CN103400115B - A kind of wireless signal finger print matching method - Google Patents

Abstract

The present invention discloses a kind of wireless signal finger print matching method, comprises the steps: S1, selectes a fingerprint recording F in fingerprint database_k; S2, determine user fingerprints F_uWith F_kPublic sign AP; S3, the sign energy force coefficient calculating public sign AP described in each; S4, according to the Received signal strength intensity difference of each public sign AP above-mentioned and described sign energy force coefficient, calculate F_uWith F_kMatching distance; S5, renewal k value, repeat above-mentioned steps, until obtaining user fingerprints F_uWith the matching distance of all fingerprint recordings in fingerprint database; S6, find out and user fingerprints F_uMatching distance is minimum and adjacent fingerprint and user fingerprints F_uMatching distance meet certain threshold value fingerprint as user fingerprints F_uTarget fingerprint, corresponding position is user position. The present invention utilizes delineation method for position, it is possible to effectively prevent the phenomenon of erroneous matching, it is to increase the accuracy of coupling; Utilize Shandong bar gauge to be about technology and solve the not high problem of fingerprint matching precision that the unstable due to AP causes.

Description

A kind of wireless signal finger print matching method

Technical field

The present invention relates to wireless indoor field of locating technology, particularly relate to a kind of wireless signal finger print matching method being applied to indoor accurate position.

Background technology

The large-scale promotion that is extensively universal and wireless device of smart mobile phone has expedited the emergence of the research located in a large number about wireless indoor. The indoor orientation method of great majority utilizes existing wireless access point (AccessPoint, AP) (access point such as wireless technologys such as WiFi, GSM, ZigBee) in environment to realize location at present. Easily acquired due to Received signal strength intensity (ReceivedSignalStrength, RSS), it becomes the most frequently used physical features. Received signal strength intensity for each AP can directly read usually Wireless Communication Equipment. Wireless signal is adopted to become, as the method for position feature, the indoor orientation method widely used most, also known as based on the indoor positioning of fingerprint or fingerprint location.

Wireless fingerprint localization method is divided into two benches usually: training and operation. In the training stage, it is necessary to localizing environment to be carried out on-the-spot exploration. In this process, slip-stick artist sets up a fingerprint database by each position record RSS fingerprint in the region of interest. Record in a database together with relative position by fingerprint. In the operation phase, when user sends the request of a position enquiring, RSS fingerprint that location algorithm utilizes this user current retrieval fingerprint database also returns the record of optimum matching, using the location estimation of the position of this record correspondence as user.

Fingerprint matching algorithm directly concerns positioning precision. Under the indoor environment of complexity, propagate impacts such as being subject to multipath, refraction, reflection due to wireless signal, add the dynamic of environment, instability, spatially unique problem in wireless signal fingerprint lifetime.Specifically, mainly show as:

1) isospace position may detect similar fingerprint. In other words, same bar user fingerprints may all have higher similarity with the fingerprint on mutually so far apart two positions in space, thus the location estimation giving user fingerprints brings uncertainty.

2) unstable of AP, Received signal strength intensity presents fluctuation by a small margin within the short period of time, presents and significantly change in for a long time. And the stability out that different AP shows is different.

Existing received signals fingerprint coupling technology does not all consider above-mentioned factor, the coupling that often makes the mistake or the problem such as matching precision is lower.

Summary of the invention

(1) technical problem solved

Technical problem to be solved by this invention is: how to provide a kind of wireless signal finger print matching method, solves and mates accuracy and the not high problem of precision in prior art.

(2) technical scheme

In order to solve the problems of the technologies described above, on the one hand, the present invention provides a kind of wireless signal finger print matching method, comprises the steps:

S1, in fingerprint database a selected fingerprint recording F_k;

S2, determine user fingerprints F_uWith F_kEach public sign AP;

S3, the sign energy force coefficient calculating public sign AP described in each;

S4, according to the Received signal strength intensity difference of each public sign AP above-mentioned and described sign energy force coefficient, calculate F_uWith F_kMatching distance;

S5, renewal k value, repeat above-mentioned steps, until obtaining user fingerprints F_uWith the matching distance of all fingerprint recordings in fingerprint database;

S6, find out and user fingerprints F_uEurope formula apart from minimum and adjacent fingerprint and user fingerprints F_uMatching distance meet certain threshold value fingerprint as user fingerprints F_uTarget fingerprint, corresponding position is user position;

The Received signal strength intensity difference of each public sign AP described is labeled as Δ F=[x_i, i=1,2 ..., Q], wherein Q is the number of public sign AP.

Preferably, step is also comprised after step S3:

S3B, to described F_uWith F_kThe Received signal strength intensity difference of corresponding public sign AP carries out stipulations.

Specifically, the specific implementation step of described S3B is:

S3B1, according to least square intermediate value theorem, to described Δ F generate regression estimates value;

The Received signal strength intensity difference of S3B2, the public sign AP that judges each whether outlier;

S3B3, the regression estimates value obtained with step S3B1 substitute the Received signal strength intensity difference of the public sign AP of corresponding exception.

Specifically, described step S3B2 judges whether outlier is specially for the Received signal strength intensity difference of each public sign AP:

Judge the Received signal strength intensity difference x of each public sign AP_i(i ∈ 1,2 ..., and Q}) whether corresponding estimation residual error exceed the threshold value of setting, if exceeding this threshold value, judges x_iFor outlier; Otherwise it is judged to non-outlier.

Specifically, described step S3B2 judges whether outlier is specially for the Received signal strength intensity difference of each public sign AP:

According to x_i(i ∈ 1,2 ..., Q}) resistance diagnosis index judge Received signal strength intensity difference x_iWhether it is outlier. x_iResistance diagnosis index RD_iIt is specially:

${\mathrm{RD}}_i=\frac{x_i}{{\mathrm{med}}_{j=1,...,Q}{x}_j}$

Wherein, med represents median. Work as RD_iDuring < 2.5, x_iDo not belong to outlier; Work as RD_iWhen >=2.5, x_iBelong to outlier.

Further, F in described step S2_uWith F_kPublic sign AP particular by perform following step obtain:

S21, delete F respectively_uWith F_kMiddle Received signal strength intensity is less than the AP of setting threshold value, the F obtained_u' and F_k', wherein F_uThe set of the AP of ' correspondence isF_kThe set of the AP of ' correspondence is

S22, rightWithSeek common ground, orderRepresent the number of AP in occuring simultaneously, if P=0, export fingerprint F_uWith described fingerprint F_kDo not mate, upgrade k value, return to step S1;

If P ≠ 0, perform S23;

S23, rightWithSeek union, and all AP concentrated are exactly user fingerprints F_uWith described F_kPublic sign AP.

Further, the sign energy force coefficient concrete steps calculating public sign AP described in each in described step S3 are as follows:

S31, Received signal strength intensity with described public sign AP represent F again_uWith F_k;

S32, public sign AP and the F that calculates each_uPhysical distance D_u=[d_u,q, q=1,2 ..., Q], d_u,qRepresent q public sign AP and F_uPhysical distance;

Calculate each public sign AP and F_kPhysical distance D_k=[d_k,q, q=1,2 ..., Q], d_k,qRepresent q public sign AP and F_kPhysical distance;

S33, the d obtained based on step S32_u,qAnd d_k,q, q=1,2 ..., Q, calculates q public sign AP respectively for user fingerprints F_uWith F_kSign energy force coefficient ρ_u,qWith ρ_k,qAs follows:

${\mathrm{\&rho;}}_{t,q}=\frac{1}{d_{t,q}{\mathrm{\&Sigma;}}_t}$

Wherein, ${\mathrm{\&Sigma;}}_t={\mathrm{\&Sigma;}}_{j=1}^Q\frac{1}{d_{t,j}},$ T ∈ { u, k};

S34, the ρ getting step S33 and obtaining_u,qAnd ρ_k,qIn bigger value, as the q F_uWith F_kThe sign energy force coefficient ρ of public sign AP_q, q=1,2 ..., Q.

Further, in described step S32, step calculates d_u,qOr d_k,qDetailed process be:

Known according to logarithm distance path fading model (Log-normalDistancePathLoss):

PL(d_q)=PL₀+10γlg(d_q)

Wherein, PL₀Representing that the physical distance with q public sign AP is Received signal strength intensity when 1 meter, γ is path attenuation index, PL (d_q) represent that the physical distance with q public sign AP is d_u,qTime Received signal strength intensity, PL (d_q) can from F_uOr F_kIn directly obtain;

Separate above-mentioned equation and can obtain d_u,qOr d_k,qValue.

Specifically, F in described step S4_uWith F_kThe matching distance of matching distance be calculated as follows:

$\mathrm{\&Phi;}(F_u,F_k)=\frac{Q}{P}\sqrt{{\mathrm{\&Sigma;}}_{q=1}^Q{\left({\mathrm{\&rho;}}_q{x}_q\right)}^2}$

Wherein, Q represents F_uWith F_kThe number of public sign AP, P represents F_uWith F_kMiddle Received signal strength intensity is more than or equal to the number of the common factor of the AP of setting threshold value, ρ_qRepresent F_uWith F_kThe sign energy force coefficient of the public sign AP of q, x_qRepresent F_uWith F_kThe difference of Received signal strength intensity of the public sign AP of q.

Concrete, threshold value described in described step S6 is the maximum value of matching distance between the fingerprint on wantonly two consecutive positions in fingerprint database.

Further, the adjacent fingerprint in described step S6 refers to: the fingerprint that in described fingerprint database, the neighbouring sample point of record is corresponding;

Preferably, adjacent fingerprint chooses the fingerprint of the neighbouring sample point of east, west, south, north four direction.

(3) useful effect

Wireless signal finger print matching method described in the present invention proposes the concept of delineation position, and the fingerprint mated with user fingerprints determined in fingerprint database, the matching distance of its adjacent fingerprint and user fingerprints also to be met some requirements. The phenomenon of erroneous matching is prevented, it is to increase the accuracy of coupling with this.

In fingerprint matching process, the method described in the present invention, does not use whole AP, and the public sign AP only selecting sign ability strong reduces the complexity of coupling process. Utilizing Shandong bar gauge to be about technology, to have screened out between to be matched two fingerprint abnormal RSS poor, thus solves the problem that fingerprint matching precision that the unstable due to AP causes is not high.

Meanwhile, the method, without any need for extra hardware device, is also surveyed or locating terminal without the need to the scene of revising, it is possible to integrated use direct with all kinds of methods general at present, has very high universality.

After reading the detailed explanation of embodiment of the present invention by reference to the accompanying drawings, other advantage and disadvantages of the present invention will become clearly.

Accompanying drawing explanation

Fig. 1 is the steps flow chart schematic diagram of wireless signal finger print matching method in one embodiment of the present invention;

Fig. 2 is one embodiment of the present invention centre circle positioning principle schematic diagram;

The difference in signal strength of adding of abnormal public sign AP is carried out Shandong bar gauge and is about schematic diagram in embodiments of the invention two by Fig. 3.

Embodiment

Below in conjunction with Figure of description and embodiment, the specific embodiment of the present invention is described in further detail. Following examples are only for illustration of the present invention, but are not used for limiting the scope of the invention.

For the ease of describing and the specific embodiment of the present invention is described, the Received signal strength intensity difference of public sign AP refers in the examples below: in be matched two fingerprints, the difference of the Received signal strength intensity of the same public sign AP recorded respectively; AP and fingerprint F_kPhysical distance refer to: the distance between physical location corresponding to the position at AP place and this fingerprint.

Embodiment one

As shown in Figure 1, the present embodiment describes a kind of wireless signal finger print matching method, comprises the steps:

S1, at fingerprint databaseIn a selected fingerprint recording F_k;

Concrete, indoor wireless networks is made up of multiple wireless AP usually, by surveying described AP Received signal strength intensity at indoor scene, it is possible to obtain the fingerprint database in localizing objects region. E.g., exploration process is: first select sampling point (every two meters arrange a sampling) according to certain density according to grid-like fashion; Secondly, each sampling point records the Received signal strength intensity of each AP, as the wireless signal fingerprint of position corresponding to this sampling point, it is stored in fingerprint database. That is, the every bar fingerprint in fingerprint database all correspond to a physical location label accurately.

Assume that locating area has N number of wireless AP and is respectively { AP₁,AP₂,...,AP_N, fingerprint F_kCan represent and be: F_k=[f_k,1,f_k,2,...,f_k,N], wherein f_k,n(1≤n≤N) represents F_kRepresented sampling point is to AP_nReceived signal strength intensity, if sampling point does not receive the signal of a certain AP, to the Received signal strength intensity of AP being designated as-100dBm.

S2, determine user fingerprints F_uWith above-mentioned F_kPublic sign AP, calculate the sign energy force coefficient of each public sign AP;

Usually, in a fingerprint, the Received signal strength intensity of AP is more strong, illustrates that this AP is the closer to the position represented by this fingerprint. At the present embodiment, do not use the whole AP in fingerprint, and only select the AP that Received signal strength intensity is stronger can effectively reduce the complexity calculating matching distance and being as public sign AP.

Finger print information when user sends position enquiring request is designated as F_u=[f_u,1,f_u,2,...,f_u,N], it is determined that user fingerprints F_uWith described F_kPublic sign AP specifically comprise the following steps:

S21, respectively deletion user fingerprints F_uWith described F_kMiddle Received signal strength intensity is less than the AP of setting threshold value σ, the F obtained_u' and F_k', wherein F_uThe set of the AP of ' correspondence isF_kThe set of the AP of ' correspondence is

The value of described σ can obtain according to empirical measurements, usually elects-95dBm as;

S22, rightWithSeek common ground, orderRepresent the number of AP in occuring simultaneously, if P=0, export fingerprint F_uWith described fingerprint F_kDo not mate, upgrade k value, return to step S1;

If P ≠ 0, perform S23;

S23, rightWithSeek union, note Wherein Q≤N, In all wireless AP be exactly user fingerprints F_uWith described F_kPublic sign AP.

The sign energy force coefficient of S3, the public sign AP that calculates each;

By the AP more near with a certain positional distance, the sign energy force coefficient of this position is more big.The representing by distance of the sign energy force coefficient of AP in the present embodiment, it is specially:

S31, to above-mentioned F_u' and F_k' expand, make F_u'=[f_u′_,1,f_u′_,2,...,f_u′_,Q], wherein f_u′_,q(1≤q≤Q) represents F_uRepresented physical location is to AP_q' Received signal strength intensity, ifThen remember f_u′_,q=-100, unit is dBm;

Similar, make F_k'=[f_k′_,1,f_k′_,2,...,f_k′_,Q], wherein f_k′_,q(1≤q≤Q) represents F_kRepresented sampling point is to AP_q' Received signal strength intensity, ifThen remember f_k′_,q=-100, unit is dBm;

Make x_q=| f_u′_,q-f_k′_,q|, q=1,2 ..., Q, represents that public sign AP is at F_uWith F_kIn Received signal strength intensity difference, note Δ F=[x₁,x₂,...,x_Q]。

S32, calculatingIn the position at each AP place and F_uThe physical distance of represented position.

Concrete, AP_q' (1≤q≤Q) is to F_uThe distance d of represented position_u,qCan calculate according to logarithm distance path fading model (Log-normalDistancePathLoss):

PL(d_u,q)=PL₀+10γlg(d_u,q)

Wherein, PL₀Represent and AP_q' distance be Received signal strength intensity when 1 meter, γ is path attenuation index, PL (d_u,q) represent and AP_q' distance be d_u,qTime Received signal strength intensity, PL (d in the present embodiment_u,q)=f_u′_,q。

With reason, can calculateIn the position at each AP place and F_kThe physical distance of represented position.

S33, the d obtained based on above-mentioned steps_u,qCalculate AP_q' (q=1,2 ..., Q) for user fingerprints F_u'=[f_u′_,1,f_u′_,2,...,f_u′_,Q] sign energy force coefficient as follows:

${\mathrm{\&rho;}}_{u,q}=\frac{1}{d_{u,q}{\mathrm{\&Sigma;}}_u}$

Wherein,It is a normalized parameter so that F_u'=[f_u′_,1,f_u′_,2,...,f_u′_,Q] in the sign energy force coefficient summation of each AP

With reason, AP can be calculated_q' (q=1,2 ..., Q) for user fingerprints F_k'=[f_k′_,1,f_k′_,2,...,f_k′_,Q] sign energy force coefficient ρ_k,q, q=1,2 ..., Q.

S34, the ρ getting step S33 and obtaining_u,qAnd ρ_k,qIn bigger value, as the q F_uWith F_kThe sign energy force coefficient of public sign AP, i.e. ρ_q=max{ ρ_u,q,ρ_k,q, q=1,2 ..., Q.

S4, calculating F_uWith F_kMatching distance;

Based on above-mentioned steps, F_uWith F_kThe matching distance of matching distance be calculated as follows:

$\mathrm{\&Phi;}(F_u,F_k)=\frac{Q}{P}\sqrt{{\mathrm{\&Sigma;}}_{q=1}^Q{\left({\mathrm{\&rho;}}_q{x}_q\right)}^2}$

S5, renewal k, repeat above-mentioned steps, until obtaining user fingerprints F_uWith the matching distance of all fingerprint recordings in fingerprint database;

S6, find out and user fingerprints F_uMatching distance is minimum and surrounding fingerprint and user fingerprints F_uMatching distance meet certain threshold value fingerprint as user fingerprints F_uTarget fingerprint, corresponding position is user position.

On the basis of above-mentioned steps, the present embodiment introduces the concept of " delineation position ", improves positioning precision. As shown in Figure 2, delineation position refers to, does not use the mode that single-point mates, (mating one by one with the fingerprint in fingerprint database by user fingerprints, find the most similar target fingerprint), and adopt multiple spot to mate in fingerprint matching process. Multiple spot mates for single-point, not only to be ensured that user fingerprints is similar to target fingerprint, to be ensured that the fingerprint that user fingerprints is corresponding with target fingerprint consecutive position is also comparatively close simultaneously when referring to coupling. Specifically, given user fingerprints F_u, to each fingerprint F in fingerprint database_k, remember that the fingerprint of sampling point on four direction around it is F_k1, F_k2, F_k3And F_k4, Φ (F can be obtained according to above-mentioned steps_u,F_x), wherein F_x∈{F_k,F_k1,F_k2,F_k3,F_k4,

When

And

$\&ForAll;F_x\&Element;\{F_{k1},F_{k2},F_{k3},F_{k4}\},\mathrm{\&Phi;}(F_u,F_x)<\mathrm{\&epsiv;}$

Time, selected F^*For target fingerprint, F^*Corresponding position is target location. Wherein, ε is adjacent fingerprint similarity threshold, usually gets fingerprint databaseIn the maximum value of matching distance between fingerprint on any two consecutive positions.

If the target fingerprint meeting above condition does not exist, then judge that user is not within the locating area scope corresponding to current fingerprint database, cannot return to the position of user.

In this step of the present embodiment, for each fingerprint F in fingerprint database_k, around it, on four direction, the fingerprint of sampling point can be sampling point adjacent on the four direction of east, south, west, north, it is also possible to be sampling point adjacent on the southeast, northeast, southwest, northwest four direction. Further, the fingerprint of four direction around can not also be got in other embodiments of the invention, change into and get the nearest fingerprint of adjacent fingerprint middle distance, or only get the fingerprint of left and right adjacent sampling point or only get the fingerprint of forward and backward adjacent sampling point. The technician of this area is it should be appreciated that can also have other modes to choosing of surrounding sample points fingerprint, it should also be appreciated that for belonging to invention which is intended to be protected.

Embodiment two

On the basis of embodiment one, also comprise the following steps before step S4 after step S3:

S3B, Received signal strength intensity difference to abnormal public sign AP carry out Shandong bar gauge about.

Owing to the stability of each AP is different, identical sampling point is surveyed by different time, and the sampling fingerprint obtained may be different. Especially, if the stability of some AP is very poor, so to the exploration the possibility of result of this AP Received signal strength intensity, there will be two kinds of situations: one is at same sampling point, there is bigger fluctuation within the short period of time in the Received signal strength intensity of this AP, another kind is position two employing points closely, the Received signal strength intensity of this AP surveyed but has very big-difference, and the matching result of user fingerprints either way can be produced a very large impact by this. The AP of this kind of stability difference is just called abnormal AP.

If abnormal AP is as the public sign AP of fingerprint to be matched, abnormal Received signal strength intensity difference will be produced, and then the matching distance of fingerprint to be matched is caused relatively big error. In order to eliminate the disadvantageous effect that abnormal AP causes, match user fingerprint accurately, it is necessary to the abnormal Received signal strength intensity difference that abnormal AP causes carries out stipulations. The present embodiment adopts Shandong bar gauge to be about technology and is revised by abnormal Received signal strength intensity difference.

As shown in Figure 3, bar gauge effect about in Shandong is when most RSS difference is all less, the great outlier of minority difference can be adjusted to less estimated value, and when most RSS difference is all bigger, the minimum outlier of minority difference is then adjusted to bigger estimated value. Bar gauge concrete steps about in Shandong are as follows:

S3B1, to received signal intensity difference vector Δ F=[x₁,x₂,...,x_Q] generate a regression estimates amountObtain to Δ F one group of estimation $\mathrm{\&Delta;}\hat{F}=[{\hat{x}}_1,{\hat{x}}_2,...,{\hat{x}}_Q],$ Wherein ${\hat{x}}_i=x_i\&CenterDot;\mathrm{\&theta;}+e,$ I=1,2 ..., Q.

According to least square intermediate value theorem, the optimum target of setting regression estimates is:

$\underset{\widehat{\mathrm{\&theta;}}}{\min }\underset{i}{\mathrm{med}}{r_i}^2$

Wherein med represents median,For estimating residual error;

There are θ value and e value that above-mentioned optimization formula can separate, thus obtainI=1,2 ..., Q.

S3B2, abnormal Received signal strength intensity difference is revised;

Judge that whether Received signal strength intensity difference is abnormal, it is possible to adopt following two kinds of modes:

One judges each Received signal strength intensity difference x_i(i ∈ 1,2 ..., and Q}) corresponding estimation residual error r_iWhether exceeding the threshold value of setting, if exceeding this threshold value, judging x_iFor outlier; Otherwise it is judged to non-outlier.

Another kind adopts resistance diagnosis (ResistantDiagnostic) index to judge Received signal strength intensity difference x_i(i ∈ 1,2 ..., and Q}) whether it is outlier.X_iResistance diagnosis index RD_iIt is specially:

${\mathrm{RD}}_i=\frac{x_i}{{\mathrm{med}}_{j=1,...,Q}{x}_j}$

Wherein, med represents median. Work as RD_iDuring < 2.5, x_iDo not belong to outlier; Work as RD_iWhen >=2.5, x_iBelong to outlier.

S3B3, according to above-mentioned two kinds of decision methods, the Received signal strength intensity difference after note stipulations is, i ∈ 1,2 ..., Q}, so

Step S4 in embodiment one changes into: based on the Received signal strength intensity difference after stipulations, and the AP calculated in conjunction with previous step characterizes energy force coefficient, F_uWith F_kThe matching distance of matching distance be calculated as follows:

$\mathrm{\&Phi;}(F_u,F_k)=\frac{Q}{P}\sqrt{{\mathrm{\&Sigma;}}_{q=1}^Q{\left({\mathrm{\&rho;}}_q{\tilde{x}}_q\right)}^2}$

Other steps in embodiment 1 are constant.

The Received signal strength intensity difference of AP is carried out stipulations, it is possible to eliminate abnormal AP to the impact of matching result, greatly improve matching precision.

The above is only the preferred embodiment of the present invention; it is noted that for those skilled in the art, under the prerequisite not departing from the technology of the present invention principle; can also making some improvement and modification, these improve and modification also should be considered as protection scope of the present invention.

Claims (9)

1. a wireless signal finger print matching method, it is characterised in that, comprise the following steps:

S1, in fingerprint database a selected fingerprint recording F_k;

S2, determine user fingerprints F_uWith F_kEach public sign AP;

S3, the sign energy force coefficient calculating public sign AP described in each;

S4, according to the Received signal strength intensity difference of each public sign AP above-mentioned and described sign energy force coefficient, calculate F_uWith F_kMatching distance;

S5, renewal k value, repeat above-mentioned steps, until obtaining user fingerprints F_uWith the matching distance of all fingerprint recordings in fingerprint database;

S6, find out and user fingerprints F_uMatching distance is minimum and adjacent fingerprint and user fingerprints F_uMatching distance meet certain threshold value fingerprint as user fingerprints F_uTarget fingerprint, corresponding position is user position;

The Received signal strength intensity difference of each public sign AP described is labeled as Δ F=[x_i, i=1,2 ..., Q], wherein Q is the number of public sign AP;

Wherein, F in described step S2_uWith F_kPublic sign AP particular by perform following step obtain:

S21, choose F respectively_uWith F_kMiddle Received signal strength intensity is more than or equal to the AP of setting threshold value, F_uThe set of corresponding AP is A_u, F_kThe set of corresponding AP is A_k;

S22, to A_uAnd A_kSeek common ground, make the number that P represents AP in common factor, if P=0, export fingerprint F_uWith described fingerprint F_kDo not mate, upgrade k value, return to step S1;

If P ≠ 0, perform S23;

S23, to A_uAnd A_kSeek union, and all AP concentrated are exactly F_uWith F_kPublic sign AP;

The sign energy force coefficient concrete steps calculating public sign AP described in each in described step S3 are as follows:

S31, Received signal strength intensity with described public sign AP represent F again_uWith F_k;

S32, public sign AP and the F that calculates each_uPhysical distance D_u=[d_u,q, q=1,2 ..., Q], d_u,qRepresent q public sign AP and F_uPhysical distance;

Calculate each public sign AP and F_kPhysical distance D_k=[d_k,q, q=1,2 ..., Q], d_k,qRepresent q public sign AP and F_kPhysical distance;

S33, the d obtained based on step S32_u,qAnd d_k,q, q=1,2 ..., Q, calculates q public sign AP respectively for user fingerprints F_uWith F_kSign energy force coefficient ρ_u,qWith ρ_k,qAs follows:

${\mathrm{\&rho;}}_{t,q}=\frac{1}{d_{t,q}{\mathrm{\&Sigma;}}_t}$

Wherein, ${\mathrm{\&Sigma;}}_t={\mathrm{\&Sigma;}}_{j=1}^Q\frac{1}{d_{t,j}},$ T ∈ { u, k};

S34, the ρ getting step S33 and obtaining_u,qAnd ρ_k,qIn bigger value, as the q F_uWith F_kThe sign energy force coefficient ρ of public sign AP_q, q=1,2 ..., Q.

2. the method for claim 1, it is characterised in that, also comprise step after step S3:

S3B, to described F_uWith F_kThe Received signal strength intensity difference of each public sign AP carry out stipulations.

3. method as claimed in claim 2, it is characterised in that, the specific implementation step of described S3B is:

S3B1, according to least square intermediate value theorem, to described Δ F generate regression estimates value;

The Received signal strength intensity difference of S3B2, the public sign AP that judges each whether outlier;

S3B3, the regression estimates value obtained with step S3B1 substitute the Received signal strength intensity difference of the public sign AP of corresponding exception.

4. method as claimed in claim 3, it is characterised in that, described step S3B2 judges whether outlier is specially for the Received signal strength intensity difference of each public sign AP:

Judge the Received signal strength intensity difference x of each public sign AP_i(i ∈ 1,2 ..., and Q}) whether corresponding estimation residual error exceed the threshold value of setting, if exceeding this threshold value, judges x_iFor outlier; Otherwise it is judged to non-outlier.

5. method as claimed in claim 3, it is characterised in that, described step S3B2 judges whether outlier is specially for the Received signal strength intensity difference of each public sign AP:

According to x_i(i ∈ 1,2 ..., Q}) resistance diagnosis index judge Received signal strength intensity difference x_iWhether it is outlier, x_iResistance diagnosis index RD_iIt is specially:

${\mathrm{RD}}_i=\frac{x_i}{{\mathrm{med}}_{j=1,...,Q}{x}_j}$

Wherein, med represents median, works as RD_i< when 2.5, x_iDo not belong to outlier; Work as RD_iWhen >=2.5, x_iBelong to outlier.

6. the method for claim 1, it is characterised in that, in described step S32, step calculates d_u,qOr d_k,qDetailed process be:

According to logarithm distance path fading model:

PL(d_q)=PL₀+10γlg(d_q)

Wherein, PL₀Representing that the physical distance with q public sign AP is Received signal strength intensity when 1 meter, γ is path attenuation index, PL (d_q) represent that the physical distance with q public sign AP is d_u,qTime Received signal strength intensity, PL (d_q) can from F_uOr F_kIn directly obtain;

Separate above-mentioned equation and can obtain d_u,qOr d_k,qValue.

7. method as according to any one of claim 1 to 6, it is characterised in that, F in described step S4_uWith F_kMatching distance be calculated as follows:

$\mathrm{\&Phi;}(F_u,F_k)=\frac{Q}{P}\sqrt{{\mathrm{\&Sigma;}}_{q=1}^Q{\left({\mathrm{\&rho;}}_q{x}_q\right)}^2}$

Wherein, P represents F_uWith F_kMiddle Received signal strength intensity is more than or equal to the number of the common factor of the AP of setting threshold value, ρ_qRepresent F_uWith F_kThe sign energy force coefficient of the public sign AP of q, x_qRepresent F_uWith F_kThe difference of Received signal strength intensity of the public sign AP of q.

8. method as according to any one of claim 1 to 6, it is characterised in that, threshold value described in described step S6 is the maximum value of matching distance between the fingerprint on wantonly two consecutive positions in fingerprint database.

9. method as according to any one of claim 1 to 6, it is characterised in that, the adjacent fingerprint in described step S6 refers to: the fingerprint that in described fingerprint database, the neighbouring sample point of record is corresponding;

Adjacent fingerprint chooses the fingerprint of the neighbouring sample point of east, west, south, north four direction.