CN106559744A - It is a kind of based on sparse identification and the body area network power-economizing method of compressed sensing - Google Patents
It is a kind of based on sparse identification and the body area network power-economizing method of compressed sensing Download PDFInfo
- Publication number
- CN106559744A CN106559744A CN201611044550.9A CN201611044550A CN106559744A CN 106559744 A CN106559744 A CN 106559744A CN 201611044550 A CN201611044550 A CN 201611044550A CN 106559744 A CN106559744 A CN 106559744A
- Authority
- CN
- China
- Prior art keywords
- signal
- node
- sparse
- sample
- area network
- 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
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/80—Services using short range communication, e.g. near-field communication [NFC], radio-frequency identification [RFID] or low energy communication
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/30—Services specially adapted for particular environments, situations or purposes
- H04W4/38—Services specially adapted for particular environments, situations or purposes for collecting sensor information
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W52/00—Power management, e.g. TPC [Transmission Power Control], power saving or power classes
- H04W52/02—Power saving arrangements
- H04W52/0209—Power saving arrangements in terminal devices
- H04W52/0225—Power saving arrangements in terminal devices using monitoring of external events, e.g. the presence of a signal
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The invention discloses a kind of based on the sparse body area network power-economizing method recognized with compressed sensing, concrete grammar is：Sample data in data with existing storehouse is carried out normally with anomaly classification and number and be built into complete dictionary D，Then dimension-reduction treatment is done to test sample y and dictionary D by subalpine forests random matrix，The rarefaction representation that compressed signal is calculated with rarefaction representation algorithm is sparse then，By judging whether test sample is useful signal，It is judged as that invalid signals are directly filtered，Node restarts sampling，It is judged as that useful signal is then checked with database flags by residual error identification signal calculated least residual position i，If normal signal，Node is directly entered resting state，If abnormal signal，Compressed signal is transferred to the reconstruct of center base station node by node，Reconstruction signal y ' is sent to into remote medical center diagnostic analysiss and to giveing first-aid measures finally，Realize body area network node is very important performance indications. dormancy energy-conservation and data compression transmission energy-conservation，Reduce body area network energy consumption.
Description
Technical field
The present invention relates to body area network technical field, and in particular to a kind of based on the sparse body area network section recognized with compressed sensing
Can method.
Background technology
Wireless body area network is centered on human body, for the purpose of gathering the various physiological parameters of human body, by being distributed in people's body surface
The cordless communication network of the sensor and personal data collecting and processing unit composition of face or implantation inside of human body.China is that population is big
State, people require to improve constantly to health medical treatment, but medical resource is limited, it is impossible to which everyone has a doctor, with people
In terms of mouth aging degree aggravation, cardiovascular, disease is increasing, the unexpected generation of many diseases and much potential disease
Do not find in time, miss optimal therapeutic time.With the fast development of medical services, body area network is widely used to remotely
The aspect such as medical diagnosiss, diseases monitoring and prevention, family's nurse.Body area network is by body surface or internal placement sensor section
Point monitoring human body health, the such as monitoring such as pulse, blood pressure, electrocardio, brain electricity, insulin concentration, to reach to the timely of disease
Prevention, find, and treatment.Body area network is generally powered using battery, is difficult frequently replacing, particularly with the sensing section of implantation human body
Point, usually requires that the working time that can maintain the several years, and body area network monitors the network of human body physiological data as a kind of long-time, must
There need be good energy supply, therefore study body area network energy-saving scheme and there is extremely important practical significance very much.Reduce body area network
The approach of power consumption is mainly by improving communication technology and reducing sampled data output.Improve communication technology and mainly design efficient
Energy-saving routing protocol, make node when no data sends in a dormant state, it is to avoid node idle listening, reduce sampled data master
If improving data transmission bauds reduces volume of transmitted data simultaneously.
Compressed sensing (Compressed Sensing, CS) is theoretical to provide new thinking for signal sampling, traditional how Kui
This special sampling thheorem requires twice of the sample frequency not less than highest frequency in analogue signal frequency spectrum, could distortionless recovery mould
Intend signal, as long as and CS theory signal have it is openness or on a certain transform domain have it is openness, it is possible to one with become
Change the incoherent observing matrix of base conversion gained high dimensional signal to be projected on one lower dimensional space, it is then excellent by solving one
Change problem just can reconstruct original signal with high probability from a small amount of projection, and enough letters of reconstruction signal are contained in projection
Breath.It is used in CS theoretical in body area network node is very important performance indications. data acquisition, while the accuracy of data transmission procedure is met, can reduces
The collection capacity of data, reduces body area network volume of transmitted data, reduces network energy consumption.
Sparse recognition algorithm (Sparse Representation-based Classication, SRC) initially come from
CS is theoretical, is widely used in machine learning, machine vision and area of pattern recognition, and SRC algorithm main thoughts are：With all
Training sample constituted complete dictionary, test sample was projected in dictionary, as test sample is in the expression of excessively complete dictionary
It is only larger with the coefficient for having corresponding sample class, so expression of the test sample on excessively complete dictionary is typically sparse, from
And sample can be classified.With scientific and technological progress, in medical apparatus and instruments detection, counted by the measurement to all kinds of patients,
Establish all kinds of physiological signal data storehouses so that SRC is used widely in association areas such as body area network identifications, such as stomach disease
Sick detection, electrocardio compression sampling, electroencephalogram etc..In body area network communications portion, the transmission of signal and more, the body that receives consumed energy
The most normal signals for plateau of signal of domain net node collection, need not be sent to center base station.Therefore sensor is adopted
The signal of collection had both wasted the limited energy of node and had wasted the medical resource of preciousness without transmission is distinguished.SRC is stopped for body area network
Dormancy energy-conservation, by all kinds physiological signal recorded in medical statisticses physiological data storehouse, is given birth to by extracting these different types
Reason data were set up complete dictionary, signal were identified by SRC as training sample, so as to stop normal signal transmission,
Make collection normal signal node in a dormant state, reduce body area network communication energy consumption.
Body area network node is very important performance indications. energy is extremely limited, is difficult frequently replacing, and the usual hardware configuration of center base station node is high, and
Battery is readily replaceable, has many scholar's research body area network power-economizing methods at present, most for one-sided research body area network dormancy dispatching
Strategy or data compression energy-conservation.
The content of the invention
In view of the drawbacks described above of prior art, the technical problem to be solved is to provide a kind of based on sparse knowledge
Body area network power-economizing method not with compressed sensing, the method consider two kinds of energy-saving squares of body area network dormancy strategy and data compression
SRC and CS theories are provided commonly for body area network energy saving research by case, by means of existing physiological data storehouse, construction physiological signal training
Dictionary, is identified to the signal that body area network node is very important performance indications. is gathered by SRC, sends the stopping of normal signal, extends node dormancy
Time, so as to reduce body area network communications portion energy, meanwhile, the abnormal signal of compression transmission identification, in center base station node root
Compressed signal is reconstructed according to CS is theoretical, so as to reduce body area network volume of transmitted data, reduce tcp data segment energy.
For achieving the above object, the present invention provides a kind of based on the sparse body area network power-economizing method recognized with compressed sensing,
Comprise the steps：
S1, extraction data base's physiological signal sample built complete dictionary D：According to the physiological signal type of node collection,
Normal and anomaly classification labelling are carried out to data base's sample of the collection signal of node, training sample was configured to into complete word
Allusion quotation D；
S2, the original test sample y of compression and excessively complete dictionary D：Subalpine forests random matrix dimensionality reduction is adopted to original physiologic signal y
CompressionEqually carry out dimensionality reduction compression to obtain compressing dictionary to excessively complete dictionary D：
S3, seek rarefaction representation coefficientBy compressed signalAnd dictionaryConsider influence of noise in reality, add noise jamming
Error term ε, according to sparse representation theoryChanging solving model is
Coefficient table is obtained up to sparse
S4, filter invalid signals：By sparse coefficientSubstitute into sparse concentration degree indexAnd obtainValue,
RelativelyWith the size of given threshold τ, whether discriminating test sample is effective sample signal, if invalid sample letter
Number, directly filter, into step S1, node restarts physiological data of sampling, and is useful signal, then into next step；
Wherein, it is describedIt is expressed as：The threshold tau
∈[0,1]；N is classification sum, i=1,2 ..., n during node Sample Establishing；δiX () representsIn i-th position be
Number；
Whether S5, identification test sample are normal signal：Normal signal is identified using residual error classification function：Calculate least residualLeast residual position i is obtained, check and correction has been numbered
Determine the effective sample generic, whether identification test signal y is normal signal, if the test signal normal signal,
Switching collection normal signal node is resting state, if it is abnormal signal that the test signal is detected result, by compression letter
NumberIt is sent to the reconstruct of center base station node；
S6, construction sensing matrix Acs：Selection makes node collection signal in the sparse sparse base Ψ of orthogonal intersection space, and described
Subalpine forests are randomIt is configured to compressed sensing sensing matrix
S7, calculating reconstruct sparse coefficient θ '：Compressed signal is received according to center base station nodeWith sensing matrix Acs, use
Compressed sensing signal reconstruction algorithm, calculates reconstruct sparse coefficient θ ',
S8, recovery and rebuilding signal y '：According to sparse coefficient θ ' and orthogonal basiss Ψ has been reconstructed, calculate reconstruction signal y '=
ψθ′；
S9, diagnostic signal：Reconstruction signal is sent to remote medical center carries out diagnostic analysiss, and in center base station node
Early warning.
Using such scheme, collection normal signal node can be made to stop sending data and entering resting state, extend section
Point dormancy time, acquisition abnormity physiological signal Node compression transmission signal reduce body area network volume of transmitted data, and in base station reconfiguration
Signal, so as to realize reducing body area network energy consumption.
To better implement the present invention, may further be：
It is in the S4, describedCompare with threshold tau is set, ifThen test sample signal is
Invalid sample signal, invalid signals are directly filtered, and node resurveys data, ifThen discriminating test sample is
Effective sample signal, initially enters next step to test sample signal identification.The technical problem that the program is directed to is：Node is at certain
Jitter is gathered in the case of a little, it is impossible to reach diagnostic analysiss standard, and some are not in the training sample of physiological data storehouse
Invalid signals are likely to be sent, and cause unnecessary energy consumption and make remote to prevent these invalid signals to be sent to center base station node
Journey medical centre causes mistaken diagnosis or warning, before to Modulation recognition, using sparse concentration degree index S CI as criterion.
By sparse coefficientSubstitute intoMiddle calculatingValue, is compared with the threshold tau for setting, will's
Invalid sample target signal filter.
In step S5, least residual is calculated with residual error function, obtain the affiliated label i of test sample, and
It is abnormal signal or normal signal that test sample is recognized by the label classified in step S1, if normal signal,
Node stops sending data, and into resting state, if abnormal signal, node directly sends compression verification to center base station node
Signal Measured value in as described S7 in the theoretical reconstruct test signals of CS.The technical problem that the program is directed to is：According to
Least residual object function identification test sample is abnormal signal or normal signal, and by normal signal node dormancy, is extended
Abnormal signal compression is sent to the reconstruct of center base station node by node dormancy time, energy efficient, reduces volume of transmitted data, section
About data transfer energy consumption.
In step S6, center base station node builds sensing matrix AcsBy institute in the sparse orthogonal basiss Ψ of signal and node
State the subalpine forests random matrix in step S2Constitute, theoretical according to CS, to enable signals to reconstruct, calculation matrix needs to meet
Limited isometry RIP, subalpine forests random matrixCalculation matrix during as CS is theoretical, compresses to test signal, and subalpine forests are random
MatrixStructure is very sparse and has been demonstrated to meet limited isometry RIP with high probability.The program mainly for technology
Problem is：The sparse identification of signal is carried out on node, and signal reconstruction is carried out on center base station node, sub- to test signal dimensionality reduction
Gaussian random matrixSubalpine forests random matrix is reconstructed with center base station nodeIt is same matrix, subalpine forests random matrixBy
It is very sparse in its structure so that the complexity that projection memory space and projection are calculated is substantially reduced, and is suitable for storing and is calculated
Resource-constrained body area network node is very important performance indications., can be set in center base station node and collection signal node respectively in advance, and subalpine forests are thrown
Shadow matrixRaising sparse signal representation calculating speed can be utilized at node, also compression transmission original test signal, together
When subalpine forests random matrixMeet limited isometry RIP and can enable signals to reconstruct, signal is enable in center base station node weight
Structure.
In step S7, the expression formula of reconstruct sparse coefficient θ ' is calculatedMiddle measured valueAs described step
It is identified as after abnormal signal, being sent directly to center base station by step S5 through dimensionality reduction test signal, test sample in rapid S2
Node.The technical problem that the program is directed to is：Excessively complete dictionary in step S3It is not to fix dictionary, through training, word
Allusion quotation and rarefaction representation coefficient can be constantly updated, and dictionary data amount is larger, and all data of dictionary are transmitted from node to base station can consume
Many energy, node can not make center base station to signal reconstruction by transmitting training dictionary, need to select suitable orthogonal basiss Ψ weights
Structure signal, subalpine forests random matrixThe matrix Φ of compositionL×mDimensionality reduction is carried out to each sample signal in excessively complete dictionary D,
Sparse signal representation calculating speed can be improved, the signal after subalpine forests random matrix dimensionality reductionIt is transmitted directly to center base station
Node is reconstructed, and as the measured value of CS reconstruction signals, had both been improve the sparse recognition speed of signal, and had been also reduced body area network signal number
According to transmission quantity, so as to reduce body area network energy consumption.
The invention has the beneficial effects as follows：Physiological signal is gathered according to body area network node is very important performance indications. and is usually plateau normal signal,
Body area network node is very important performance indications. finite energy, sends normal signal and wastes node energy, and the usual hardware configuration of center base station node is high, and
Battery is readily replaceable, considers body area network dormancy strategy and data compressing energy-saving scheme, by SRC algorithms and CS theoretical common use
In body area network energy saving research, by means of existing physiological data storehouse, physiological signal training dictionary is constructed, by CS theories to test specimens
The projection compression of this signal, carries out normal or anomalous identification by SRC, sends the stopping of normal signal, extend to compressed signal
The node dormancy time, make the abnormal signal compression transmission of node collection give center base station node, reduce body area network transmitted data amount,
And according to CS restructing algorithms in base station reconfiguration signal, so as to realize reducing body area network energy consumption.
Description of the drawings
Fig. 1 is work structuring figure of the present invention；
Fig. 2 is flow chart of the present invention；
Fig. 3 is particular flow sheet of the present invention.
Specific embodiment
The invention will be further described with specific embodiment below in conjunction with the accompanying drawings, as shown in Figure 1, Figure 2 and Figure 3, a kind of
Based on sparse identification and the body area network power-economizing method of compressed sensing, concretely comprise the following steps,
Step 1：Built complete dictionary
Dictionary is configured with two ways, and one kind is that existing training sample is directly constructed dictionary by mathematical model, this
The mode of kind is it cannot be guaranteed that signal is sparse enough；It is for second by learning to training sample obtaining the word for adapting to signal characteristic
Allusion quotation, constantly updates rarefaction representation coefficient and dictionary atom by dictionary learning, finds the new dictionary that signal can be made most sparse, so as to
Discrimination and computation rate are improved, the excessively complete dictionary D of training sample construction, the data volume of dictionary are big, are to reduce dictionary data
Amount, improves calculating speed, is compressed dimensionality reduction to signal using random matrix, and accidental projection has that computation complexity is low, distance
High dimensional data can be projected to lower dimensional space using accidental projection matrix by retentivity advantage, realize that Data Dimensionality Reduction compresses, it is assumed that body
Domain net interior joint collection n dimension high dimensional data x, can project to low-dimensional of a m dimension by m × n accidental projection matrix Φ
Space (m<<N), it is expressed as：
In formula (1)The projection of initial data x is gathered for node.
In CS is theoretical, random matrix is generally chosen as the calculation matrix of CS, the node collection life of random matrix compression
Reason data are the measured value that compressed sensing reconstructs part, therefore the selection of random matrix is most important with physiological data reconstruct.
In order to be able to reconstruct n dimensional signals from m dimension measurement signals, calculation matrix Φ answers meet the constraint isometry condition to Cand é s et al.
(Restricted Isometry Property, RIP) condition.That is matrix Φ meets s rank RIP properties, there is constant δs∈(0,
1), following formula sets up：
In formulaAnd | Λ |≤s, ΦΛThe size constituted for the related column indicated by Λ is indexed in Φ is s
The submatrix of × | Λ |, s be signal degree of rarefication, δsFor RIP constants.
Gaussian random matrix, Bernoulli Jacob's random matrix, subalpine forests random matrix are proved to meet RIP bars with high probability
Part, wherein subalpine forests random matrix are very sparse due to its structure so that the complexity that projection memory space and projection are calculated is big
It is big to reduce, it is suitable for the body area network node is very important performance indications. for storing and computing resource is limited.
Body area network interior nodes are distributed in human body different parts, measure different types of physiological data respectively, node collection
Each type physiological data has corresponding data sample and signal Analysis standard in data base, first by physiological data storehouse
Each sample carry out the classification and marking of normal and abnormal signal, and extract in data base sample number in each type data
According to as training sample, it is assumed that the physiological data species of certain node test includes w types, and this w kind is selected from data base
Type physiological signal di, diThe i-th type physiological data in data base is represented, k training sample in each type, is included： K-th training sample in the i-th type is represented, if each training sample dimension is m,
aiki∈Rm, and this type w kind physiological signal was configured to into complete dictionary D：
D ∈ R in formulam×n, n=n1+n2+…+nw。
Step 2：Compress original test sample y and excessively complete dictionary D
Excessively complete dictionary D data volumes directly by physiological data storehouse training sample construction are big, in order to reduce data volume, adopt
Dimensionality reduction compression can be carried out with subalpine forests random matrix to D, compression classification speed is improved：
In formula (4), L represents the dimension after dictionary D dimensionality reductions, therefore the dimension of dictionary D drops to L dimensions from m dimensions.ΦL×mFor subalpine forests
Random matrixThe matrix of composition, carries out dimensionality reduction to each sample in dictionary, and the original physiologic signal y that node is collected is same
It is also required to use dimension-reduction treatment, i.e.,：
Step 3：Seek rarefaction representation coefficient
In sparse representation theory, according to the excessively complete dictionary of compressionAnd test sampleTest sample can be tried to achieve
Sparse expression coefficientBut often contain noise during gathered data in real node, it is therefore desirable to plus an error term
ε, its solving model is：
Wherein ε >=0, represents allowable error.Formula (6) is a convex optimization problem, and existing many algorithms can be solved,
Test sample compressed signal is obtainedRarefaction representation coefficient
Step 4：Filter invalid signals
Body area network node is very important performance indications. collection signal kinds are more, and signal is easily disturbed, and when body area network is just opened, gather swinging of signal
It is fixed, it is impossible to reach diagnostic analysiss standard and some signals not in the training sample of physiological data storehouse, it is also possible to sent, this
A little invalid signals are sent to base station or remote medical center, have not only wasted node energy but also may cause mistaken diagnosis and warning, to carry
High RST classification accuracy, before signal identification, using sparse concentration degree index (Sparsity Centration Index,
SCI) invalid sample is excluded：
Wherein n is classification sum, i=1,2 ..., n during node Sample Establishing.δiX () representsIn i-th position
Coefficient.By the sparse coefficient of test sample yCalculate in substituting into SCI (x)Value, ifThen representing should
Test sample has only used one training sample of a certain apoplexy due to endogenous wind to illustrate test sample；IfThe test specimens are represented then
This is invalid sample signal across all of class.Therefore a threshold tau, τ ∈ [0,1], if SCI are set during differentiation
X () >=τ then thinks the test sample for effective sample, if SCI (x) is ＜ τ, test sample is invalid sample signal, invalid sample
This signal is directly filtered, and node restarts sampling, if being determined as effective sample, also need further to carry out signal it is normal with it is different
General knowledge is other, signal judge process such as Fig. 3.
Step 5：Whether identification test sample is normal signal
Due to effect of noise,Nonzero element can intersperse among between many classes, known using following residual error classification function
Not：
Wherein i=1,2 ..., n, thereforeExpression Minimum Residual difference, as final classification result, obtain signal point
Class-mark i, the normal and abnormal numbering according to each type data in classification just can recognize that whether test sample is normal letter
Number, if recognition result is that soon envoy's point enters resting state to normal physiological signal, if testing result is abnormal signal, just will
The test signal y ' of compression is sent to center base station node.Data are compressed with reconstruct in center base station node.
Step 6：Construction sensing matrix Acs
Compressive sensing theory mainly has rarefaction representation, random observation, three parts of reconstruct, and the rarefaction representation of signal is application
The priori conditions of compressed sensing, signal has openness generally on certain orthogonal intersection space, and needing to find certain orthogonal basis Ψ makes
Signal is sparse on the base：
Y=Ψ θ (9)
In formula, y is acquired original signal, and θ is rarefaction representation coefficient, and Ψ is the sparse base of correspondence species signal, commonly uses sparse base
There are Fourier's base, wavelet basiss, cosine basis etc..The sparse identification of signal is carried out in node, and signal reconstruction is carried out in base station, excessively complete
Dictionary constructs part, and training dictionary is not the dictionary that immobilizes, and learns through training, and dictionary atom and rarefaction representation coefficient can not
Disconnected to update, dictionary data amount is larger, dictionary data is transmitted from node to base station and can consume more multi-energy.Therefore in signal reconstruction portion
Point, the dictionary that node can not be waited by transmitting training makes center base station to signal reconstruction, needs to select suitable orthogonal basiss Ψ reconstruct
Signal.
Know that the signal made by compression transmission can be reconstructed by formula (2), the observing matrix of signal needs the sight for meeting RIP
Matrix is surveyed, in the sparse identification division of signal, chosen subalpine forests random matrixIt is as signal accidental projection matrix therefore former
Dimension-reduction treatment of the beginning signal through formula (5), signalAs compressed sensing reconstructs the observation of part, and formula (9) can be write as：
In formula (10)Center base station Node compression abnormal signal, i.e. compressed sensing observation are sent to for body area network node is very important performance indications.,For subalpine forests accidental projection matrix, Ψ is the orthogonal sparse base of signal correspondence, ifFor compressed sensing sensing matrix,
Therefore the change of formula (10) expression formula is turned to：
Step 7：Calculate reconstruct sparse coefficient θ '
Known measurementsWith sensing matrix Acs, sparse coefficient θ ' is obtained by restructing algorithm, has been had compared with multiple pressure at present
Contracting perceptual signal restructing algorithm, predominantly convex optimized algorithm and greedy tracing algorithm.Convex optimized algorithm is common base tracing algorithm
BP, interior point method IP, gradient projection method GPSR and iteration threshold algorithm, greedy algorithm be mainly matching pursuit algorithm MP, orthogonal
With tracing algorithm OMP, regularization constraint algorithm ROMP, and hybrid algorithm.
Step 8：Recovery and rebuilding signal y '
Sparse coefficient θ ' is obtained, then signal y ' just can be calculated by formula (9), such as following formula：
Y '=ψ θ ' (12)
Step 9：Diagnostic signal
After obtaining reconstruction signal y ', reported to the police and while signal is issued to tele-medicine end in body area network center base station
End diagnostic analysiss simultaneously gather emergency treatment.So far, whole body area network energy-saving scheme system design, whole body area network energy-conservation are completed
System work such as Fig. 1, realizes body area network system energy-conservation using method such as Fig. 2, and sparse identification is used in and normal signal is recognized, makes
Node stops normal signal and sends, and reduces radio communication energy consumption, and compressive sensing theory is for the dimensionality reduction to primary signal and in base
Stand reconstruct, reduce body area network wirelessly transmitting data amount, so as to reduce body area network energy consumption.
In sum, the invention provides a kind of body area network power-economizing method based on compressive sensing theory more than rarefaction representation,
First, sample data in data with existing storehouse is carried out normally with anomaly classification and number and be built into complete dictionary D, Ran Houtong
Cross subalpine forests random matrix and dimension-reduction treatment is done to test sample y and dictionary D, compressed signal is calculated with rarefaction representation algorithm's
Rarefaction representation is sparseThen, pass throughJudge whether test sample is useful signal, be judged as that invalid signals are direct
Filter, node restarts sampling, be judged as useful signal then by residual error identification signal calculated least residual position i and data
The verification of storehouse labelling, if normal signal, node is directly entered resting state, if abnormal signal, node is by compressed signalPass
The reconstruct of center base station node is defeated by, reconstruction signal y ' is sent to into remote medical center diagnostic analysiss finally and is taken first aid to arrange
Apply, realize body area network node is very important performance indications. dormancy energy-conservation and data compression transmission energy-conservation, reduce body area network energy consumption.
The preferred embodiment of the present invention described in detail above.It should be appreciated that one of ordinary skill in the art without
Need creative work just can make many modifications and variations with design of the invention.Therefore, all technologies in the art
Personnel are available by logical analysis, reasoning, or a limited experiment on the basis of existing technology under this invention's idea
Technical scheme, all should be in the protection domain being defined in the patent claims.
Claims (5)
1. it is a kind of based on the sparse body area network power-economizing method recognized with compressed sensing, it is characterised in that to comprise the steps：
S1, extraction data base's physiological signal sample built complete dictionary D：According to the physiological signal type of node collection, to section
Data base's sample of the collection signal of point carries out normal and anomaly classification labelling, and training sample was configured to complete dictionary D；
S2, the original test sample y of compression and excessively complete dictionary D：Subalpine forests random matrix dimensionality reduction pressure is adopted to original physiologic signal y
ContractingEqually carry out dimensionality reduction compression to obtain compressing dictionary to excessively complete dictionary D：
S3, seek rarefaction representation coefficientBy compressed signalAnd dictionaryConsider influence of noise in reality, add noise jamming to miss
Difference item ε, according to sparse representation theoryChanging solving model isObtain and be
Number expression is sparse
S4, filter invalid signals：By sparse coefficientSubstitute into sparse concentration degree indexAnd obtainValue, than
Compared withWith the size of given threshold τ, whether discriminating test sample is effective sample signal, if invalid sample signal, directly
Connect and filter, into step S1, node restarts physiological data of sampling, and is useful signal, then into next step；
Wherein, it is describedIt is expressed as：Threshold tau ∈ [0,
1]；N is classification sum, i=1,2 ..., n during node Sample Establishing；δiX () representsIn i-th position coefficient；
Whether S5, identification test sample are normal signal：Normal signal is identified using residual error classification function：Calculate least residualLeast residual position i is obtained, check and correction has been numbered
Determine the effective sample generic, whether identification test signal y is normal signal, if the test signal normal signal,
Switching collection normal signal node is resting state, if it is abnormal signal that the test signal is detected result, by compression letter
NumberIt is sent to the reconstruct of center base station node；
S6, construction sensing matrix Acs：Selection makes node collection signal in the sparse sparse base Ψ of orthogonal intersection space, with described Asia height
This is randomIt is configured to compressed sensing sensing matrix
S7, calculating reconstruct sparse coefficient θ '：Compressed signal is received according to center base station nodeWith sensing matrix Acs, with compression
Perceptual signal restructing algorithm, calculates reconstruct sparse coefficient θ ',
S8, recovery and rebuilding signal y '：According to sparse coefficient θ ' and orthogonal basiss Ψ has been reconstructed, reconstruction signal y '=ψ θ ' are calculated；
S9, diagnostic signal：Reconstruction signal is sent to remote medical center carries out diagnostic analysiss, and in center base station node early warning
Prompting.
2. according to claim 1 a kind of based on the sparse body area network power-economizing method recognized with compressed sensing, its feature exists
In：It is in the S4, describedCompare with threshold tau is set, ifThen test sample signal is invalid
Sample signal, invalid signals are directly filtered, and node resurveys data, ifThen discriminating test sample is effective
Sample signal, initially enters next step to test sample signal identification.
3. according to claim 1 a kind of based on the sparse body area network power-economizing method recognized with compressed sensing, its feature exists
In：In step S5, least residual is calculated with residual error function, obtain the affiliated label i of test sample, and pass through institute
Come to recognize test sample be abnormal signal or normal signal, if normal signal, node stops the label classified in stating step S1
Data are sent only, into resting state, if abnormal signal, node directly sends compression verification signal to center base station node Measured value in as described S7 in the theoretical reconstruct test signals of CS.
4. according to claim 1 a kind of based on the sparse body area network power-economizing method recognized with compressed sensing, its feature exists
In：In step S6, center base station node builds sensing matrix AcsWalked by described in the sparse orthogonal basiss Ψ of signal and node
Subalpine forests random matrix in rapid S2Constitute, theoretical according to CS, to enable signals to reconstruct, calculation matrix needs to meet limited
Isometry RIP, subalpine forests random matrixCalculation matrix during as CS is theoretical, compresses to test signal, subalpine forests random matrixStructure is very sparse and has been demonstrated to meet limited isometry RIP with high probability.
5. according to claim 1 a kind of based on the sparse body area network power-economizing method recognized with compressed sensing, its feature exists
In：In step S7, the expression formula of reconstruct sparse coefficient θ ' is calculatedMiddle measured valueAs described step S2
It is middle through dimensionality reduction test signal, test sample is identified as center base station section is sent directly to after abnormal signal by step S5
Point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611044550.9A CN106559744A (en) | 2016-11-24 | 2016-11-24 | It is a kind of based on sparse identification and the body area network power-economizing method of compressed sensing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611044550.9A CN106559744A (en) | 2016-11-24 | 2016-11-24 | It is a kind of based on sparse identification and the body area network power-economizing method of compressed sensing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106559744A true CN106559744A (en) | 2017-04-05 |
Family
ID=58444525
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611044550.9A Pending CN106559744A (en) | 2016-11-24 | 2016-11-24 | It is a kind of based on sparse identification and the body area network power-economizing method of compressed sensing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106559744A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108852340A (en) * | 2018-05-17 | 2018-11-23 | 广东工业大学 | A kind of ECG Gathering System, device and computer readable storage medium |
| CN109275138A (en) * | 2018-07-27 | 2019-01-25 | 四川大学 | A kind of method of signal compression acquisition and data encryption |
| CN109391931A (en) * | 2018-11-01 | 2019-02-26 | 北京京航计算通讯研究所 | High reliability data transmission system applied to emergency disposal |
| CN109993105A (en) * | 2019-03-29 | 2019-07-09 | 北京化工大学 | A kind of improved adaptive sparse sampling Fault Classification |
| CN110974223A (en) * | 2019-12-13 | 2020-04-10 | 河北工业大学 | Surface electromyogram signal compression reconstruction method based on improved KSVD algorithm |
| CN112019530A (en) * | 2020-08-17 | 2020-12-01 | 宁波工程学院 | Physiological signal safe compression method and system suitable for body area network |
| CN113627544A (en) * | 2021-08-16 | 2021-11-09 | 沈阳工业大学 | Machine tool milling cutter state identification method based on multi-source heterogeneous data fusion |
| CN114034755A (en) * | 2021-10-13 | 2022-02-11 | 郑州航空工业管理学院 | Abnormal particulate matter detection method based on engine gas circuit electrostatic signal |
| CN115618201A (en) * | 2022-10-09 | 2023-01-17 | 湖南万脉医疗科技有限公司 | Breathing machine signal processing method based on compressed sensing and breathing machine |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103944579A (en) * | 2014-04-10 | 2014-07-23 | 东华大学 | Coding and decoding system for compressed sensing reconstitution |
| CN104122540A (en) * | 2014-07-31 | 2014-10-29 | 电子科技大学 | Matching dictionary and compressive sensing based radar range profile object identification method |
| CN104579361A (en) * | 2014-10-29 | 2015-04-29 | 上海大学 | Low-power-consumption multi-source physiological signal mixed compression method for wireless body area network |
| CN104779960A (en) * | 2015-03-20 | 2015-07-15 | 南京邮电大学 | A signal reconstruction method based on block compressed sensing |
| CN105099462A (en) * | 2014-05-22 | 2015-11-25 | 北京邮电大学 | Signal processing method based on compressive sensing |
| CN106100860A (en) * | 2016-06-16 | 2016-11-09 | 湖北文理学院 | A kind of body area network node dormancy method based on rarefaction representation |
-
2016
- 2016-11-24 CN CN201611044550.9A patent/CN106559744A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103944579A (en) * | 2014-04-10 | 2014-07-23 | 东华大学 | Coding and decoding system for compressed sensing reconstitution |
| CN105099462A (en) * | 2014-05-22 | 2015-11-25 | 北京邮电大学 | Signal processing method based on compressive sensing |
| CN104122540A (en) * | 2014-07-31 | 2014-10-29 | 电子科技大学 | Matching dictionary and compressive sensing based radar range profile object identification method |
| CN104579361A (en) * | 2014-10-29 | 2015-04-29 | 上海大学 | Low-power-consumption multi-source physiological signal mixed compression method for wireless body area network |
| CN104779960A (en) * | 2015-03-20 | 2015-07-15 | 南京邮电大学 | A signal reconstruction method based on block compressed sensing |
| CN106100860A (en) * | 2016-06-16 | 2016-11-09 | 湖北文理学院 | A kind of body area network node dormancy method based on rarefaction representation |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108852340A (en) * | 2018-05-17 | 2018-11-23 | 广东工业大学 | A kind of ECG Gathering System, device and computer readable storage medium |
| CN109275138A (en) * | 2018-07-27 | 2019-01-25 | 四川大学 | A kind of method of signal compression acquisition and data encryption |
| CN109391931B (en) * | 2018-11-01 | 2021-03-05 | 北京京航计算通讯研究所 | High-reliability data transmission system applied to emergency disposal |
| CN109391931A (en) * | 2018-11-01 | 2019-02-26 | 北京京航计算通讯研究所 | High reliability data transmission system applied to emergency disposal |
| CN109993105A (en) * | 2019-03-29 | 2019-07-09 | 北京化工大学 | A kind of improved adaptive sparse sampling Fault Classification |
| CN110974223A (en) * | 2019-12-13 | 2020-04-10 | 河北工业大学 | Surface electromyogram signal compression reconstruction method based on improved KSVD algorithm |
| CN112019530A (en) * | 2020-08-17 | 2020-12-01 | 宁波工程学院 | Physiological signal safe compression method and system suitable for body area network |
| CN112019530B (en) * | 2020-08-17 | 2022-05-31 | 宁波工程学院 | Physiological signal safe compression method and system suitable for body area network |
| CN113627544A (en) * | 2021-08-16 | 2021-11-09 | 沈阳工业大学 | Machine tool milling cutter state identification method based on multi-source heterogeneous data fusion |
| CN113627544B (en) * | 2021-08-16 | 2024-03-08 | 沈阳工业大学 | Machine tool milling cutter state identification method based on multi-source heterogeneous data fusion |
| CN114034755A (en) * | 2021-10-13 | 2022-02-11 | 郑州航空工业管理学院 | Abnormal particulate matter detection method based on engine gas circuit electrostatic signal |
| CN114034755B (en) * | 2021-10-13 | 2024-01-12 | 郑州航空工业管理学院 | Abnormal particulate matter detection method based on engine gas circuit electrostatic signals |
| CN115618201A (en) * | 2022-10-09 | 2023-01-17 | 湖南万脉医疗科技有限公司 | Breathing machine signal processing method based on compressed sensing and breathing machine |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106559744A (en) | It is a kind of based on sparse identification and the body area network power-economizing method of compressed sensing | |
| CN107103182A (en) | A kind of heart disease Warning System and method based on deep learning algorithm | |
| CN107788976A (en) | Sleep monitor system based on Amplitude integrated electroencephalogram | |
| CN109247936B (en) | Abnormal electroencephalogram behavior monitoring system and method for whole night sleep monitoring | |
| CN107742534A (en) | Patient's viability forecasting system | |
| CN103637784B (en) | Based on the physiological parameter acquisition system of ZigBee technology | |
| CN110960211A (en) | Embedded-based real-time electrocardio monitoring system | |
| Jin et al. | Classification of normal and abnormal ECG records using lead convolutional neural network and rule inference | |
| Fyntanidou et al. | IoT-based smart triage of Covid-19 suspicious cases in the Emergency Department | |
| CN106446765A (en) | Health state evaluation system based on multidimensional physiological big data depth learning | |
| CN203133863U (en) | Digital cardiovascular accident early diagnosis system | |
| CN201227272Y (en) | Dress type electro-cardio and respiration rate monitoring device based on Zigbee | |
| CN111370124A (en) | Health analysis system and method based on facial recognition and big data | |
| CN206443702U (en) | A kind of noninvasive dynamics monitoring device | |
| CN113288168A (en) | Wearable fatigue monitoring of intelligence and early warning system | |
| CN106691498A (en) | Borborygmus processing system | |
| CN105943022B (en) | A kind of cardioelectric monitor system that there are three leads to reconstruct 12 lead function | |
| CN111643074A (en) | Lossy compression electrocardiosignal atrial fibrillation screening system | |
| CN109662708A (en) | A kind of postoperative patient behavior monitoring method based on deep neural network | |
| Rodríguez-Jorge et al. | Internet of things-assisted architecture for QRS complex detection in real time | |
| CN117133464B (en) | Intelligent monitoring system and monitoring method for health of old people | |
| Li et al. | Detection of muscle fatigue by fusion of agonist and synergistic muscle semg signals | |
| CN206301393U (en) | A kind of physical signs dynamic monitoring system of suitable home environment | |
| CN110960207A (en) | Tree model-based atrial fibrillation detection method, device, equipment and storage medium | |
| CN110693466A (en) | Rehabilitation evaluation device based on remote control induction technology and implementation method thereof |
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: 20170405 |