CN104977561A - Positioning algorithm based on virtual tags and cyclic check - Google Patents

Abstract

The invention discloses an indoor positioning algorithm based on virtual tags and cyclic check. The positioning algorithm comprises steps of dividing a positioning region; using a radio signal loss model; adopting K virtual reference tags away from a to-be-positioned tag; and correcting a positioning result based on cyclic check. The positioning algorithm is improved based on a VIRE algorithm. The method of the positioning algorithm comprises steps of introducing the radio signal loss model, and obtaining a RSSI value of a virtual tag according to an RSSI value of an introduced tag and a distance between the virtual tag and the introduced tag; using a K nearest neighbor method to calculate the position of the to-be-positioned tag; calculating the virtual position of a reference tag in reverse using cyclic iteration check; and performing position correction according to a difference between the virtual position and the practical position of the reference tag. The invention is mainly used for detecting, tracking and controlling the position and information of an object, and solves problems that the practical position of a to-be-positioned tag influences positioning precision and an obstacle in the positioning environment increases positioning errors in VIRE.

Description

A kind of location algorithm based on virtual label and cyclic check

Technical field

The present invention relates to indoor positioning field, particularly relate to a kind of location algorithm based on virtual label and cyclic check.

Background technology

For many industries, as communications and transportation, manufacture, the application of logistics and health care standard and personalization, positional information is a vital information on services.It is also to provide basis that is personalized and location Based service simultaneously.Positioning system can be divided into indoor and outdoors to locate according to the difference of localizing environment.Under the promotion that the E-911 proposed in US Federal Communication Committee (FCC) requires, the comparatively morning of outdoor positioning algorithm starting, rapidly, achievement is obvious in development.Comparatively speaking, the proposition of indoor positioning algorithms is more late, and be very important field, development prospect is large.In recent years, REID (RadioFrequency Identification, RFID) social various fields start application, to improve people quality of life, improve Business Economic Benefit, enhancing public security and improving Social Informatization Level creates material impact.Current RFID technique has been widely used in the every field such as monitoring, false proof, safety, gate inhibition, logistics, military affairs.

RFID localization method can be divided into: distance estimations method, scene analysis method, leash law.Based in the indoor positioning of RFID, scene analysis method have more again extensibility strong, comparatively complex environment condition, positioning precision comparatively high can be adapted to.Its principle of work is: in reader coverage, passive electronic label receives the radiofrequency signal that reader sends, then the information stored in induction acquisition energy loading electronic tag is on electromagnetic wave, be sent back on reader, for active electronic label, launch radiofrequency signal with information on one's own initiative to reader, after reader decodes, read information, be sent to microprocessor and carry out information data process, complete and identify electronic tag coordinate information.RFID indoor positioning mainly comprises SpotON positioning system, LANDMARC algorithm, VIRE location algorithm etc.

In conjunction with associated materials, by analysis, mainly there is following problem in existing indoor positioning algorithms:

1., in former LANDMARC system, electronic tag launches a radiofrequency signal in every 7.5 seconds, and this long interval result in the delay of location.

2., for the signal intensity of electronic tag, reader provide only 8 corresponding energy levels, instead of concrete intensity level, therefore in positioning precision, there is certain error.

3. location depends on reference label, if will improve positioning precision, needs the density increasing reference label.Not only increase system cost like this, and between electronic tag, signal disturbing can make positioning precision reduce on the contrary.

4. the physical location of label to be positioned also can have influence on positioning precision.When label to be positioned is among numerous reference label, positioning error is less; And when label to be positioned is in the edge of reference label coverage, error is then relatively much bigger.

5. the barrier in localizing environment is also the factor causing error larger.When having barrier between reader and electronic tag, the label signal that reader receives can become less than normal value because of electromagnetic reflection; When there being mobile object between reader and electronic tag, RSSI value also can be undergone mutation; Even sometimes there will be the situation that RSSI value is lost.

Summary of the invention

Based on the problems referred to above, the present invention introduces a kind of indoor positioning algorithms based on virtual label and cyclic check, object is the RSSI value being calculated virtual reference label by a kind of more reasonably mode, by reducing the impact of environment radio frequency signal and then improving positioning precision.The last technology adopting iteration correction again, the result that innovatory algorithm calculates carries out correcting process, makes result more accurate.Solve the delay that long interval causes locating, reduce the dependence to reference label density, reduce due to electronic tag signal intensity, the comparatively big error existed in the location that in label physical location to be positioned and localizing environment, barrier difference causes.

The present invention proposes a kind of location algorithm based on virtual label and cyclic check, concrete grammar:

A. locating area is divided into a plane grid, makes entity reference label be evenly scattered in various piece in grid, laterally identical with the interval of longitudinal direction;

B. divide plane grid with the virtual grid of m*n, get the coordinate position of each virtual net center of a lattice as a reference label, represent the virtual reference label covering thus of this virtual grid;

C. by logarithm distance loss model formula

$P=P_0+10\mathrm{nlg}\left(\frac{d_{i,j}}{d_0}\right)+\mathrm{\ξ}$

Push away

$n=\frac{p_x-p_y}{10\lg \left(\frac{d_x}{d_y}\right)}$

$\mathrm{\ξ}=P-P_0-10\mathrm{nlg}\left(\frac{d_{i,j}}{d_0}\right)$

D. RSSI value and the distance of choosing adjacent four labels substitute into above-mentioned formula, try to achieve 3 results;

E. average as the loss model parameter in this region;

F. the RSSI value of parameter and a label is substituted into the RSSI value of formulae discovery virtual reference label;

G. the RSSI vector of defining virtual reference label and label to be positioned

$\stackrel{\→}{{\mathrm{\θ}}_i}=({\mathrm{\θ}}_{i,1},{\mathrm{\θ}}_{i,2},...,{\mathrm{\θ}}_{i,m})$

$\stackrel{\→}{S_j}=(S_{j,1},S_{j,2},...,S_{j,m})$

Label p to be positioned is to the distance of each virtual reference label

$E_j=\sqrt{\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{({\mathrm{\θ}}_i-S_i)}^2}$

H. K the virtual reference label that chosen distance label to be positioned is nearest;

I. weighted sum calculates the position of label to be positioned, passes through formula

$(x,y)=\underset{i=1}{\overset{k}{\mathrm{\Σ}}}{w}_i(x_i,y_i)$

Wherein

$w_i=\frac{\frac{1}{{E_i}^2}}{{\mathrm{\Σ}}_{i=1}^k\frac{1}{{E_i}^2}}$

w ₁+w ₂+…+w _K=1

J. the coordinate of each virtual reference label is drawn successively by backwards calculation;

K. the coordinate figure of label to be positioned is upgraded;

L. difference and certain threshold value of coordinate figure in K and J coordinate is compared;

If be M. more than or equal to this threshold value, after the coordinate renew result in K, return J;

If be N. less than this threshold value, K is final positioning result.

Beneficial effect of the present invention is as follows:

In the process of location, this algorithm calculates the RSSI value of virtual reference label by more accurate mode, if need to obtain more accurate positioning precision, only needs more careful for initial planar mesh, and do not need the density increasing reference label, thus reduce costs.Use iterative correction techniques, the positioning error because the barrier in the physical location of positioning label and localizing environment causes can be reduced.The positioning result that the present invention draws is more accurate.

Accompanying drawing explanation

In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, be briefly described obtaining the required accompanying drawing used in description of the prior art to embodiment below.

Fig. 1 is the location algorithm process flow diagram that the present invention is based on virtual label and cyclic check

Fig. 2 is the virtual reference label distribution plan (getting m=10, n=12 in this figure) being divided into m*n

Fig. 3 is adjacent four virtual reference labels and a label schematic diagram to be positioned

Embodiment

The invention provides a kind of location algorithm based on virtual label and cyclic check, for making object of the present invention, technical scheme and effect clearly, clearly, below in conjunction with accompanying drawing, the present invention is described in more detail.Should be appreciated that specific embodiment described herein only in order to explain the present invention, be not intended to limit the present invention.

The present invention introduces a kind of indoor positioning algorithms based on virtual label and cyclic check, is calculated the RSSI value of virtual reference label by a kind of more reasonably mode, by reducing the impact of environment radio frequency signal and then improving positioning precision.

Refer to Fig. 1, Fig. 1 is the location algorithm process flow diagram that the present invention is based on virtual label and cyclic check, and described method comprises:

S1. locating area is divided into a plane grid, entity reference label is evenly scattered in the various piece in grid, laterally identical with longitudinal interval;

S2. plane grid is divided with the virtual grid of m*n, get the coordinate position of each virtual net center of a lattice as a reference label, represent the virtual reference label covering thus of this virtual grid, each virtual net center of a lattice is as the coordinate position of a reference label, represent this virtual grid thus virtual reference label covered, refer to Fig. 2, Fig. 2 is the virtual reference label distribution plan (getting m=10, n=12 in this figure) being divided into m*n;

S3. choose RSSI value and the distance of adjacent four labels, substitute into the n pushed away by logarithm distance loss model formula, ξ expression formula, obtains 3 results;

S4. the loss model parameter n of mean value as this region of 3 results is got, ξ value;

The RSSI value of the n S5. will tried to achieve, ξ parameter value and a label substitutes into the RSSI value of formulae discovery virtual reference label;

S6. K the virtual reference label that chosen distance label to be positioned is nearest, is calculated the position of label to be positioned, refers to Fig. 3 by weighted sum, Fig. 3 is adjacent four virtual reference labels and a label schematic diagram to be positioned;

S7. the coordinate of each virtual reference label is drawn successively by backwards calculation;

S8. the coordinate figure of label to be positioned is upgraded;

S9. the difference of tag coordinate value to be positioned and reference label coordinate figure is calculated;

S10. more whether difference is less than threshold value;

S11. if the coordinate getting label to be positioned is final positioning result;

S12. if not, return S7.

Claims (9)

1. a location algorithm, is characterized in that, comprises step:

S1. locating area is divided;

S2. the RSSI value of each virtual reference label to each reader is drawn according to the entity reference tag computation nearest apart from it;

S3. use k nearest neighbor algorithm to treat positioning label weighted sum and position calculating;

S4. the result of location is revised by the algorithm of cyclic check.

2. localization method according to claim 1, is characterized in that, described step S1 also comprises:

S11. locating area is regarded as a plane grid, entity reference label disperses within a grid uniformly, laterally identical with longitudinal interval;

S12. this network is divided into the virtual network of m*n, each virtual net center of a lattice as the coordinate position of a reference label, represent this virtual grid thus virtual reference label covered.

3. localization method according to claim 1, is characterized in that, described step S2 also comprises:

S21. logarithm distance loss model formula is adopted calculate;

S22. by two parameter n, ξ in friendly neighbour's label cooperated computing formula;

S23. the RSSI value of virtual reference label is calculated.

4. localization method according to claim 1, is characterized in that, described step S3 also comprises:

S31. the RSSI vector of defining virtual reference label and label to be positioned;

S32. K the virtual reference label that chosen distance label to be positioned is nearest, positions calculating by weighted sum.

5. localization method according to claim 1, is characterized in that, described step S4 also comprises:

S41. the label of computing reference coordinate;

S42. iteration more new target location.

6. localization method according to claim 3, is characterized in that, described step S22 also comprises:

S221. formula is used calculate;

S222. RSSI value and the distance of choosing four labels respectively substitute into above-mentioned formula, try to achieve 3 results;

S223. average as the loss model parameter in this region.

7. localization method according to claim 4, is characterized in that, described step S31 also comprises:

S311.

S312. adopt Euclidean distance represent each independently label p to be positioned to the distance of each virtual reference label

。

8. localization method according to claim 4, is characterized in that, described step S32 also comprises:

S331. the position of label to be positioned

S332. weight factor and w ₁+ w ₂+ ... + w _k=1.

9. localization method according to claim 5, is characterized in that, described step S42 also comprises:

S421. the coordinate of each virtual reference label is drawn successively by backwards calculation;

S422. the coordinate figure of label to be positioned is upgraded;

S423. when the value in S422 and S421 differ be less than certain threshold value time, S422 is final positioning result, otherwise returns S421 after upgrading result with the coordinate figure of S422.