CN107196738A - A kind of compressed sensing method of data capture based on the dynamic norms of L_p - Google Patents
A kind of compressed sensing method of data capture based on the dynamic norms of L_p Download PDFInfo
- Publication number
- CN107196738A CN107196738A CN201710276348.7A CN201710276348A CN107196738A CN 107196738 A CN107196738 A CN 107196738A CN 201710276348 A CN201710276348 A CN 201710276348A CN 107196738 A CN107196738 A CN 107196738A
- Authority
- CN
- China
- Prior art keywords
- signal
- degree
- norm
- data
- rarefication
- 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.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L1/00—Arrangements for detecting or preventing errors in the information received
- H04L1/0001—Systems modifying transmission characteristics according to link quality, e.g. power backoff
- H04L1/0014—Systems modifying transmission characteristics according to link quality, e.g. power backoff by adapting the source coding
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M7/00—Conversion of a code where information is represented by a given sequence or number of digits to a code where the same, similar or subset of information is represented by a different sequence or number of digits
- H03M7/30—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/02—Resource partitioning among network components, e.g. reuse partitioning
- H04W16/10—Dynamic resource partitioning
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W84/00—Network topologies
- H04W84/18—Self-organising networks, e.g. ad-hoc networks or sensor networks
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Quality & Reliability (AREA)
- Theoretical Computer Science (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
On the premise of degree of rarefication is unknown, if conventional reconstruction algorithm will make restructing algorithm convergence rate slack-off the degree of rarefication overestimate of signal, the complexity of algorithm becomes big, if to the degree of rarefication underrating of signal, will occur missing inspection.The present invention proposes to be based in the case where signal degree of rarefication degree is unknownThe compressed sensing method of data capture of dynamic norm, this method is utilizedNorm reconstructs primary signal, and the reconstructed error based on signal dynamically updated using mean square error approximate gradient methodValue, further detects the degree of rarefication and supported collection of signal.The present invention can be high-quality in the case where information source degree of rarefication is unknown to restore primary signal according to the transmission information of information source, with extensive adaptability.
Description
Technical field
The present invention relates in the case where signal degree of rarefication is unknown, propose that one kind is based onThe compressed sensing number of dynamic norm
According to collection method, belong to the communication technology and data assembling sphere.
Background technology
Wireless sensor network is a kind of brand-new information acquisition platform, and the purpose is to collaboratively perceive, gather and handle
The information of perceptive object in network's coverage area, and observer is sent to, have been widely used for military field, environmental monitoring, doctor
Treat the association areas such as nursing.Traditional method of data capture is that all data for gathering sensor node are passed by leader cluster node
It is defeated to be handled to base station.However, in the data acquisition of sensor network, typically multiple sensor nodes are to same
Event carries out perception compression, wherein carrying substantial amounts of redundant data, occupies network communication bandwidth during transmission significantly, brings
Unnecessary energy consumption.
Compressed sensing (CompressedSensing, CS) proposition collected for wireless sensor network data open it is new
Thinking, it is breached the limitation of conventional Nyquist sampling thheorem and accurately weighed using less sample information there is provided a kind of
The method that structure goes out primary signal.The theory by lower dimensional space, low resolution, owes Nyquist using the compressibility of signal
The perception of high dimensional signal is realized in the irrelevant observation of sampled data.Compressive sensing theory is segmented into three processes：Sampling, survey
Amount and reconstruct.Sampling process：Sparse data is sampled；Measurement process：Measurement is compressed to the data that sampling is obtained, obtained
To observation vector；Restructuring procedure：Initial data is reduced by observation vector data.Compressive sensing theory eliminates real letter
Number processing traditional sampling theory between contradiction, had a wide range of applications in field of signal processing.
For convenience of explanation, two key concepts related to compressive sensing theory are first defined below：
Degree of rarefication：If signalCan only it be used on another number fieldIndividual vector carrys out linear expression（Much smaller than the length of signal
Degree）, then the degree of rarefication of this signal is exactly。
Measure number：According to compressive sensing theory, when signal has sparse property, signal only can be passed throughIndividual line
Property measurement recovers signal in high precision（Much smaller than the length of signal）, hereIt is known as measuring number.
These work before reviewing, the problem of being primarily present following two aspects：First, existing research is set up in number
According to degree of rarefication known under the conditions of, and this point is often unpractical in practical situations both；Second, in the sparse of data
Degree is in unknown situation, and existing reconstructing method effect is not ideal：Measurement number is too high to cause the waste of communications cost, and
Measurement number is very few and can not accurately recover data.
In summary, high-quality Data Collection is realized for how to combine compressive sensing theory, and by measuring square
Battle array dynamic optimization reduces the collecting amount of data, becomes one of key technology difficulty of urgent need to resolve.
The content of the invention
For defect of the prior art, the present invention is, in the case where signal degree of rarefication is unknown, to propose that one kind is based on
The compressed sensing method of data capture of dynamic norm, comprises the following steps：
Step 1: wireless sensor node is arranged around multiple information sources, by the sensor section nearest apart from each information source position
Point is defined as cluster head, and sub-clustering is carried out to the sensor node in the wireless sensor network centered on each cluster head；
Step 2: generation gaussian random matrix, the data in each cluster are gathered, it is determined that the measurement number needed；
Step 3: each sensor node in network sends the leader cluster node that is weighted to of its data, leader cluster node is according to must
To data characteristics using being based onThe restructing algorithm of norm carries out the reconstruct of signal, and the degree of rarefication of data is gone out according to residual detection；
Step 4: according to reconstructed error, dynamic renewalValue, continues reconstruction signal, until reaching required precision.
In summary, compared with the conventional method, the invention has the advantages that：
1）Pass through designDynamic regulation, can be achieved high-quality Data Collection；
2）Quality can be collected according to the information density and network bandwidth state dynamic adjusting data of information source, with extensive adaptation
Property.
Brief description of the drawings
Fig. 1 is the flow chart of the present invention.
Embodiment
The present invention proposes one kind and is based in the case where signal degree of rarefication is unknownThe compressed sensing data of dynamic norm
Collection method, with reference to Fig. 1, the specific implementation method of Data Collection is as follows：
Step 1: 300 wireless sensor nodes are arranged around 20 information sources, by the sensing apart from each information source position recently
Device node is defined as cluster head, and sub-clustering is carried out to the sensor node in the wireless sensor network centered on each cluster head；
Step 2: generation gaussian random matrix, the data in each cluster are gathered, it is determined that the measurement number needed；
Step 3: initialization, setFor empty set, for the primary signal of input, make initial sparse degree, it is high
This random matrix is；
Step 4: initializationUtilize() normCalculate：
；（1）
Step 5: setting cyclic variable；
Step 6: right, calculate：
（2）
（3）
HereIt is that a diagonal element isDiagonal matrix, i.e.,：
（4）
Obtain；
Step 7: calculating, that is, calculate, whereinRepresent calculation matrixRow to
Amount, wherein：
（5）
ChooseIn from big to smallIndividual element, by value correspondence whereinRow sequence numberConstitute set, order；
Step 8: calculation error, wherein：
（6）
And update sparse angle value, order；
If Step 9:More than given precision, dynamic renewalValue, is adjusted according to mean square errorBe worth size and, i.e.,：
,（7）
And six are gone to step, testing result is otherwise exported, ten are gone to step, hereIt is an invariant, it is near for controlling
The step-length declined like Gradient Iteration, wherein：
（8）
Approximate gradient is defined as：
；（9）
Step 10: according to measuring assemblyIn element descending, it is rightObservation collection sequence, be still designated as after sequence, fromProgressively reject back to frontElement, it is ensured thatIn it is remainingElement energy Perfect Reconstruction goes out signal；
Step 11: terminating.
Claims (4)
1. one kind is based onThe compressed sensing method of data capture of dynamic norm, it is characterised in that methods described includes following step
Suddenly：
Step 1: some wireless sensor nodes are arranged around multiple information sources, by the sensing apart from each information source position recently
Device node is defined as cluster head, and sub-clustering is carried out to the sensor node in the wireless sensor network centered on each cluster head；
Step 2: generation gaussian random matrix, the data in each cluster are gathered, it is determined that the measurement number needed；
Step 3: each sensor node in network sends the leader cluster node that is weighted to of its data, leader cluster node is according to obtaining
Data characteristics using being based onThe restructing algorithm of norm carries out the reconstruct of signal, and the degree of rarefication of data is gone out according to residual detection；
Step 4: according to reconstructed error, dynamic renewalValue, continues reconstruction signal, until reaching required precision.
2. one kind is based onThe compressed sensing method of data capture of dynamic norm, it is characterised in that each sensor section in network
Point sends the leader cluster node that is weighted to of its data, and leader cluster node is used according to obtained data characteristics and is based onThe reconstruct of norm
Algorithm carries out the reconstruct of signal, and the degree of rarefication of data is gone out according to residual detection, at least also comprise the following steps：
1）Initialization, set, for the primary signal of input, make initial sparse degree, calculation matrixFor
Gaussian random matrix；
2）Leader cluster node receives observation vector, and using being based onThe restructing algorithm of norm solves primary signalEstimation letter
Number；
3）Calculate, that is, calculate, whereinRepresentColumn vector,,
ChooseIn from big to smallIndividual element, element correspondenceRow sequence numberConstitute set, order；
4）Judge reconstructed error, dynamic renewalValue, and update sparse angle value, order；
5）Continue reconstruction signal, until reaching required precision, then export testing result, otherwise return to 3）, whereinRepresent the
Secondary primary signalEstimation,To detect obtained Sparse degree；
6）According to measuring assemblyThe descending of middle element, it is rightObservation collection sequence, be still designated as after sequence, fromFrom it is rear toward
It is preceding progressively to reject its element, until ensureingIn it is remainingElement energy Perfect Reconstruction goes out signal.
3. one kind is based onThe compressed sensing method of data capture of dynamic norm, it is characterised in that leader cluster node receives observation vector, and using being based onThe restructing algorithm of norm solves primary signalEstimation signal, at least also comprise the following steps：
1）Initialization, utilize() normCalculate
；
2）Cyclic variable is set,；
3）It is right, calculateWith, hereIt is one right
Angle element isDiagonal matrix, i.e.,。
4. one kind is based onThe compressed sensing method of data capture of dynamic norm, it is characterised in that according to reconstructed error, dynamic
UpdateValue, continues reconstruction signal, until reaching required precision, at least also comprises the following steps：
1）Calculation error；
2）IfMore than given precision, adjusted according to mean square errorBe worth size and, i.e.,： ,, and turn the step 2 in power 2）, hereIt is an invariant, for controlling the decline of approximate gradient iteration
Step-length, wherein, approximate gradient is defined as：, otherwise
Terminate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710276348.7A CN107196738A (en) | 2017-04-25 | 2017-04-25 | A kind of compressed sensing method of data capture based on the dynamic norms of L_p |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710276348.7A CN107196738A (en) | 2017-04-25 | 2017-04-25 | A kind of compressed sensing method of data capture based on the dynamic norms of L_p |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107196738A true CN107196738A (en) | 2017-09-22 |
Family
ID=59872651
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710276348.7A Pending CN107196738A (en) | 2017-04-25 | 2017-04-25 | A kind of compressed sensing method of data capture based on the dynamic norms of L_p |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107196738A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112534427A (en) * | 2018-08-07 | 2021-03-19 | 昕诺飞控股有限公司 | System and method for compressing sensor data using clustering and shape matching in edge nodes of a distributed computing network |
| CN114841370A (en) * | 2022-04-29 | 2022-08-02 | 杭州锘崴信息科技有限公司 | Processing method and device of federal learning model, electronic equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101841932A (en) * | 2010-05-10 | 2010-09-22 | 南京邮电大学 | Distributed compression sensing method based on dynamic clustering in wireless sensor network |
| CN104270156A (en) * | 2014-06-12 | 2015-01-07 | 湘潭大学 | Method for constructing tracking, reducing and compensating mechanism measurement matrix in compressed sensing |
| CN105050105A (en) * | 2015-08-21 | 2015-11-11 | 湘潭大学 | High-energy-efficiency low-information-density data collecting method based on compressed sensing |
-
2017
- 2017-04-25 CN CN201710276348.7A patent/CN107196738A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101841932A (en) * | 2010-05-10 | 2010-09-22 | 南京邮电大学 | Distributed compression sensing method based on dynamic clustering in wireless sensor network |
| CN104270156A (en) * | 2014-06-12 | 2015-01-07 | 湘潭大学 | Method for constructing tracking, reducing and compensating mechanism measurement matrix in compressed sensing |
| CN105050105A (en) * | 2015-08-21 | 2015-11-11 | 湘潭大学 | High-energy-efficiency low-information-density data collecting method based on compressed sensing |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112534427A (en) * | 2018-08-07 | 2021-03-19 | 昕诺飞控股有限公司 | System and method for compressing sensor data using clustering and shape matching in edge nodes of a distributed computing network |
| CN114841370A (en) * | 2022-04-29 | 2022-08-02 | 杭州锘崴信息科技有限公司 | Processing method and device of federal learning model, electronic equipment and storage medium |
| CN114841370B (en) * | 2022-04-29 | 2022-12-09 | 杭州锘崴信息科技有限公司 | Processing method and device of federal learning model, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107730451B (en) | Compressed sensing reconstruction method and system based on depth residual error network | |
| CN107784676B (en) | Compressed sensing measurement matrix optimization method and system based on automatic encoder network | |
| CN103295198B (en) | Based on redundant dictionary and the sparse non-convex compressed sensing image reconstructing method of structure | |
| CN105827250A (en) | Electric-energy quality data compression and reconstruction method based on self-adaptive dictionary learning | |
| CN102938649A (en) | Self-adaptive reconstruction and uncompressing method for power quality data based on compressive sensing theory | |
| CN103124179A (en) | Electric power system data reconfiguration decompressing method based on orthogonal matching pursuit | |
| CN103036574B (en) | A kind of self checking degree of rarefication Adaptive matching tracing algorithm based on compression sensing | |
| CN105050105A (en) | High-energy-efficiency low-information-density data collecting method based on compressed sensing | |
| CN108832934A (en) | A kind of two-dimensional quadrature match tracing optimization algorithm based on singular value decomposition | |
| CN109684314A (en) | A kind of wireless sensor network missing value estimation method based on space structure | |
| CN107196738A (en) | A kind of compressed sensing method of data capture based on the dynamic norms of L_p | |
| CN107561416A (en) | A kind of local discharge signal acquisition system and method based on compressed sensing | |
| CN103124180A (en) | Data reconfiguration and decompression method of power system based on projection pursuit | |
| CN112085062A (en) | Wavelet neural network-based abnormal energy consumption positioning method | |
| CN115452376A (en) | Bearing fault diagnosis method based on improved lightweight deep convolution neural network | |
| CN111093166A (en) | Compressed data collection system using sparse measurement matrix in internet of things | |
| CN114964782A (en) | Rolling bearing fault diagnosis acoustic emission signal processing method and system based on compressed sensing | |
| Zeng et al. | Image reconstruction of IoT based on parallel CNN | |
| CN110311685A (en) | Timing Bayes compression sampling and signal decompression reconstructing method and loss of data restoration methods | |
| CN106559670A (en) | A kind of improved piecemeal video compress perception algorithm | |
| CN104639398B (en) | Method and system based on the compression measurement test system failure | |
| CN106887026A (en) | Many compression of images and the method rebuild are realized based on compressed sensing and orthogonal modulation | |
| Zhai et al. | Data reconstructing algorithm in unreliable links based on matrix completion for heterogeneous wireless sensor networks | |
| CN109115350B (en) | Wavefront detection system based on compressed sensing | |
| CN108596851A (en) | A kind of compressed sensing image processing algorithm based on Wavelet transformation |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170922 |