CN110516878A - A kind of localization method and system of the sparse signal representation model based on space-time restriction - Google Patents
A kind of localization method and system of the sparse signal representation model based on space-time restriction Download PDFInfo
- Publication number
- CN110516878A CN110516878A CN201910806397.6A CN201910806397A CN110516878A CN 110516878 A CN110516878 A CN 110516878A CN 201910806397 A CN201910806397 A CN 201910806397A CN 110516878 A CN110516878 A CN 110516878A
- Authority
- CN
- China
- Prior art keywords
- client
- signal
- site
- sparse
- current location
- 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
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
- G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0639—Item locations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/02—Banking, e.g. interest calculation or account maintenance
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Game Theory and Decision Science (AREA)
- Remote Sensing (AREA)
- Technology Law (AREA)
- Radar, Positioning & Navigation (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Position Fixing By Use Of Radio Waves (AREA)
Abstract
The localization method and system of the invention discloses a kind of sparse signal representation model based on space-time restriction, method include: to obtain client in the signal of site current location；Signal based on client in site current location, and by the sparse signal representation model of time constraint and space constraint factor optimizing, client is obtained in the current location information of site.The present invention can be accurately positioned according to the signal of client current location and by the sparse signal representation model of time constraint and space constraint factor optimizing to client currently in the accurate location of site, accurately be analyzed convenient for subsequent customer action.
Description
Technical field
The present invention relates to signal processing technology field more particularly to a kind of sparse signal representation models based on space-time restriction
Localization method and system.
Background technique
Client is relatively long on average in the time of bank outlets' transacting business at present, in the time for waiting transacting business
Interior, client waits in Accreditation Waiting Area sometimes, the product that experience bank releases in experience area sometimes.It can thus be seen that in visitor
Family waited in the time of transacting business, if it is possible to know each client in the specific location of bank outlets, to analysis mining visitor
The potential professional ability at family has certain help.
Therefore, how effective location client is a urgent problem to be solved in the specific location of bank outlets.
Summary of the invention
In view of this, the localization method of the present invention provides a kind of sparse signal representation model based on space-time restriction, energy
It is enough effectively to navigate to client in the specific location of bank outlets.
The localization method of the present invention provides a kind of sparse signal representation model based on space-time restriction, comprising:
Client is obtained in the signal of site current location；
Signal based on the client in site current location, and it is excellent by time constraint and the space constraint factor
The sparse signal representation model of change obtains the client in the current location information of site.
Preferably, the method also includes:
Current location information of the client based on acquisition in site analyzes the behavior of client, exports client's row
To analyze result.
It is preferably, described to obtain client before the signal of site current location, further includes:
Collect the signal strength on the different location of site known coordinate；
Different location and corresponding signal strength based on known coordinate construct received signals fingerprint library.
Preferably, the method also includes:
Sparse signal representation model is constructed by simulation client in the path of site based on the received signals fingerprint library；
It is sparse that signal described in time constraint and space constraint factor pair is added in the sparse signal representation model
Indicate that model optimizes, the sparse signal representation model after being optimized.
A kind of positioning system of the sparse signal representation model based on space-time restriction, comprising:
Module is obtained, for obtaining client in the signal of site current location；
Locating module, for the signal based on the client in site current location, and by time constraint and
The sparse signal representation model of space constraint factor optimizing obtains the client in the current location information of site.
Preferably, the system also includes:
Analysis module, the current location information for the client based on acquisition in site divide the behavior of client
Analysis exports customer behavior analysis result.
Preferably, the system also includes:
Information collection module, the signal strength on different location for collecting site known coordinate；
Fingerprint base constructs module, for different location and corresponding signal strength building signal based on known coordinate
Fingerprint base.
Preferably, the system also includes:
Model construction module, by simulation client in the path of site, constructs signal for being based on the received signals fingerprint library
Sparse representation model；
Optimized model module, for added in the sparse signal representation model time constraint and space constraint because
Son optimizes the sparse signal representation model, the sparse signal representation model after being optimized.
In conclusion the localization method of the invention discloses a kind of sparse signal representation model based on space-time restriction, when
Need to be accurately positioned client at the position of site, acquisition client is then based on client in the signal of site current location first
Signal in site current location, and the sparse signal representation mould by time constraint and space constraint factor optimizing
Type obtains client in the current location information of site.The present invention can be according to the signal of client current location and by the time
The sparse signal representation model of constraint factor and space constraint factor optimizing is accurately positioned to client currently in the accurate position of site
It sets, customer action is accurately analyzed convenient for subsequent.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is a kind of localization method embodiment 1 of the sparse signal representation model based on space-time restriction disclosed by the invention
Method flow diagram；
Fig. 2 is a kind of localization method embodiment 2 of the sparse signal representation model based on space-time restriction disclosed by the invention
Method flow diagram；
Fig. 3 is a kind of positioning system embodiment 1 of the sparse signal representation model based on space-time restriction disclosed by the invention
Structural schematic diagram；
Fig. 4 is a kind of positioning system embodiment 2 of the sparse signal representation model based on space-time restriction disclosed by the invention
Structural schematic diagram.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
As shown in Figure 1, being a kind of localization method of the sparse signal representation model based on space-time restriction disclosed by the invention
The method flow diagram of embodiment 1, the method may include following steps:
S101, client is obtained in the signal of site current location；
When client enters bank outlets, and the specific location to client in site is needed to be accurately positioned, obtain first
Signal to client in bank outlets current location.
Specifically, can be got by client when entering site obtaining client in the signal of site current location
Queue number board get.Wherein, queue number board needs to carry out in advance special processing and production, so that number slip can be with
The collection and storage of wireless signal are carried out, and is able to carry out the transmission of signal upon activation.
S102, the signal based on client in site current location, and pass through time constraint and the space constraint factor
The sparse signal representation model of optimization obtains client in the current location information of site.
When getting client after the signal of site current location, further about according to process time constraint and space
The sparse signal representation model of beam factor optimizing is determined approximate in the signal strength of site current location with client in a model
Signal strength indication, and corresponding position determines the client in the present bit of site in a model according to approximate signal strength indication
It sets.
It should be noted that rarefaction representation, which refers to, indicates letter with atom as few as possible in given super complete dictionary
Number, the more succinct representation of signal can be obtained, to more easily obtain the information that is contained in signal, be more convenient into
One step is processed signal.
In conclusion in the above-described embodiments, when needing to be accurately positioned position of the client in site, obtaining client first
Signal in site current location is then based on client in the signal of site current location, and by time constraint and
The sparse signal representation model of space constraint factor optimizing obtains client in the current location information of site.The present invention being capable of root
Signal according to client current location and the sparse signal representation model by time constraint and space constraint factor optimizing
Client is accurately positioned currently in the accurate location of site, customer action is accurately analyzed convenient for subsequent.
As shown in Fig. 2, being a kind of localization method of the sparse signal representation model based on space-time restriction disclosed by the invention
The method flow diagram of embodiment 2, the method may include following steps:
S201, collect site known coordinate different location on signal strength；
S202, the different location based on known coordinate and corresponding signal strength construct received signals fingerprint library；
S203, it is based on received signals fingerprint library, by simulation client in the path of site, constructs sparse signal representation model；
S204, time constraint and space constraint factor pair sparse signal representation are added in sparse signal representation model
Model optimizes, the sparse signal representation model after being optimized；
S205, client is obtained in the signal of site current location；
S206, the signal based on client in site current location, and pass through time constraint and the space constraint factor
The sparse signal representation model of optimization obtains client in the current location information of site；
S207, the current location information based on the client of acquisition in site analyze the behavior of client, export client
Behavioural analysis result.
Specifically, in the above-described embodiments, rarefaction representation is in recent years about fields such as image recognition, computer visions
Research hotspot, it indicates the main information of original signal by non-zero coefficient as few as possible, to make the solution of signal processing
Journey becomes simpler and is easy.It is exactly compression sampling in an important application of field of signal processing, rarefaction representation.
Solving Sparse Problems includes the selection of basic function dictionary matrix ψ and the derivation algorithm of Sparse Problems.General dictionary ψ
All it was complete dictionary, and met the properties such as orthogonal；For the derivation algorithm of Sparse Problems, since the essence of rarefaction representation is convex ask
The solution of topic, therefore directly can go to solve using the algorithm that the generality of such as interior point method solves convex optimization problem, but it is common
Algorithm not can be well solved sparsity problem, so, for the particularity of sparse solution, propose that such as GP algorithm, IST are calculated
Method, ALM algorithm go to solve Sparse Problems, on the one hand these algorithms solve the solution of Sparse Problems, have also pushed simultaneously sparse
Indicate the application in practical problem.
The mathematical model of rarefaction representation approximately as: to an one-dimensional discrete signal x, it is made of time-limited real value,
It can be regarded as the column vector tieed up in the M × 1 in the space RM, element is x [n], n=1,2 ..., M.Any signal in the space RM
The orthogonal basal orientation that can be tieed up with N × 1Linear combination indicate, then x is expressed as x=ψ θ.
VectorThe orthogonal basis dictionary matrix of M × N is formed as column vectorAppoint
Meaning signal x may be expressed as (1-1)
Wherein θ is weighting coefficientThe column vector of N × 1 of composition.Obviously, x and θ is same
The equivalent representation of signal, x are expression of the signal in real domain, and θ is expression of the signal in the domain ψ.N > M is required in above formula, according to linear
The knowledge of algebra according to sparse condition, can be chosen in all feasible solutions it is found that sparse equation has the solution of infinite multiple groups
The least solution of nonzero element, that is, meet sparsity.Then following mathematical model (1-2) is obtained:
min||x||0S.t. x=ψ θ (1-2)
It is zero norm constraint in objective function, is NP problem, Terry is proved, under certain condition, zero norm problem and one
Norm problem be it is of equal value, then above-mentioned model conversation be (1-3):
min||x||1S.t. x=ψ θ (1-3)
Existing sparse signal representation method can be divided into two class of orthogonal basis rarefaction representation and redundant dictionary rarefaction representation.Cause
It is the non-sparse natural sign in time domain in the transformation by certain transform domains, can be converted into sparse.Orthogonal basis sparse table
Show this characteristic that signal is utilized, mainly project to signal on orthogonal transformation basic function, obtains sparse or approximate sparse
Transformation vector.
When being unable to rarefaction representation signal with orthogonal basis, so that it may replace basic function with redundancy functions, be carried out to signal
Rarefaction representation.Super complete redundancy functions are also referred to as redundant dictionary, and element therein is referred to as dictionary atom.The choosing of dictionary
Selecting be to meet the structure for being reconstructed signal, and the sparse bayesian learning process of signal is exactly to find from the redundant dictionary of building and original
Beginning signal has the K item atom of optimum linear combination.Because collection obtains in the experiment scene of design in actual location
The number of resampled finger point is always far longer than the number of test point, thus received signals fingerprint library of the invention using redundant dictionary come
Building.
In the present invention, received signals fingerprint library is constructed using the method for redundant dictionary.Specific process is as follows, in off-line training
Stage, it is assumed that a total of M AP (Access Point can emit the equipment of signal, usually router) in scene, and
N number of known coordinate point in scene, i.e., (Received Signal Strength, wireless signal are strong by the RSS of acquisition AP on fingerprint point
Degree) value, the matrix ψ ∈ R of a M*N available in this wayM×N.The fingerprint point number of usual RSS acquisition is much larger than the quantity of AP,
Therefore the matrix that fingerprint collecting obtains has redundancy in column vector, can be used as the redundant dictionary of next step rarefaction representation, is denoted as
I.e. Indicate the signal strength indication that i-th of AP is observed on j-th of fingerprint point,
Each column vectorIndicate j-th of position (xj,yj) on the wire size intensity value of M AP observed, will
Signal strength and its position are known as scene fingerprint, are denoted asIn specific experiment, if in this position
It sets, does not receive the RSS value of some AP, then the value of RSS is set as -100dBm, can guarantee the complete of finger print data in this way
Property.Because the fingerprint point number collected can be much larger than the location point number being collected into test phase, received signals fingerprint library can
Using as the redundant dictionary in rarefaction representation algorithm.
Then need to acquire the signal strength on the action trail of client.Test signal YtQueue number board is represented in t
The signal strength indication for all AP transmitting that a location point is collected into, indicates such as formula (1-5):
Yt=[yt1,yt2,...ytM]T (1-5)
Wherein ytiIt indicates to be collected into the RSS value of i-th of AP transmitting t-th of time point, M is the number of AP.
Matching relationship between received signals fingerprint library and actual test signal in order to obtain can be gone by rarefaction representation algorithm
It solves, can be solved to obtain θ by formula (1-6), θ is the sparse vector solution for being rarefaction representation, is exactly with considerably less non-zero coefficient
Indicate that signal Y, Y are shown in formula (1-5) that ψ is the signal strength data library of all fingerprint points, and i-th of element in θ, corresponds to ψ here
In the i-th column signal intensity value, be equivalent to corresponding i-th of coordinate value.Therefore by solving θ, so that it may estimated location.
In test phase, one section of continuous signal Υ is acquired, Υ is divided into n sections to get Υ=[Y1,...,Yn], Yt=
[yt1,yt2,...ytM]T, t=1 ..., T represents the vector in t-th of time point all M AP signal strength indications.Pass through in this way
Formula (1-7) is availableIt can be obtained by the location information of every column signal in Υ, to realize the positioning of all positions.In
It can be solved using CVX kit in the present invention formula (1-7).
WhereinIndicate the rarefaction representation matrix of Υ, θt=[θt1,...,θTN]T∈RNIt is Yt
Rarefaction representation vector；It isOptimal estimation.Consider that actual signal intensity is represented by other, especially peripheral location
The linear combination of signal strength, therefore introduce the nonnegativity of sparse coefficient and the constraint of linear combination
Once obtaining rarefaction representation coefficient of the observation signal in fingerprint redundant dictionaryIt then can use fingerprint signal
The position of the location information estimation observation signal of intensity, sees formula (1-8):
Wherein (xn,yn) in the coordinate value of location point n, r is the threshold value of non-zero rarefaction representation coefficient, the sparse table greater than r
Show that the corresponding fingerprint signal of coefficient is to think fingerprint point relevant to Current observation signal.It is finally real by location estimation above
The positioning based on rarefaction representation is showed.
By obtaining the study found that the positioning result of many matching algorithms does not consider continuity problem at present
Effect picture is not a continuous line, but common people can be continuous when walking.For test phase walking path
On the RSS data that is collected into, adjacent location point signal strength would not differ too big, i.e., for each path point, it upper
The distance between one step and path point of next step differ will not be very big.For the path of a continuous walking, we can know
The fluctuation of path is without departing from certain range, and existing location algorithm, in many cases all it cannot be guaranteed that point and point
Between distance to a very small extent, be easy to appear king-sized fluctuation, lead to positioning result poor effect.
So the present invention adds time constraint condition on the basis of rarefaction representation algorithm.It is corresponding for the same AP
Signal value, the upper position signal message difference corresponding with the position of next point in walking should be smaller, corresponding to sit
The distance difference of punctuate will not be especially big, for observation Υ=[Y of Time Continuous1,...,YT], due to the continuity of its data
Determine its rarefaction representation coefficientAlso there is temporal continuity, therefore here to dilute in (1-7)
Dredging indicates that model introduces the constraint of time continuity.Specifically, the corresponding sparse coefficient θ of adjacent observation upper for the timetWith
θt+1, it is desirable to they have continuity, i.e., | | θt-θt+1||F 2Want small, so as to construct following matrix, T is shown in formula (1-9)
Analysis by the sparse vector solved to front, it can be found that the sampled point far from anchor point can also
There can be very big weight, i.e., far point is very big on the influence of the coordinate of true location point, and such case Producing reason may
It is that will cause positioning result in this way because the complexity of indoor environment, noise, multipath cause signal similarity-rough set high and generate very
Big error, so, for this problem, the present invention proposes space constraints on the basis of time-constrain rarefaction representation.
Space constraints are mainly to constrain the position distribution of θ value in sparse vector, and the signal obtained for space position (x, y) is strong
Spend Y(x,y), it is believed that the signal strength that only position adjacent thereto obtains is related, that is, there is a neighborhood O of (x, y)(x,y)So thatTherefore, the rarefaction representation coefficient of observation signal also has in sparse representation model
The corresponding position of relative signal strength of such spatial continuity, i.e. its non-zero sparse coefficient selection should belong to observation station and exist
In one neighborhood of position.Due to the Location-Unknown of observation signal intensity, can be retouched with fingerprint neighborhood of a point nearest therewith
This spatial locality matter is stated, that is, there is j, so thatWherein
Indicate that said combination is in fingerprint point (xj,yj) neighborhoodMiddle progress.In other words, signal Mr. Yu position obtained
Intensity must be the linear combination of fingerprint in a neighborhood of some fingerprint point, i.e., its corresponding rarefaction representation coefficient is in its neighbour
The element non-zero on fingerprint point in domain, and outside field it is zero.For this purpose, defining the pact that can reflect above-mentioned spatial continuity
Beam matrix S, its specific element definition are that S is shown in formula (1-10)
Generalized time constraint and the space constraint factor, obtain new model (1-11) on the basis of sparse representation model:
Formula (1-11) can be converted to (1-12):
Because of θiNot only 1 sparse vector, i.e. more than one position are non-zero, so one threshold value r of setting, value
Several positions bigger than r are made last position and are calculated, such as formula (1-13):
Since formula (1-12) is a more complicated model, direct solution is relatively difficult, it is therefore possible to use ADMM
The solution of method progress model.
It, can also further working as in site according to the client of acquisition after obtaining current location information of the client in site
Front position information analyzes the behavior of client, for example, the current location information according to client in site can analyze out visitor
Family be currently at Accreditation Waiting Area still experience area, when analyze client be currently at experience area when, product related personnel can and
When the experience area that goes to be that client carries out related introduction, and then realizes to the precision marketing of client.
In conclusion the present invention can be according to the signal of client current location and by time constraint and space about
The sparse signal representation model of beam factor optimizing is accurately positioned to client currently in the accurate location of site, and according to the visitor of acquisition
Current location information of the family in site analyzes the behavior of client, can be realized the essence to client according to behavioural analysis result
Quasi- marketing.
As shown in figure 3, being a kind of positioning system of the sparse signal representation model based on space-time restriction disclosed by the invention
The structural schematic diagram of embodiment 1, the system may include:
Module 301 is obtained, for obtaining client in the signal of site current location；
When client enters bank outlets, and the specific location to client in site is needed to be accurately positioned, obtain first
Signal to client in bank outlets current location.
Specifically, can be got by client when entering site obtaining client in the signal of site current location
Queue number board get.Wherein, queue number board needs to carry out in advance special processing and production, so that number slip can be with
The collection and storage of wireless signal are carried out, and is able to carry out the transmission of signal upon activation.
Locating module 302 for the signal based on client in site current location, and passes through time constraint and sky
Between constraint factor optimize sparse signal representation model, obtain client site current location information.
When getting client after the signal of site current location, further about according to process time constraint and space
The sparse signal representation model of beam factor optimizing is determined approximate in the signal strength of site current location with client in a model
Signal strength indication, and corresponding position determines the client in the present bit of site in a model according to approximate signal strength indication
It sets.
It should be noted that rarefaction representation, which refers to, indicates letter with atom as few as possible in given super complete dictionary
Number, the more succinct representation of signal can be obtained, to more easily obtain the information that is contained in signal, be more convenient into
One step is processed signal.
In conclusion in the above-described embodiments, when needing to be accurately positioned position of the client in site, obtaining client first
Signal in site current location is then based on client in the signal of site current location, and by time constraint and
The sparse signal representation model of space constraint factor optimizing obtains client in the current location information of site.The present invention being capable of root
Signal according to client current location and the sparse signal representation model by time constraint and space constraint factor optimizing
Client is accurately positioned currently in the accurate location of site, customer action is accurately analyzed convenient for subsequent.
As shown in figure 4, being a kind of positioning system of the sparse signal representation model based on space-time restriction disclosed by the invention
The structural schematic diagram of embodiment 2, the system may include:
Information collection module 401, the signal strength on different location for collecting site known coordinate；
Fingerprint base constructs module 402, for different location and corresponding signal strength building letter based on known coordinate
Number fingerprint base；
Model construction module 403, by simulation client in the path of site, constructs signal for being based on received signals fingerprint library
Sparse representation model；
Optimized model module 404, for added in sparse signal representation model time constraint and space constraint because
Son optimizes sparse signal representation model, the sparse signal representation model after being optimized；
Module 405 is obtained, for obtaining client in the signal of site current location；
Locating module 406 for the signal based on client in site current location, and passes through time constraint and sky
Between constraint factor optimize sparse signal representation model, obtain client site current location information；
Analysis module 407, the current location information for the client based on acquisition in site divide the behavior of client
Analysis exports customer behavior analysis result.
Specifically, in the above-described embodiments, rarefaction representation is in recent years about fields such as image recognition, computer visions
Research hotspot, it indicates the main information of original signal by non-zero coefficient as few as possible, to make the solution of signal processing
Journey becomes simpler and is easy.It is exactly compression sampling in an important application of field of signal processing, rarefaction representation.
Solving Sparse Problems includes the selection of basic function dictionary matrix ψ and the derivation algorithm of Sparse Problems.General dictionary ψ
All it was complete dictionary, and met the properties such as orthogonal；For the derivation algorithm of Sparse Problems, since the essence of rarefaction representation is convex ask
The solution of topic, therefore directly can go to solve using the algorithm that the generality of such as interior point method solves convex optimization problem, but it is common
Algorithm not can be well solved sparsity problem, so, for the particularity of sparse solution, propose that such as GP algorithm, IST are calculated
Method, ALM algorithm go to solve Sparse Problems, on the one hand these algorithms solve the solution of Sparse Problems, have also pushed simultaneously sparse
Indicate the application in practical problem.
The mathematical model of rarefaction representation approximately as: to an one-dimensional discrete signal x, it is made of time-limited real value,
It can be regarded as the column vector tieed up in the M × 1 in the space RM, element is x [n], n=1,2 ..., M.Any signal in the space RM
The orthogonal basal orientation that can be tieed up with N × 1Linear combination indicate, then x is expressed as x=ψ θ.
VectorThe orthogonal basis dictionary matrix of M × N is formed as column vectorAppoint
Meaning signal x may be expressed as (1-1)
Wherein θ is weighting coefficientThe column vector of N × 1 of composition.Obviously, x and θ is same
The equivalent representation of signal, x are expression of the signal in real domain, and θ is expression of the signal in the domain ψ.N > M is required in above formula, according to linear
The knowledge of algebra according to sparse condition, can be chosen in all feasible solutions it is found that sparse equation has the solution of infinite multiple groups
The least solution of nonzero element, that is, meet sparsity.Then following mathematical model (1-2) is obtained:
min||x||0S.t. x=ψ θ (1-2)
It is zero norm constraint in objective function, is NP problem, Terry is proved, under certain condition, zero norm problem and one
Norm problem be it is of equal value, then above-mentioned model conversation be (1-3):
min||x||1S.t. x=ψ θ (1-3)
Existing sparse signal representation method can be divided into two class of orthogonal basis rarefaction representation and redundant dictionary rarefaction representation.Cause
It is the non-sparse natural sign in time domain in the transformation by certain transform domains, can be converted into sparse.Orthogonal basis sparse table
Show this characteristic that signal is utilized, mainly project to signal on orthogonal transformation basic function, obtains sparse or approximate sparse
Transformation vector.
When being unable to rarefaction representation signal with orthogonal basis, so that it may replace basic function with redundancy functions, be carried out to signal
Rarefaction representation.Super complete redundancy functions are also referred to as redundant dictionary, and element therein is referred to as dictionary atom.The choosing of dictionary
Selecting be to meet the structure for being reconstructed signal, and the sparse bayesian learning process of signal is exactly to find from the redundant dictionary of building and original
Beginning signal has the K item atom of optimum linear combination.Because collection obtains in the experiment scene of design in actual location
The number of resampled finger point is always far longer than the number of test point, thus received signals fingerprint library of the invention using redundant dictionary come
Building.
In the present invention, received signals fingerprint library is constructed using the method for redundant dictionary.Specific process is as follows, in off-line training
Stage, it is assumed that a total of M AP (Access Point can emit the equipment of signal, usually router) in scene, and
N number of known coordinate point in scene, i.e., (Received Signal Strength, wireless signal are strong by the RSS of acquisition AP on fingerprint point
Degree) value, the matrix ψ ∈ R of a M*N available in this wayM×N.The fingerprint point number of usual RSS acquisition is much larger than the quantity of AP,
Therefore the matrix that fingerprint collecting obtains has redundancy in column vector, can be used as the redundant dictionary of next step rarefaction representation, is denoted as
I.e. Indicate the signal strength indication that i-th of AP is observed on j-th of fingerprint point,
Each column vectorIndicate j-th of position (xj,yj) on the wire size intensity value of M AP observed, will
Signal strength and its position are known as scene fingerprint, are denoted asIn specific experiment, if in this position
It sets, does not receive the RSS value of some AP, then the value of RSS is set as -100dBm, can guarantee the complete of finger print data in this way
Property.Because the fingerprint point number collected can be much larger than the location point number being collected into test phase, received signals fingerprint library can
Using as the redundant dictionary in rarefaction representation algorithm.
Then need to acquire the signal strength on the action trail of client.Test signal YtQueue number board is represented in t
The signal strength indication for all AP transmitting that a location point is collected into, indicates such as formula (1-5):
Yt=[yt1,yt2,...ytM]T (1-5)
Wherein ytiIt indicates to be collected into the RSS value of i-th of AP transmitting t-th of time point, M is the number of AP.
Matching relationship between received signals fingerprint library and actual test signal in order to obtain can be gone by rarefaction representation algorithm
It solves, can be solved to obtain θ by formula (1-6), θ is the sparse vector solution for being rarefaction representation, is exactly with considerably less non-zero coefficient
Indicate that signal Y, Y are shown in formula (1-5) that ψ is the signal strength data library of all fingerprint points, and i-th of element in θ, corresponds to ψ here
In the i-th column signal intensity value, be equivalent to corresponding i-th of coordinate value.Therefore by solving θ, so that it may estimated location.
In test phase, one section of continuous signal Υ is acquired, Υ is divided into n sections to get Υ=[Y1,...,Yn], Yt=
[yt1,yt2,...ytM]T, t=1 ..., T represents the vector in t-th of time point all M AP signal strength indications.Pass through in this way
Formula (1-7) is availableIt can be obtained by the location information of every column signal in Υ, to realize the positioning of all positions.In
It can be solved using CVX kit in the present invention formula (1-7).
WhereinIndicate the rarefaction representation matrix of Υ, θt=[θt1,...,θTN]T∈RNIt is Yt
Rarefaction representation vector；It isOptimal estimation.Consider that actual signal intensity is represented by other, especially peripheral location
The linear combination of signal strength, therefore introduce the nonnegativity of sparse coefficient and the constraint of linear combination
Once obtaining rarefaction representation coefficient of the observation signal in fingerprint redundant dictionaryIt then can use fingerprint signal
The position of the location information estimation observation signal of intensity, sees formula (1-8):
Wherein (xn,yn) in the coordinate value of location point n, r is the threshold value of non-zero rarefaction representation coefficient, the sparse table greater than r
Show that the corresponding fingerprint signal of coefficient is to think fingerprint point relevant to Current observation signal.It is finally real by location estimation above
The positioning based on rarefaction representation is showed.
By obtaining the study found that the positioning result of many matching algorithms does not consider continuity problem at present
Effect picture is not a continuous line, but common people can be continuous when walking.For test phase walking path
On the RSS data that is collected into, adjacent location point signal strength would not differ too big, i.e., for each path point, it upper
The distance between one step and path point of next step differ will not be very big.For the path of a continuous walking, we can know
The fluctuation of path is without departing from certain range, and existing location algorithm, in many cases all it cannot be guaranteed that point and point
Between distance to a very small extent, be easy to appear king-sized fluctuation, lead to positioning result poor effect.
So the present invention adds time constraint condition on the basis of rarefaction representation algorithm.It is corresponding for the same AP
Signal value, the upper position signal message difference corresponding with the position of next point in walking should be smaller, corresponding to sit
The distance difference of punctuate will not be especially big, for observation Υ=[Y of Time Continuous1,...,YT], due to the continuity of its data
Determine its rarefaction representation coefficientAlso there is temporal continuity, therefore here to dilute in (1-7)
Dredging indicates that model introduces the constraint of time continuity.Specifically, the corresponding sparse coefficient θ of adjacent observation upper for the timetWith
θt+1, it is desirable to they have continuity, i.e., | | θt-θt+1||F 2Want small, so as to construct following matrix, T is shown in formula (1-9)
Analysis by the sparse vector solved to front, it can be found that the sampled point far from anchor point can also
There can be very big weight, i.e., far point is very big on the influence of the coordinate of true location point, and such case Producing reason may
It is that will cause positioning result in this way because the complexity of indoor environment, noise, multipath cause signal similarity-rough set high and generate very
Big error, so, for this problem, the present invention proposes space constraints on the basis of time-constrain rarefaction representation.
Space constraints are mainly to constrain the position distribution of θ value in sparse vector, and the signal obtained for space position (x, y) is strong
Spend Y(x,y), it is believed that the signal strength that only position adjacent thereto obtains is related, that is, there is a neighborhood O of (x, y)(x,y)So thatTherefore, the rarefaction representation coefficient of observation signal also has in sparse representation model
The corresponding position of relative signal strength of such spatial continuity, i.e. its non-zero sparse coefficient selection should belong to observation station and exist
In one neighborhood of position.Due to the Location-Unknown of observation signal intensity, can be retouched with fingerprint neighborhood of a point nearest therewith
This spatial locality matter is stated, that is, there is j, so thatWherein
Indicate that said combination is in fingerprint point (xj,yj) neighborhoodMiddle progress.In other words, signal Mr. Yu position obtained
Intensity must be the linear combination of fingerprint in a neighborhood of some fingerprint point, i.e., its corresponding rarefaction representation coefficient is in its neighbour
The element non-zero on fingerprint point in domain, and outside field it is zero.For this purpose, defining the pact that can reflect above-mentioned spatial continuity
Beam matrix S, its specific element definition are that S is shown in formula (1-10)
Generalized time constraint and the space constraint factor, obtain new model (1-11) on the basis of sparse representation model:
Formula (1-11) can be converted to (1-12):
Because of θiNot only 1 sparse vector, i.e. more than one position are non-zero, so one threshold value r of setting, value
Several positions bigger than r are made last position and are calculated, such as formula (1-13):
Since formula (1-12) is a more complicated model, direct solution is relatively difficult, it is therefore possible to use ADMM
The solution of method progress model.
It, can also further working as in site according to the client of acquisition after obtaining current location information of the client in site
Front position information analyzes the behavior of client, for example, the current location information according to client in site can analyze out visitor
Family be currently at Accreditation Waiting Area still experience area, when analyze client be currently at experience area when, product related personnel can and
When the experience area that goes to be that client carries out related introduction, and then realizes to the precision marketing of client.
In conclusion the present invention can be according to the signal of client current location and by time constraint and space about
The sparse signal representation model of beam factor optimizing is accurately positioned to client currently in the accurate location of site, and according to the visitor of acquisition
Current location information of the family in site analyzes the behavior of client, can be realized the essence to client according to behavioural analysis result
Quasi- marketing.
Each embodiment in this specification is described in a progressive manner, the highlights of each of the examples are with other
The difference of embodiment, the same or similar parts in each embodiment may refer to each other.For device disclosed in embodiment
For, since it is corresponded to the methods disclosed in the examples, so being described relatively simple, related place is said referring to method part
It is bright.
Professional further appreciates that, unit described in conjunction with the examples disclosed in the embodiments of the present disclosure
And algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware and
The interchangeability of software generally describes each exemplary composition and step according to function in the above description.These
Function is implemented in hardware or software actually, the specific application and design constraint depending on technical solution.Profession
Technical staff can use different methods to achieve the described function each specific application, but this realization is not answered
Think beyond the scope of this invention.
The step of method described in conjunction with the examples disclosed in this document or algorithm, can directly be held with hardware, processor
The combination of capable software module or the two is implemented.Software module can be placed in random access memory (RAM), memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology
In any other form of storage medium well known in field.
The foregoing description of the disclosed embodiments enables those skilled in the art to implement or use the present invention.
Various modifications to these embodiments will be readily apparent to those skilled in the art, as defined herein
General Principle can be realized in other embodiments without departing from the spirit or scope of the present invention.Therefore, of the invention
It is not intended to be limited to the embodiments shown herein, and is to fit to and the principles and novel features disclosed herein phase one
The widest scope of cause.
Claims (8)
1. a kind of localization method of the sparse signal representation model based on space-time restriction characterized by comprising
Client is obtained in the signal of site current location；
Signal based on the client in site current location, and by time constraint and space constraint factor optimizing
Sparse signal representation model obtains the client in the current location information of site.
2. the method according to claim 1, wherein further include:
Current location information of the client based on acquisition in site analyzes the behavior of client, output customer action point
Analyse result.
3. according to the method described in claim 2, it is characterized in that, described obtaining client in the signal of site current location
Before, further includes:
Collect the signal strength on the different location of site known coordinate；
Different location and corresponding signal strength based on known coordinate construct received signals fingerprint library.
4. according to the method described in claim 3, it is characterized by further comprising:
Sparse signal representation model is constructed by simulation client in the path of site based on the received signals fingerprint library；
Sparse signal representation described in time constraint and space constraint factor pair is added in the sparse signal representation model
Model optimizes, the sparse signal representation model after being optimized.
5. a kind of positioning system of the sparse signal representation model based on space-time restriction characterized by comprising
Module is obtained, for obtaining client in the signal of site current location；
Locating module for the signal based on the client in site current location, and passes through time constraint and space
The sparse signal representation model of constraint factor optimization, obtains the client in the current location information of site.
6. system according to claim 5, which is characterized in that further include:
Analysis module, the current location information for the client based on acquisition in site analyze the behavior of client,
Export customer behavior analysis result.
7. system according to claim 6, which is characterized in that further include:
Information collection module, the signal strength on different location for collecting site known coordinate；
Fingerprint base constructs module, for different location and corresponding signal strength building received signals fingerprint based on known coordinate
Library.
8. system according to claim 7, which is characterized in that further include:
Model construction module, by simulation client in the path of site, it is sparse to construct signal for being based on the received signals fingerprint library
Indicate model；
Optimized model module, for adding time constraint and space constraint factor pair in the sparse signal representation model
The sparse signal representation model optimizes, the sparse signal representation model after being optimized.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910806397.6A CN110516878A (en) | 2019-08-28 | 2019-08-28 | A kind of localization method and system of the sparse signal representation model based on space-time restriction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910806397.6A CN110516878A (en) | 2019-08-28 | 2019-08-28 | A kind of localization method and system of the sparse signal representation model based on space-time restriction |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110516878A true CN110516878A (en) | 2019-11-29 |
Family
ID=68628823
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910806397.6A Pending CN110516878A (en) | 2019-08-28 | 2019-08-28 | A kind of localization method and system of the sparse signal representation model based on space-time restriction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110516878A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060105776A1 (en) * | 2004-11-17 | 2006-05-18 | Cingular Wireless Ii, Llc | Method and system for providing location information for emergency services |
| CN103139907A (en) * | 2013-02-04 | 2013-06-05 | 北京工业大学 | Indoor wireless positioning method by utilizing fingerprint technique |
| US20130207840A1 (en) * | 2010-11-04 | 2013-08-15 | Jung Kan Mook | System and method for estimating indoor location using satellite signal generation device |
| CN104519571A (en) * | 2014-12-26 | 2015-04-15 | 北京工业大学 | Indoor positioning method based on RSS (Received Signal Strength) |
| US20160337808A1 (en) * | 2015-05-15 | 2016-11-17 | Motorola Mobility Llc | Method and Apparatus for Indoor Location Estimation Among Peer-To-Peer Devices |
| CN108668249A (en) * | 2018-07-10 | 2018-10-16 | 北京物资学院 | A kind of mobile terminal indoor orientation method and device |
| CN109461033A (en) * | 2018-11-12 | 2019-03-12 | 中国银行股份有限公司 | A kind of analysis method and device of customer action |
-
2019
- 2019-08-28 CN CN201910806397.6A patent/CN110516878A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060105776A1 (en) * | 2004-11-17 | 2006-05-18 | Cingular Wireless Ii, Llc | Method and system for providing location information for emergency services |
| US20130207840A1 (en) * | 2010-11-04 | 2013-08-15 | Jung Kan Mook | System and method for estimating indoor location using satellite signal generation device |
| CN103139907A (en) * | 2013-02-04 | 2013-06-05 | 北京工业大学 | Indoor wireless positioning method by utilizing fingerprint technique |
| CN104519571A (en) * | 2014-12-26 | 2015-04-15 | 北京工业大学 | Indoor positioning method based on RSS (Received Signal Strength) |
| US20160337808A1 (en) * | 2015-05-15 | 2016-11-17 | Motorola Mobility Llc | Method and Apparatus for Indoor Location Estimation Among Peer-To-Peer Devices |
| CN108668249A (en) * | 2018-07-10 | 2018-10-16 | 北京物资学院 | A kind of mobile terminal indoor orientation method and device |
| CN109461033A (en) * | 2018-11-12 | 2019-03-12 | 中国银行股份有限公司 | A kind of analysis method and device of customer action |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhou et al. | Adaptive genetic algorithm-aided neural network with channel state information tensor decomposition for indoor localization | |
| CN107358293B (en) | Neural network training method and device | |
| Willett et al. | Minimax optimal level-set estimation | |
| CN110458187B (en) | Malicious code family clustering method and system | |
| CN103426434B (en) | Separated by the source of independent component analysis in conjunction with source directional information | |
| CN103426437B (en) | The source using the independent component analysis utilizing mixing multivariate probability density function separates | |
| CN105844635A (en) | Sparse representation depth image reconstruction algorithm based on structure dictionary | |
| CN110443287B (en) | Crowd moving stream drawing method based on sparse trajectory data | |
| CN103116761A (en) | Dynamic texture recognition method on basis of image sequence | |
| CN110781413A (en) | Interest point determining method and device, storage medium and electronic equipment | |
| Valera et al. | Infinite factorial dynamical model | |
| CN111046737A (en) | Efficient intelligent sensing acquisition method for microseism signal detection | |
| CN106503631A (en) | A kind of population analysis method and computer equipment | |
| CN110377752A (en) | A kind of knowledge base system applied to the operation of government affairs hall | |
| CN115034261A (en) | Method and equipment for extracting features between pulses of radar radiation source signal and storage medium | |
| CN114423034A (en) | Indoor personnel action identification method, system, medium, equipment and terminal | |
| CN112884062B (en) | Motor imagery classification method and system based on CNN classification model and generated countermeasure network | |
| CN109801073A (en) | Risk subscribers recognition methods, device, computer equipment and storage medium | |
| CN109615577A (en) | High spectrum image super-resolution processing method based on convolutional network | |
| CN110516878A (en) | A kind of localization method and system of the sparse signal representation model based on space-time restriction | |
| Cao et al. | An adaptive self-organizing migration algorithm for parameter optimization of wavelet transformation | |
| CN110505573A (en) | A kind of localization method and system of the sparse signal representation model based on space constraint | |
| Voss et al. | Densities of shortest path lengths in spatial stochastic networks | |
| CN110519704A (en) | A kind of localization method and system of the sparse signal representation model based on time-constrain | |
| CN103778439A (en) | Body contour reconstruction method based on dynamic time-space information digging |
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 |
Application publication date: 20191129 |
|
| RJ01 | Rejection of invention patent application after publication |