CN105548958A - Fine grain multi-target passive positioning method on the basis of RFID - Google Patents

Abstract

The present invention discloses a fine-grain multi-target passive positioning method on the basis of a RFID. The method provided by the invention is able to take the disturbance of multipath information as fingerprint information, and a multi-target positioning method in the prior art remove multipath or neglect multipath information, so that the method provided by the invention is able to realize the calculation of fine grain position, and other methods can only indicate the substantial position of an object. The method provided by the invention is able to accurately perform fine grain multi-target positioning and fill up defects of the multi-target positioning algorithm in the aspect of positioning precision, and has good dynamic adaptability of the environment and low economic cost.

Description

A kind of fine granularity multiple goal passive type localization method based on RFID

Technical field

The invention belongs to wireless positioning field, relate to a kind of fine granularity multi-target orientation method, be specifically related to a kind of fine granularity multiple goal passive type localization method based on RFID.

Background technology

In a lot of application scenarios now, as anti-intrusion detections, mall navigation, Falls Among Old People detection etc., the positional information of acquisition target is most basic content, has attracted large quantities of scholar to carry out correlative study work in these in the past 20 years.Wherein, passive type (without carrying other equipment) localization method has important effect, especially for multiobject localization method in the application of countering intrusions always.

Existing passive type localization method uses sensor node, infrared facility, senior camera or even specialized equipment etc., and these equipment prices are higher, are difficult to meet large scale deployment demand.The favor that current RFID passive label is located because of cheap (one 5 ~ 10 cents), but the label deployment density used based on the passive type localization method of RFID is at present excessive, and cost is higher, and can not be applicable to multiobject situation.RFID (RadioFrequencyIdentification) technology, refers to unlimited REID.

In current multiple goal passive type localization method, all only consider the impact of people on los path (LoS) signal, and have ignored the situation of multipath in reality, must multipath conditions be avoided even as far as possible.In fact, the situation of multipath is unescapable, and as shown in Figure 1, people can not only affect the los path signal of transmitting terminal to receiving end, and other path signals also can affect.Therefore these methods can only point out the Position Approximate of target, accurately can not calculate target location, can not meet the demand of people.In addition, existing localization method mainly adopts RSSI technology, but RSSI technology is because of by the impact of antenna gain, can pin down by multipath, thus the location of coarseness can only be reached.

To sum up, the following defect of existing passive type localization method ubiquity: 1) multiobject location can not be realized in monitored area; 2) equipment is not cheap; 3) precision can not meet people's demand.

The list of references quoted:

[1]J.Wang,B.Xie,D.Fang,L.C.Chen,Xiaojiang,T.Xing,andW.Nie.Accuratedevice-freelocalizationwithlittlehumancost.InProc.ACMMobiComWorkshop.2015.

[2]D.Zhang,Y.Liu,X.Guo,andL.M.Ni.RASS:Areal-time,accurate,andscalablesystemfortrackingtransceiver-freeobjects.IEEETrans.ParallelandDistributedSystems,24(5):996-1008,2013.

[3]K.Joshi,D.Bharadia,M.Kotaru,andS.Katti.WiDeo:Fine-graineddevice-freemotiontracingusingrfbackscatter.InProc.USENIXNSDI,pages189-204,2015.

Summary of the invention

For the defect existed in above-mentioned prior art or deficiency, the object of the invention is to, a kind of fine granularity multiple goal passive type localization method based on RFID is provided, the method can carry out fine-grained Multi-target position accurately in surveillance area, fill up the defect of Multi-target position algorithm in positioning precision, better to the dynamic adaptable of environment, and financial cost is low.

In order to realize above-mentioned task, the present invention adopts following technical scheme to be achieved:

Based on a fine granularity multiple goal passive type localization method of RFID, specifically comprise the following steps:

Step one, disposes RFID passive label and multiple rfid interrogator in monitored area;

Step 2, obtains off-line phase finger print information and also stores, the multi-path information of label that off-line phase finger print information comprises the target location be in off-line phase monitored area, the label ID be affected and is affected;

Step 3, obtain on-line stage corresponding with each rfid interrogator be affected tag set, determine multiplely to be affected label subregion according to being affected tag set, and then determine target numbers to be detected; Obtain the online finger print information that each target to be detected is corresponding;

Step 4, the online finger print information that the target each to be detected that the off-line finger print information utilizing step 2 to obtain and step 3 obtain is corresponding, calculates the positional information of target to be detected.

Particularly, the specific implementation of described step 2 comprises:

Step 2.1:RFID read write line is provided with two, be respectively X and Y, two rfid interrogators send signal simultaneously, the aerial array of each rfid interrogator is measured and is recorded the phase value of all labels when not having target, and utilizes the phase value obtained to obtain the multi-path information of each rfid interrogator to all labels by SAR method;

Step 2.2: guarded region is divided into the square net that size is identical, the people selecting physique medium stand on each grid, obtains the multi-path information of each rfid interrogator to all labels;

Step 2.3: adopt DTW method and consider the impact that the multi-path information change of system noise on label produces, step 2.1 is obtained there is no a target time multi-path information and multi-path information during target that has that obtains of step 2.2 compare, obtain by the label of each object effects respectively; Using target location, be stored in server database as off-line phase finger print information by the label ID of object effects, the multi-path information being affected label and influence degree.

Particularly, the specific implementation of described step 3 comprises:

Step 3.1: several people to be detected enter surveyed area, after static, two rfid interrogators, read the phase value of all labels, and adopt SAR method to calculate the multi-path information of two rfid interrogators to each label respectively, finally these information are stored in server database;

Step 3.2: two rfid interrogators have different routing informations to each label, namely has two set: set one is the multi-path information of first rfid interrogator to all labels, and set two is that second rfid interrogator is to all label multi-path information; Utilize the DTW algorithm in step 2.3 and consider the impact that the multi-path information change of system noise on label produces, label multi-path information in set one and set two is compared with the label multi-path information in no one's situation respectively, obtains two and be affected tag set;

Step 3.3: for each rfid interrogator, be affected in tag set corresponding with it has several to be affected label subregion, calculates target numbers according to subregion number; For each target, the multi-path information that be affected label ID and these labels corresponding with it forms the online finger print information of this target.

Particularly, the specific implementation of described step 4 comprises:

The label of each object effects is different, then between each target without any relation, so the account form of each target location is identical with process, the computing formula of a target location is as follows:

M(R,F)＝Nλ _match-d _R(R,F)+n _lack·W _lack(9)

$d_R(R,F)=\frac{\sqrt{{{\mathrm{SLT}}_1}^2+{{\mathrm{SLT}}_2}^2+\mathrm{...}+{{\mathrm{SLT}}_N}^2}}{N}---\left(10\right)$

$L_{t\arg et}=\underset{L}{\arg \max }\left(M(R,F_i)\right),i\∈(1,...,N_{rf})---\left(11\right)$

In formula:

R represents the finger print information that on-line stage calculates, and comprises the multi-path information of label ID that each target affects and these labels, a corresponding one group of online finger print information in target location;

F represents all target information set that off-line phase calculates, namely target location, be affected label ID and be affected the multi-path information of label;

What M (R, F) represented the finger print information of the finger print information of on-line stage and one group of off-line phase mates mark, and coupling mark is higher, and both explanations are more close;

N represents the same label number of on-line stage and each group off-line phase;

λ _matchrepresent identical and be affected the weight of label in the matching process of position;

D _r(R, F) represents that the finger print information of on-line stage and one group of off-line phase finger print information are in all identical comprehensive differences be affected on the multi-path information of label;

SLT represents identical and is affected the difference of label on multi-path information;

N _lackrepresent same antenna signal in off-line phase and on-line stage incomplete be affected number of labels;

W _lackrepresent the weight that aerial signal is incomplete;

L _targetrepresent for each target to be detected, calculate M (R, the F) value of online fingerprint corresponding to this target to be detected and all off-line fingerprints respectively, obtain maximum M (R, F) the off-line fingerprint that value is corresponding, then the target location in this off-line fingerprint is required position.

Particularly, obtain in described step 2.3 being comprised by the specific implementation of the label of each object effects:

Suppose have two multi-path information files, i.e. B _αand B _β,

B _α＝B _α(0),…B _α(i),…B _α(180)(2)

B _β＝B _β(0),…B _β(i),…B _β(180)(3)

Have L=180 value in each multi-path information file, then the difference value of these two multi-path information is

${\mathrm{SLT}}_{\mathrm{\α},\mathrm{\β}}=min\underset{i,j=1}{\overset{L}{\Σ}}|B_{\mathrm{\α}}\left(i\right)-B_j\left(j\right)|---\left(4\right)$

In computation process, i needs not be equal to j, so the difference of two multi-path information is smallest match value;

Because system noise also can cause the multi-path information of label to change, in order to judge that multi-path information change is impact due to noise or the impact due to target: suppose that α was that a upper moment is for judging the threshold value whether label multi-path information is affected by people, SLT _newrepresent current record judge the threshold value whether label multipath is affected by people, m represents threshold deviation, then first calculate the difference diff of present threshold value and previous moment threshold value,

diff＝SLT _new-α(5)

Use current time threshold value to upgrade α and m, obtain

α＝(1-0.125)·α+0.125·SLT _new(6)

m＝(1-0.25)·m+0.25·|diff|(7)

Wherein, 0.125 and 0.25 is up-to-date threshold value and the proportion shared by deviation, and abbreviation formula (6) and (7), draw,

α＝α+0.125·diff(8)

m＝m+0.25·(|diff|-m)(9)

Thus up-to-date threshold value is,

Th＝α+4m(10)

If record the multi-path information difference value SLT of label when having people and no one _{α, β}be greater than threshold value Th, then think that this label is for being affected label.

The present invention compared with prior art, has the following advantages:

1, localization method of the present invention only needs a small amount of cheap RFID passive label just can locate multiple unknown object, and namely required financial cost is low, breaches in traditional sense and uses expensive device could realize multiobject location.

2, localization method of the present invention using people to the disturbance of multi-path information as finger print information, then remove multipath or ignore multi-path information in multi-target orientation method at present, so the present invention can realize fine-grained position calculation, and additive method can only point out the approximate location of target.

3, the present invention adopts selectivity matching strategy, if current online fingerprint and certain off-line fingerprint do not have identically to be affected label, then without the need to coupling, reduces time complexity to a great extent;

4, the present invention ingenious utilize rfid interrogator and label in communication process target block the situation that the signal that causes cannot receive, instead of try every possible means and avoid this situation, there is universality and practicality.

5, localization method of the present invention is very good to the dynamic adaptable of environment, therefore, it is possible to well adapt to the dynamic change of application scenarios, this localization method has certain universality.

Accompanying drawing explanation

Fig. 1 is a multi-path information key diagram.

Fig. 2 is the principle key diagram positioned according to multi-path information.

Fig. 3 is positioning flow figure of the present invention.

Fig. 4 is backscatter communication schematic diagram.

Four examples in Fig. 5 object count process.Fig. 5 (a) for target range far away time method of counting schematic diagram; Fig. 5 (b) is the schematic diagram occurring vanishing target situation under a rfid interrogator; Fig. 5 (c) is the schematic diagram eliminating vanishing target situation under two rfid interrogators; Fig. 5 (d) for target range nearer time schematic diagram.

Fig. 6 is 4 targets, the true experiment scene graph under 5*5 label matrix.

Fig. 7 is counting precision and the positioning error of true experiment under 480cm*480cm, 600cm*600cm, 720cm*600cm monitored area.

Fig. 8 is counting precision under 480cm*480cm-1.2m, 720cm*600cm-1.2m, 480cm*480cm-0.6m, 720cm*600cm-0.6m truly tests and positioning error, and wherein 0.6m and 1.2m is label spacing.

When Fig. 9 is in sighting distance situation, the positioning error of the present invention and Alico, RASS, WiDeo contrasts, and wherein Fig. 9 (a) is positioning error CDF figure, and Fig. 9 (b) is deployment diagram.

When Figure 10 is in non line of sight situation, the positioning error of the present invention and Alico, RASS, WiDeo contrasts, and wherein Figure 10 (a) is positioning error CDF figure, and Figure 10 (b) is deployment diagram.

Below in conjunction with drawings and Examples particular content of the present invention explained in further detail and illustrate.

Embodiment

See Fig. 1, multi-path information always exists, and between transmitting terminal to receiving end, there is a more than bars travel path, and the quantity of information that every paths is propagated is different, then many transmission paths situation is called as multi-path information.Appearance due to people can affect the signal on multiple path, thus the change of rear multi-path information appears in description people, can judge the position of people.

In order to provide cardinal principle of the present invention more clearly, first with a simple example, principle of the present invention is described, as shown in Figure 2.Dispose schematic diagram 2 (a) display, a rfid interrogator and 12 RFID passive labels are deployed in surveyed area.Multi-path information Fig. 2 (b) display label Tag1 unmanned, people in G1 position, people in G2 position time multi-path information graphic differences very large, so people is different at diverse location on the impact being affected label.In addition, as can be seen from Fig. 2 (c), people is at same position G1, and the multi-path information of different label is different.Therefore we can carry out target localization according to target to the influence degree of the multi-path information of different label.The fine granularity multiple goal passive type localization method that the present invention is based on RFID utilizes multi-path information and fingerprint matching to come realize target location, and only utilize cheap RFID passive label, is a kind of not by the new definition method inquired in wireless positioning field.

Overview ground describes the localization method in the present invention: the multiple goal passive type localization method based on RFID proposed in the present invention mainly comprises two stages: destination number calculates and fine granularity location, as shown in Figure 3.

In the object count stage, whether the label multi-path information that the present invention adopts DTW algorithm to detect in monitored area is affected by people, and continuous print is affected label is divided into subregion, different targets has and different is affected label subregion, then two rfid interrogators and Edge detected length is used to eliminate vanishing target situation, divide and thinner be affected label subregion, the number being finally affected label subregion is object count, thus realizes multiple goal counting.

At fine granularity positioning stage, first, the finger print information of each target that the present invention collects according to on-line stage, namely label and the multi-path information being affected label is affected, finger print informations all for off-line phase is divided into two set, in one of them set, each finger print information has at least one identical to be affected label compared with current on-line stage finger print information.Secondly, for each off-line phase finger print information in above-mentioned set, calculate their identical comprehensive differences being affected label with current on-line stage finger print information respectively.Finally, calculating off-line phase finger print information mates mark with on-line stage finger print information, and mark is high, and represent that this off-line finger print information mates with current online finger print information most, then the position in off-line finger print information is the position of current goal.

Fine granularity multiple goal passive type localization method based on RFID of the present invention, specifically comprises the following steps:

Step one, scene setting

RFID passive label is disposed in monitored area, adjacent label is with 1.2m spacing composition label matrix, and 2 rfid interrogators are positioned over label matrix arbitrary neighborhood both sides, each RFID reader connects one group of aerial array, the omnidirectional antenna that aerial array is 0.15m by 4 spacing forms.

Step 2, obtains off-line phase finger print information and stores

Step 2.1:2 rfid interrogator sends signal simultaneously, every root antenna receives the signal that each tag reflection is returned, the aerial array of each rfid interrogator is measured and is recorded the phase value of all labels when not having target, and utilizes the phase value obtained to obtain the multi-path information of each rfid interrogator to all labels by SAR (syntheticapertureradar) method.

Wherein, the computing method of described all label phase values are as follows: rfid system adopts backscatter communication mode, as shown in Figure 4, rfid interrogator uses four omnidirectional antennas to continue to send radio-frequency information, label receives radiofrequency signal and reflects back, and after four antennas of rfid interrogator receive reflected signal, is calculated that the label ID that each antenna receives numbers, phase value θ by rfid interrogator, signal strength values RSSI etc., phase value θ computing formula is:

$\left\{\begin{array}{c}\mathrm{\θ}=(\frac{2\mathrm{\π}}{\mathrm{\λ}}\×2d+n)\mathrm{mod}2\mathrm{\π},\\ n={\mathrm{\θ}}_T+{\mathrm{\θ}}_R+{\mathrm{\θ}}_{tag},\end{array}\right.---\left(1\right)$

Wherein, the wavelength of λ radiofrequency signal, n is system noise, θ _t, θ _r, θ _tagbe respectively the reflection coefficient of the radiating circuit of rfid interrogator, the receiving circuit of rfid interrogator and label.

Owing to there is system noise in phase value, then the multi-path information calculated also has deviation, and performance phenomenon is that the multi-path information under Same Scene in different time has random fluctuation.

Step 2.2: guarded region is divided into the square net that size is identical, sizing grid is 0.4m.The people selecting physique medium stand on each grid, obtains the multi-path information of each rfid interrogator to all labels.Now, when someone stand on grid, the acquisition methods of the multi-path information of all labels is identical with the acquisition methods of the multi-path information of all labels when nobody stand on grid in step 2.1.

Step 2.3: adopt DTW (DynamicTimeWarping) method and consider the impact that the multi-path information change of system noise on label produces, step 2.1 is obtained there is no a target time multi-path information and multi-path information during target that has that obtains of step 2.2 compare, obtain respectively by the label of each object effects, specific implementation is as follows:

Suppose have two multi-path information files, i.e. B _αand B _β,

B _α＝B _α(0),…B _α(i),…B _α(180)(2)

B _β＝B _β(0),…B _β(i),…B _β(180)(3)

Have L=180 value in each multi-path information file, then the difference value of these two multi-path information is

${\mathrm{SLT}}_{\mathrm{\α},\mathrm{\β}}=min\underset{i,j=1}{\overset{L}{\Σ}}|B_{\mathrm{\α}}\left(i\right)-B_j\left(j\right)|---\left(4\right)$

In computation process, i needs not be equal to j, so the difference of two multi-path information is smallest match value.

Because system noise also can cause the multi-path information of label to change, in order to judge that multi-path information change is impact due to noise or the impact due to target, the present invention uses for reference the retransmission timeout timer in congestion control, a real-time threshold value is set, for judging that multi-path information change is that noise or target cause under current environment.Concrete methods of realizing is:

Suppose that α was that a upper moment is for judging the threshold value whether label multi-path information is affected by people, SLT _newrepresent current record judge the threshold value whether label multipath is affected by people, m represents threshold deviation, then first calculate the difference diff of present threshold value and previous moment threshold value,

diff＝SLT _new-α(5)

Then use current time threshold value to upgrade α and m, obtain

α＝(1-0.125)·α+0.125·SLT _new(6)

m＝(1-0.25)·m+0.25·|diff|(7)

Wherein, 0.125 and 0.25 is up-to-date threshold value and the proportion shared by deviation, and abbreviation formula (6) and (7), draw,

α＝α+0.125·diff(8)

m＝m+0.25·(|diff|-m)(9)

Thus up-to-date threshold value is,

Th＝α+4m(10)

Therefore, if record the multi-path information difference value SLT of label when having people and no one _{α, β}be greater than threshold value Th, then think that this label is for being affected label.

Now, target location, the label ID be affected, the multi-path information of label be affected and influence degree are stored in server database as off-line phase finger print information.

Step 3, obtain on-line stage corresponding with each rfid interrogator be affected tag set, determine multiplely to be affected label subregion according to being affected tag set, and then determine target numbers to be detected; Obtain the online finger print information of each target to be detected simultaneously.Concrete methods of realizing comprises:

Step 3.1: several people to be detected enter surveyed area, after static, use two rfid interrogator X and Y, read the phase value of all labels, and adopt SAR method to calculate the multi-path information of two rfid interrogators to each label respectively, finally these information are stored in server database.

Step 3.2: two rfid interrogators have different routing informations to each label, namely has two set: set one is the multi-path information of first rfid interrogator to all labels, and set two is that second rfid interrogator is to all label multi-path information; Set one and the label multi-path information gathered in two compare with the label multi-path information in no one's situation respectively, utilize the DTW algorithm in step 2.3 and consider the impact that the multi-path information change of system noise on label produces, obtaining two and be affected tag set.

Step 3.3: for each rfid interrogator, be affected in tag set corresponding with it has several to be affected label subregion, and subregion number is target numbers to be detected.

As shown in Fig. 5 (a), for rfid interrogator X, what there are two separation is affected label subregion, then represent existence two targets to be detected.

But there will be certain in real process and be affected the excessive special circumstances of label subregion, namely the number of labels in this subregion is too much, experiment shows, if be affected number of labels to be greater than upper limit R=6, then may there is the situation of vanishing target in this region, as having two targets in subregion I in Fig. 5 (b), now, need to use another one rfid interrogator Y, as shown in Fig. 5 (c), original subregion I can be divided into subregion I and subregion IV by rfid interrogator Y, and such vanishing target can be detected.

If rfid interrogator Y can not detect, as shown in Fig. 5 (d), rfid interrogator can not detect in subregion I two targets, and now need to calculate the longest length of side D of this subregion, then destination number computation process is as follows:

D＝d _tag·(n _tag-1)(11)

$N_{t\arg et}=\frac{D}{1.2}---\left(12\right)$

Wherein, d _tagrepresent the spacing of adjacent two labels; n _tagrepresent the number of labels in subregion longest edge length; N _targetrepresent destination number.

Said process obtain target numbers to be detected and each target to be detected corresponding be affected label subregion, for each target, corresponding with it multi-path information being affected label ID and these labels forms the online finger print information of this target.

Principle uses rfid interrogator more, the number of vanishing target is fewer.But use two or more rfid interrogators can produce read write line conflict simultaneously, cause the acquisition phase value time longer, can not real-time requirement be met.Reason is that passive label processing power is more weak, when label receives the signal of two or more rfid interrogators simultaneously, cannot process, and can only be used as abnormal signal and discard, and can require that the long period obtains the phase value of same quantity like this.The present invention weighs time cost and rfid interrogator quantity, adopts two rfid interrogators to carry out object count.Meanwhile, for reducing the time cost conflicting and bring, the present invention controls the emission signal frequency of two RFID reader, makes in same time, obtain phase value as much as possible, and concrete grammar is as follows:

The present invention uses for reference ALOHA agreement, and this agreement is a kind of effective anticollision protocol.It is Poisson process that ALOHA agreement sends signal depending on transmitting terminal within the unit interval, and tag reflection signal is as Poisson process in the present invention.After label filters out collision signal, rfid interrogator is returned in the signal reflex do not conflicted, then rfid interrogator success sampling rate F _wfor

F _w＝W·P(13)

Wherein, W is by the probability of label success reflected signal, and P is defined as

P＝e ^-2W(14)

Then rfid interrogator success sampling rate F _wcan be expressed as

F _w＝W·e ^-2W(15)

In addition, number of labels is n, and in the unit interval, average each label success reflected signal is λ, then can be expressed as by the probability W of label success reflected signal

W＝n·λ(16)

Thus, rfid interrogator success sampling rate F _wfor

F _w＝nλ·e ^-2nλ(17)

Theoretical according to data sampling, and if only if when sampling interval S meets certain condition, and the data of collection could reconstruct target distribution, then satisfy condition into

$S=\frac{1}{F_w}---\left(18\right)$

In fact, S is defined as

S＝n·e·lnn+r(19)

Due to the random number that r is between 1 ~ n, then

$n\&CenterDot;e\&CenterDot;\ln n<\frac{e^{2n\mathrm{\&lambda;}}}{n\mathrm{\&lambda;}}---\left(20\right)$

Therefore, the present invention, by controlling the transmission frequency of RFID reader, adjusts suitable λ, just can obtain maximum phase values within the unit interval, reduce the time of real-time counting as far as possible.

Step 4, fine granularity Multi-target position

Utilize the off-line finger print information obtained in the online finger print information and step 2 obtained in step 3, find out the off-line finger print information mated most with the online finger print information of target, the positional information of this target can be calculated.Each target location, because the label of each object effects is different, then between each target without any relation, so the account form of each target location is identical with process.

First off-line finger print information classifies by the present invention, extracts only containing the off-line finger print information that be affected label identical with online finger print information, mates for next stage.Because the off-line finger print information after extraction is far fewer than all off-line finger print informations, thus greatly reduce match time.Secondly, DTW algorithm is used to calculate the online finger print information multi-path information similarity SLT that be affected label identical with in off-line finger print information.Finally, use matching formula, obtain the coupling mark of online finger print information R and off-line finger print information F respectively.In fact, people's blocks the phase value that rfid interrogator can be caused can not to read some label, namely can not construct the multi-path information of this label, cannot carry out target localization.Matching formula considers this phenomenon, but not avoids this phenomenon, and concrete matching formula is as follows:

M(R,F)＝Nλ _match-d _R(R,F)+n _lack·W _lack(21)

Wherein, the finger print information of the finger print information that M (R, F) is on-line stage and off-line phase mate mark, coupling mark is higher, and both are more close; N represents the same label number of on-line stage and off-line phase; λ _matchrepresent identical and be affected the weight of label in the matching process of position; n _lackrepresent same antenna signal in off-line phase and on-line stage incomplete be affected number of labels; W _lackrepresent the weight that aerial signal is incomplete; d _r(R, F) is defined as follows:

$d_R(R,F)=\frac{\sqrt{{{\mathrm{SLT}}_1}^2+{{\mathrm{SLT}}_2}^2+\mathrm{...}+{{\mathrm{SLT}}_N}^2}}{N}---\left(22\right)$

SLT represents identical and is affected the similarity of label on multi-path information.

Thus this target location is the position of target in the off-line fingerprint of the highest coupling mark,

$L_{t\arg et}=\underset{L}{\arg \max }\left(M(R,F_i)\right),i\&Element;(1,...,N_{rf})---\left(23\right)$

L _targetrepresent for each target to be detected, calculate M (R, the F) value of online fingerprint corresponding to this target to be detected and all off-line fingerprints respectively, obtain maximum M (R, F) the off-line fingerprint that value is corresponding, then the target location in this off-line fingerprint is required position.

Embodiment:

Use two general rfid systems to complete test in the laboratory of Northwest University's information science and technical college, be placed with multiple desk, shelf, desktop computer etc. in laboratory, form complicated indoor environment.In this experiment, always have three kinds of different experiments, for verifying validity of the present invention, robustness and the advantage compared with additive method, often kind of experiment has different experiment scenes, completed by 7 volunteers, continue 14 hours altogether, collect 1054400 data.

The present embodiment adopt experiment I first scene in, in scene one, be made up of the label matrix of 5*5 25 RFID label tag, often pair of label be spaced apart 1.2m, thus formation 480cm*480cm monitored area.Two rfid interrogators and corresponding aerial array thereof place the adjacent both sides of monitored area.In this scene, always have 4 volunteers, as shown in Figure 6.

The fine granularity multiple goal passive type localization method based on RFID of the present embodiment, specifically carries out according to following steps:

Step one, scene setting

RFID passive label is disposed in monitored area, adjacent label is with 1.2m spacing composition label matrix, and 2 rfid interrogators are positioned over label matrix arbitrary neighborhood both sides, each rfid interrogator connects one group of aerial array, the omnidirectional antenna that aerial array is 0.15m by 4 spacing forms.

Label matrix is that 5 row 5 arrange, and label E PC is numbered 0001 ~ 0025, arranged in sequence.The rfid system communication distance of passive type is limited, is generally 10 ~ 20 meters, though here label spacing is 1.2m, for the communication range ability of rfid system, has been reduce lower deployment cost as much as possible.In aerial array, antenna distance is 0.15m, corresponding with the programming generating multi-path information.

Step 2, obtains off-line phase finger print information and stores

Step 2.1:2 rfid interrogator sends signal simultaneously, every root antenna receives the signal that each tag reflection is returned, the aerial array of each rfid interrogator is measured and is recorded the phase value of all labels when not having target, and utilizes the phase value obtained to obtain the multi-path information of each reader to all labels by SAR (syntheticapertureradar) method.

Step 2.2: guarded region is divided into the square net that size is identical, sizing grid is 0.4m.Before the people selecting physique medium stand on each grid, obtain the multi-path information of all labels relative to each reader.

Step 2.3: adopt DTW (DynamicTimeWarping) method and consider the impact that the multi-path information change of system noise on label produces, step 2.1 is obtained there is no a target time multi-path information and multi-path information during target that has that obtains of step 2.2 compare, obtain respectively by the label of each object effects, specific implementation is as follows:

But because neighbourhood noise also can cause the multi-path information of label to change, so need the label multi-path information change that the suitable threshold value of setting one causes for judging people or neighbourhood noise.Concrete threshold calculations process is as follows:

diff＝SLT _new-α(24)

α＝(1-0.125)·α+0.125·SLT _new(25)

m＝(1-0.25)·m+0.25·|diff|(26)

Thus draw,

α＝α+0.125·diff(27)

m＝m+0.25·(|diff|-m)(28)

Up-to-date threshold value is,

Th＝α+4m(29)

According to above-mentioned calculating, the scene lower threshold value Th obtaining the present embodiment is 0.4161.

If record the multi-path information difference value SLT of label when having people and no one _{α, β}be greater than threshold value Th, then think that this label is for being affected label.

In order to reduce error as far as possible, the medium people of physique is selected to obtain the finger print information of off-line phase.Because during on-line stage location, the build of multiple target is different, if people's difference on build that target now and off-line phase use is too large, larger positioning error can be caused.Off-line phase finger print information form is: { position (x, y), is affected label (tag1, tag2 ...), label multi-path information (tag1.dat, tag2.dat ...).In addition, monitored area is 0.48m*0.48m, and sizing grid is 0.4m, then always have 12*12=144 bar off-line finger print information.

Step 3, what acquisition on-line stage was corresponding with each rfid interrogator is affected tag set, determines that several are affected label subregion, and then determine target numbers to be detected according to being affected tag set; Obtain the online finger print information of each target to be detected simultaneously.Concrete methods of realizing comprises:

Step 3.1:4 people to be detected enters surveyed area, after static, use two rfid interrogator X and Y, read the phase value of all labels, and adopt SAR method to calculate the multi-path information of two rfid interrogators to each label respectively, finally these information are stored in server database.

Step 3.2: two rfid interrogators have different routing informations to each label, namely has two set: set one is the multi-path information of first rfid interrogator to all labels, and set two is that second rfid interrogator is to all label multi-path information; Set one and the label multi-path information gathered in two compare with the label multi-path information in no one's situation respectively, utilize the DTW algorithm in step 2.3 and consider the impact that the multi-path information change of system noise on label produces, obtaining two and be affected tag set.

According to above-mentioned analysis and statistics, the tag set that is affected obtaining first rfid interrogator corresponding is

{0001,0002,0006,0007,00011,0012,0016,0019,0021,0022,0004,0005,0009,0010,0014,0015,0019,0020,0024,0024}，

The tag set that is affected that second rfid interrogator is corresponding is { 0004,0005,0003,0002,0001,0006,0007,0008,0009,00010,00016,0017,0018,0019,0020,0021,0022,0023,0024,0025}.

Step 3.3: for each rfid interrogator, be affected in tag set corresponding with it has several to be affected label subregion, and subregion number is target numbers to be detected.

What first rfid interrogator X was corresponding be affected tag set can be divided into two separated subregions, are namely affected label subregion I{0001,0002,0006,0007,00011,0012,0016,0019,0021,0022}, is affected label subregion II{0004,0005,0009,0010,0014,0015,0019,0020,0024,0024}.

For rfid interrogator X, two separated subregions indicate two targets, but are all greater than 6 due to the label number that is affected of every sub regions, then every sub regions all exists and is hidden target.Now, can be calculated second rfid interrogator corresponding be affected label subregion III{0004,0005,0003,0002,0001,0006,0007,0008,0009,00010}, subregion IV{00016,0017,0018,0019,0020,0021,0022,0023,0024,0025}.Now subregion I with II can be divided into four intersecting areas be separated, i.e. A{0001,0002,0006,0007}, B{0021,0022,0016,0017}, C{0004,0005,0009,0010}, D{0019,0020,0025,0024} by subregion III with IV.Label number in each intersecting area is all less than 6, then do not exist and be hidden label, now target number is the number of intersecting area, is 4; And the label in A ~ D tetra-intersecting areas be each target be affected label, corresponding label multi-path information can extract from any one RFID reader.

Step 4, fine granularity Multi-target position

The computing formula of a target location is as follows:

M(R,F)＝Nλ _match-d _R(R,F)+n _lack·W _lack(30)

$d_R(R,F)=\frac{\sqrt{{{\mathrm{SLT}}_1}^2+{{\mathrm{SLT}}_2}^2+\mathrm{...}+{{\mathrm{SLT}}_N}^2}}{N}---\left(31\right)$

$L_{t\arg et}=\underset{L}{\arg \max }\left(M(R,F_i)\right),i\&Element;(1,...,N_{rf})---\left(32\right)$

According to above-mentioned calculating, 4 target locations obtained in this example are F7 (1,0.2), F32 (4.2,0.6), and F114 (0.6,4.6), F143 (4.6,4.4.2), wherein true origin is 0001 label position.Fn represents the station location marker in off-line fingerprint, and n represents sequence number, always has 144 off-line finger print informations in this scene, and F7 represents that position (1,0.2) is the position in the 7th article of off-line finger print information.

Locating effect is verified:

Experiment I:

The target of experiment I is to verify validity of the present invention.First, set up 3 monitored areas varied in size, i.e. 480cm*480cm, 600cm*600cm, 720cm*600cm, target numbers, number of tags and matrix information are as shown in table 1.In experimentation, gathering phase information corresponding to each label 200, therefrom selecting to occur that the value of maximum probability is for calculating multi-path information.Interval simultaneously between Offered target is 0.3m ~ 1.5m, and label spacing is 1.2m.

Table 1 experiment parameter is arranged

The test effect of experiment I:

1) object count precision: under calculating different target spacer conditions, obtain different object count precision.Experimental result is as shown in Fig. 7 (a), and along with the increase of target distance, the counting precision of three monitored areas all constantly promotes, and when target distance is greater than 1m, object count precision is 100%.The present invention regards the target of multiple hypotelorism as a target, but this shortcoming of the present invention can be tolerated, because of in problem ubiquity positioning system (such as infrared ray) for this reason.

2) target location accuracy: Fig. 7 (b) provides positioning error cumulative distribution (CDF) curve of three kinds of different monitored areas situations, and positioning error is all less than 0.42m, and namely the present invention can realize fine-grained Multi-target position.

Experiment II:

The target of experiment II is to verify robustness of the present invention.This experiment is carry out in the monitored area of 480cm*480cm, 720cm*720cm in size.Interval between Offered target is 0.3m ~ 1.5m, and label spacing is 0.6m, 1.2m.Target numbers, number of tags and matrix information are as shown in table 1.

The test effect of experiment II:

1) object count precision: as shown in Fig. 8 (a), 480*480-1.2m and 720*600-1.2m is illustrated respectively in the counting precision of two monitored areas when label is spaced apart 1.2m, similarly, 480*480-0.6m and 720*600-0.6m is illustrated respectively in the counting precision of two monitored areas when label is spaced apart 0.6m.Can draw from figure, reduce label spacing and improve counting precision not obvious, therefore, object count does not have mandatory requirement in label spacing, so the present invention has robustness in object count.

2) target location accuracy: Fig. 8 (a) provides positioning error cumulative distribution (CDF) curve of two kinds of different monitored areas situations.Can draw from figure, reducing label spacing can not significantly improve positioning precision, therefore, is positioned at label spacing aspect and does not have mandatory requirement, so the present invention has robustness in Multi-target position.

Experiment III:

In experiment III, the present invention contrasts Alice [1], RASS [2] and WiDeo [3] positioning error under sighting distance, nlos environment.Monitoring of environmental size is 720cm*720cm, and destination number is 7, interval 1.2m between label.

The test effect of experiment III:

Fig. 9 (a) and Figure 10 (a) is respectively the positioning error CDF curve in sighting distance and non line of sight situation, and FISCP of the present invention is far superior to other three kinds of methods as can be seen from Figure.In sighting distance situation, positioning precision of the present invention is 3.11 times, 5.24 times, 6.6 times of Alico, RASS, WiDeo respectively; In non line of sight situation, positioning precision of the present invention is 4.55 times, 7.61 times, 13.3 times of Alico, RASS, WiDeo respectively.Visible, additive method is work efficiency drop 1.5 times ~ 2 times in non line of sight situation.And the error of the present invention under sighting distance and non line of sight situation is all less than 0.45m, average error is 0.33m; And other three kinds of methods in sighting distance situation 80% error can only arrive 1.31m, 2.2m, 2.78m.

Claims (5)

1., based on a fine granularity multiple goal passive type localization method of RFID, it is characterized in that, specifically comprise the following steps:

Step one, disposes RFID passive label and multiple rfid interrogator in monitored area;

Step 2, obtains off-line phase finger print information and also stores, the multi-path information of label that off-line phase finger print information comprises the target location be in off-line phase monitored area, the label ID be affected and is affected;

Step 3, obtain on-line stage corresponding with each rfid interrogator be affected tag set, determine multiplely to be affected label subregion according to being affected tag set, and then determine target numbers to be detected; Obtain the online finger print information that each target to be detected is corresponding;

Step 4, the online finger print information that the target each to be detected that the off-line finger print information utilizing step 2 to obtain and step 3 obtain is corresponding, calculates the positional information of target to be detected.

2., as claimed in claim 1 based on the fine granularity multiple goal passive type localization method of RFID, it is characterized in that, the specific implementation of described step 2 comprises:

Step 2.1:RFID read write line is provided with two, be respectively X and Y, two rfid interrogators send signal simultaneously, the aerial array of each rfid interrogator is measured and is recorded the phase value of all labels when not having target, and utilizes the phase value obtained to obtain the multi-path information of each rfid interrogator to all labels by SAR method;

Step 2.2: guarded region is divided into the square net that size is identical, the people selecting physique medium stand on each grid, obtains the multi-path information of each rfid interrogator to all labels;

Step 2.3: adopt DTW method and consider the impact that the multi-path information change of system noise on label produces, step 2.1 is obtained there is no a target time multi-path information and multi-path information during target that has that obtains of step 2.2 compare, obtain by the label of each object effects respectively; Using target location, be stored in server database as off-line phase finger print information by the label ID of object effects, the multi-path information being affected label and influence degree.

3., as claimed in claim 2 based on the fine granularity multiple goal passive type localization method of RFID, it is characterized in that, the specific implementation of described step 3 comprises:

Step 3.1: several people to be detected enter surveyed area, after static, two rfid interrogators, read the phase value of all labels, and adopt SAR method to calculate the multi-path information of two rfid interrogators to each label respectively, finally these information are stored in server database;

Step 3.2: two rfid interrogators have different routing informations to each label, namely has two set: set one is the multi-path information of first rfid interrogator to all labels, and set two is that second rfid interrogator is to all label multi-path information; Utilize the DTW algorithm in step 2.3 and consider the impact that the multi-path information change of system noise on label produces, label multi-path information in set one and set two is compared with the label multi-path information in no one's situation respectively, obtains two and be affected tag set;

Step 3.3: for each rfid interrogator, be affected in tag set corresponding with it has several to be affected label subregion, calculates target numbers according to subregion number; For each target, the multi-path information that be affected label ID and these labels corresponding with it forms the online finger print information of this target.

4., as claimed in claim 3 based on the fine granularity multiple goal passive type localization method of RFID, it is characterized in that, the specific implementation of described step 4 comprises:

The label of each object effects is different, then between each target without any relation, so the account form of each target location is identical with process, the computing formula of a target location is as follows:

M(R,F)＝Nλ _match-d _R(R,F)+n _lack·W _lack(9)

$d_R(R,F)=\frac{\sqrt{{{\mathrm{SLT}}_1}^2+{{\mathrm{SLT}}_2}^2+\mathrm{...}+{{\mathrm{SLT}}_N}^2}}{N}---\left(10\right)$

$L_{t\arg et}=\underset{L}{\mathrm{argmax}}\left(M\right(R,F_i\left)\right),i\&Element;(1,...,N_{rf})---\left(11\right)$

In formula:

R represents the finger print information that on-line stage calculates, and comprises the multi-path information of label ID that each target affects and these labels, a corresponding one group of online finger print information in target location;

F represents all target information set that off-line phase calculates, namely target location, be affected label ID and be affected the multi-path information of label;

What M (R, F) represented the finger print information of the finger print information of on-line stage and one group of off-line phase mates mark, and coupling mark is higher, and both explanations are more close;

N represents the same label number of on-line stage and each group off-line phase;

λ _matchrepresent identical and be affected the weight of label in the matching process of position;

D _r(R, F) represents that the finger print information of on-line stage and one group of off-line phase finger print information are in all identical comprehensive differences be affected on the multi-path information of label;

SLT represents identical and is affected the difference of label on multi-path information;

N _lackrepresent same antenna signal in off-line phase and on-line stage incomplete be affected number of labels;

W _lackrepresent the weight that aerial signal is incomplete;

L _targetrepresent for each target to be detected, calculate M (R, the F) value of online fingerprint corresponding to this target to be detected and all off-line fingerprints respectively, obtain maximum M (R, F) the off-line fingerprint that value is corresponding, then the target location in this off-line fingerprint is required position.

5., as claimed in claim 2 based on the fine granularity multiple goal passive type localization method of RFID, it is characterized in that, obtain in described step 2.3 being comprised by the specific implementation of the label of each object effects:

Suppose have two multi-path information files, i.e. B _αand B _β,

B _α＝B _α(0),…B _α(i),…B _α(180)(2)

B _β＝B _β(0),…B _β(i),…B _β(180)(3)

Have L=180 value in each multi-path information file, then the difference value of these two multi-path information is

${\mathrm{SLT}}_{\mathrm{\&alpha;},\mathrm{\&beta;}}=min\underset{i,j=1}{\overset{L}{\&Sigma;}}|B_{\mathrm{\&alpha;}}\left(i\right)-B_j\left(j\right)|---\left(4\right)$

In computation process, i needs not be equal to j, so the difference of two multi-path information is smallest match value;

Because system noise also can cause the multi-path information of label to change, in order to judge that multi-path information change is impact due to noise or the impact due to target: suppose that α was that a upper moment is for judging the threshold value whether label multi-path information is affected by people, SLT _newrepresent current record judge the threshold value whether label multipath is affected by people, m represents threshold deviation, then first calculate the difference diff of present threshold value and previous moment threshold value,

diff＝SLT _new-α(5)

Use current time threshold value to upgrade α and m, obtain

α＝(1-0.125)·α+0.125·SLT _new(6)

m＝(1-0.25)·m+0.25·|diff|(7)

Wherein, 0.125 and 0.25 is up-to-date threshold value and the proportion shared by deviation, and abbreviation formula (6) and (7), draw,

α＝α+0.125·diff(8)

m＝m+0.25·(|diff|-m)(9)

Thus up-to-date threshold value is,

Th＝α+4m(10)

If record the multi-path information difference value SLT of label when having people and no one _{α, β}be greater than threshold value Th, then think that this label is for being affected label.