncovering routing dynamics in deployed sensor networks with multi-hop network tomography | |
| US20150066881A1 (en) | Compressing trace data | |
| Kumar et al. | 2scent: An efficient algorithm to enumerate all simple temporal cycles | |
| CN112118174B (en) | Software defined data gateway | |
| Debouk et al. | On the effect of communication delays in failure diagnosis of decentralized discrete event systems | |
| CN104615748B (en) | Internet of Things Web event-handling methods based on Watir | |
| CN115858796A (en) | Fault knowledge graph construction method and device | |
| CN114153980A (en) | Knowledge graph construction method and device, inspection method and storage medium | |
| CN109993390B (en) | Alarm association and order dispatching optimization method, device, equipment and medium | |
| CN112444858A (en) | Edge cloud cooperation-based earthquake real-time monitoring AI (Artificial intelligence) computing method, device and medium | |
| Mohapatra et al. | Artificial immune system based fault diagnosis in large wireless sensor network topology | |
| Mo et al. | Network simplification and k-terminal reliability evaluation of sensor-cloud systems | |
| CN103716824B (en) | Radio sensing network message digging system and method for digging based on correlation rule | |
| CN102883359A (en) | Method, device and system for measuring key nodes of wireless sensor network | |
| CN112699281B (en) | Alarm event rule mining method and system based on gspan algorithm | |
| CN104750812A (en) | Automatic data collecting method based on webpage label analysis | |
| CN107454089A (en) | A kind of network safety situation diagnostic method based on multinode relevance | |
| CN107517474B (en) | Network analysis optimization method and device | |
| CN105429793A (en) | Communication network weighted link importance degree assessment method | |
| CN106550387B (en) | A kind of wireless sensor network routing layer QoS evaluating method | |
| CN106296420A (en) | A kind of community discovery method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |