CN110650436A - WiFi data-based position fingerprint database establishing and fitting method - Google Patents

WiFi data-based position fingerprint database establishing and fitting method Download PDF

Info

Publication number
CN110650436A
CN110650436A CN201910557380.1A CN201910557380A CN110650436A CN 110650436 A CN110650436 A CN 110650436A CN 201910557380 A CN201910557380 A CN 201910557380A CN 110650436 A CN110650436 A CN 110650436A
Authority
CN
China
Prior art keywords
data
fingerprint database
wifi
fingerprint
fitting
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201910557380.1A
Other languages
Chinese (zh)
Other versions
CN110650436B (en
Inventor
门华江
王祖林
王宁君
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
SHENZHEN XINFENG WEIYE TECHNOLOGY Co Ltd
Beijing University of Aeronautics and Astronautics
Original Assignee
SHENZHEN XINFENG WEIYE TECHNOLOGY Co Ltd
Beijing University of Aeronautics and Astronautics
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by SHENZHEN XINFENG WEIYE TECHNOLOGY Co Ltd, Beijing University of Aeronautics and Astronautics filed Critical SHENZHEN XINFENG WEIYE TECHNOLOGY Co Ltd
Priority to CN201910557380.1A priority Critical patent/CN110650436B/en
Publication of CN110650436A publication Critical patent/CN110650436A/en
Application granted granted Critical
Publication of CN110650436B publication Critical patent/CN110650436B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02—Services making use of location information
    • H04W4/021—Services related to particular areas, e.g. point of interest [POI] services, venue services or geofences
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B17/00—Monitoring; Testing
    • H04B17/30—Monitoring; Testing of propagation channels
    • H04B17/309—Measuring or estimating channel quality parameters
    • H04B17/318—Received signal strength
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B17/00—Monitoring; Testing
    • H04B17/30—Monitoring; Testing of propagation channels
    • H04B17/391—Modelling the propagation channel
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/30—Services specially adapted for particular environments, situations or purposes
    • H04W4/33—Services specially adapted for particular environments, situations or purposes for indoor environments, e.g. buildings

Abstract

The method is suitable for the technical field of WiFi data positioning. The invention discloses a WiFi data-based position fingerprint database establishing and fitting method, which comprises a position fingerprint database establishing step and a fingerprint database fitting step, wherein the position fingerprint database establishing step comprises the steps of grid fingerprint database division, grid area division is carried out on the latitude of a fingerprint database area to be divided, two points are used as the diagonal line of a rectangular area, and a blank grid fingerprint is obtained; and (4) fingerprint database filling, namely filling the WiFi signal intensity data uploaded by the user in an accumulated manner and the measured WiFi signal intensity data into grids at corresponding positions according to the signal intensity data positions. Due to the adoption of the plane grid fingerprint library, the problem of positioning accuracy reduction caused by uneven distribution of fingerprint information can be effectively avoided. Meanwhile, Gaussian filtering is carried out on the original collected data, so that errors caused by data collection are reduced, and a solid foundation is laid for further improving indoor positioning accuracy subsequently.

Description

WiFi data-based position fingerprint database establishing and fitting method
Technical Field
The invention relates to the technical field of mobile communication and Internet of things, in particular to a WiFi data-based position fingerprint database establishing and fitting method.
Background
In recent years, attention is paid to a WiFi indoor positioning scheme which relies on indoor hotspots for positioning, and the positioning mode has the advantages of wide indoor coverage range and low deployment cost. The invention provides an indoor position fingerprint database establishing and fitting scheme based on WiFi data, which is a part of an indoor positioning system of a WiFi fingerprint database and mainly explains the establishing and fitting process of the fingerprint database.
In an indoor environment, the WiFi signal strength value received at each location is closely related to the distance from an AP (Access Point) and the surrounding environment, where the location Point is generally referred to as a fingerprint Point, and the fingerprint Point and the WiFi signal strength value received at the Point form a fingerprint. A series of fingerprint sets collected in a region to be estimated are called a fingerprint library, and during real-time positioning, the signal intensity of a point to be measured is compared with the data of the fingerprint library so as to obtain the position of the point to be estimated. The key to the above process is how to establish a location fingerprint database, which is also a hotspot of various WiFi indoor positioning studies. At present, two common fingerprint database establishing methods are mainly used, namely a direct statistical method and a propagation model estimation method.
1. Direct statistical method for position fingerprint database
The direct statistical method obtains abundant longitude and latitude position points and WiFi signal RSSI values at all positions by carrying out a large amount of manual field measurement on a position fingerprint database, and finally establishes the position points and the WiFi signal RSSI of the points as the position fingerprint database. The principle of construction of the location-specific fingerprint library is shown in fig. 1. Fig. 1 shows that a single terminal can generally receive signals transmitted from multiple APs at a certain location, location point (x)i，yj) Its fingerprint information may be expressed as { RSSIij1，RSSIij2，...，RSSIijpIndicating the position of the terminal as (x)i，yj) Signal strength of p APs received on the grid.
Although the direct statistical method for the position fingerprint database can establish the fingerprint database, the direct statistical method is limited by the high cost of manual data acquisition, and sufficient fingerprint data cannot be obtained, so that the direct statistical method for the position fingerprint database cannot be popularized and applied in a large quantity. Meanwhile, the position fingerprint database is established by using direct sample data acquisition, and the data of the position fingerprint points is difficult to avoid the phenomenon of uneven data distribution due to the limitation of manual data acquisition. During the sample collection process, the fingerprint points of the sample positions in a certain area are not uniformly distributed, as shown in fig. 2.
2. Location fingerprint library propagation model inference method
In order to expand the number of location fingerprints under limited data conditions, signal strength prediction methods based on signal propagation models are proposed. The signal propagation models are various, and the propagation model of formula (1) is frequently used.
RSSI＝Pt-K-10αlog10d (1)
Where α is called path loss exponent, Pt is transmission power, and K is a constant depending on the environment and frequency. d denotes a signal acquisition terminal position point (x)i，yj) Distance from the AP location. When the parameters of the formula (1) are determined, the RSSI value of each position point can be calculated.
The propagation model-based inference method has very poor generalization to environments, and different environments need to update a set of model parameters or even change a set of propagation models. Therefore, if a single propagation model is used to predict the RSSI values of fingerprint points, the model parameters are incorrect or the environment changes, which may cause the amplification of the error in the fingerprint database, and ultimately affect the accuracy of indoor positioning.
The existing WiFi data indoor position fingerprint database establishing method has the following defects:
first, the distribution of the position fingerprint point information is not uniform
Because the position fingerprint database is built by using direct sample data acquisition, and the limitation of indoor manual data acquisition, the data of the position fingerprint points is difficult to avoid presenting the phenomenon of uneven data distribution. In the sample collection process, the non-uniform distribution of the position points can cause defects in two aspects, on one hand, the redundancy of fingerprint data occurs in partial areas, and the storage and calculation cost is increased. On the other hand, the position fingerprint information of a partial area is sparse, so that the final positioning accuracy is reduced.
Secondly, the accuracy of the fitting mode of the fingerprint database is low
Fitting a WiFi fingerprint library is required to save data acquisition costs. For example, patent document CN104853317A discloses a method for constructing and updating a fingerprint database in WiFi indoor positioning, specifically, a method for tracing WiFi data and directly constructing a fingerprint database based on an anchor node is required in an early stage of fitting. The accuracy of the fingerprint database established by the method greatly depends on the number of the tracing nodes, and the tracing nodes are required to be densely acquired before regression is carried out. However, the acquisition of the tracing nodes requires a lot of manpower, and the method is not suitable for large-scale indoor positioning scenes.
Patent document CN105338498A discloses a method for constructing a fingerprint database in a WiFi indoor positioning system, specifically a fingerprint database filling method, which includes selecting reference points in a grid manner in a positioning area, performing on-site signal strength information acquisition on some reference points, and calculating the propagation process of signal strength information of other reference points which are not acquired by using the strength information of the acquired reference points, so as to implement signal strength filling in an unknown area. However, in practical applications, wireless channels in indoor environments are very complex, different propagation environments have different propagation models, and a single propagation model lacks effective generalization. Therefore, the propagation model estimation method described in patent document CN105338498A is not easy to be generalized, nor is it an optimal fingerprint library fitting method.
Disclosure of Invention
The technical problem mainly solved by the invention is to provide a WiFi data-based position fingerprint database establishing and fitting method, which overcomes the defects caused by non-uniform distribution of fingerprint points, establishes a planar grid fingerprint database, overcomes the defects caused by a fixed model and improves the adaptability and positioning accuracy of a propagation model.
In order to solve the above problems, the present invention provides a method for establishing and fitting a WiFi data-based location fingerprint database, the method for establishing and fitting an indoor location fingerprint database based on WiFi data comprising,
the method comprises the steps of establishing a position fingerprint library and fitting the fingerprint library, wherein the establishing step of the position fingerprint library comprises the steps of dividing a grid fingerprint library, dividing a grid area into latitudes of fingerprint library areas to be divided, setting two points as diagonal lines of a rectangular area, and obtaining blank grid fingerprints;
and (4) fingerprint database filling, namely filling the WiFi signal intensity data uploaded by the user in an accumulated manner and the measured WiFi signal intensity data into grids at corresponding positions according to the signal intensity data positions.
Further, the fingerprint database is a planar rectangular grid fingerprint database.
Further, before the fingerprint database filling step, filtering processing is performed on the WiFi signal strength data, but excluding singular APs.
Further, after filtering the WiFi signal strength data, the filtering process further includes calculating a filtered signal strength mean value of each raster data.
Further, the filtering process includes gaussian filtering.
Further, the latitude and longitude area to be divided is set to be (30.00, 120.00); (30.50, 120.50).
Further, the blank grid fingerprint obtaining step includes setting a segmentation distance, setting segmentation lengths in longitude and latitude directions, and obtaining two basic lists;
and row-column splicing, namely splicing rows and columns after row splicing, or splicing rows and columns after column splicing, and outputting blank grid fingerprints.
Further, each grid point of the fingerprint library corresponds to one signal strength data.
Further, the probability density function of the gaussian filtering is,
wherein the content of the first and second substances,
Figure BDA0002107246490000042
Figure BDA0002107246490000043
n is the total number of samples of RSSI data, RSSIiFor the measured signal strength of one of the total samples, data in the range of (μ -2S, μ +2S) is retained for all RSSIs, μ being the mean of the sample set and S being the standard deviation.
Further, the mean value of the signal strength data set is calculated by the formula,
Figure BDA0002107246490000051
wherein P is the number of APs received by the fingerprint point.
Further, the fingerprint library fitting step includes
Establishing a polynomial fitting model to learn the existing data to obtain the fitting model, and setting a fitting grid signal intensity representation function f (x, y) on a local subdomain of the fitting region to be represented as
Figure BDA0002107246490000052
Wherein a (x, y) ═ a1(x，y)，a2(x，y)，...，am(x，y)]TThe coefficient matrix to be solved is a function of coordinates (x, y), wherein x represents a longitude variable, y represents a latitude variable, and the linear basis p (x, y) of the longitude and latitude coordinates is [1, x, y ]]The quadratic group p (x, y) ═ 1, x, y, x2，xy，y2]。
Further, the linear basis p (x, y) of the latitude and longitude coordinates is [1, x, y ]]The quadratic group p (x, y) ═ 1, x, y, x2，xy，y2]The weight function constructs a loss function:
Figure BDA0002107246490000053
where n is the number of nodes in the region of influence, f (x, y) is the fitting function, z1Is x + x1Node x ofiWeight function w (x-x)i)，
Further, the weight function is a gaussian weight function.
Further, the coefficients a (x, y) are obtained,
Figure BDA0002107246490000054
a(x，y)＝A-1(x，y)B(x，y)z，
wherein the content of the first and second substances,
Figure BDA0002107246490000055
further, instead of the pseudo-inverse,
B＝[w(x-x1)p(x1)，w(x-x2)p(x2)，w(x-xn)p(xn)]
zT＝[z1，z2，...，zn]
zTfor the RSSI value data set of the sample, let formula a (x, y) be A-1(x, y) B (x, y) z, into formula
Figure BDA0002107246490000061
A final fit function is obtained that is,
Figure BDA0002107246490000062
the position fingerprint database establishing and fitting method based on WiFi data comprises a position fingerprint database establishing step and a fingerprint database fitting step, wherein the position fingerprint database establishing step comprises grid fingerprint database division, grid areas are divided according to the latitude of a fingerprint database area to be divided, two points are set as the diagonal line of a rectangular area, and a blank grid fingerprint is obtained; and (4) fingerprint database filling, namely filling the WiFi signal intensity data uploaded by the user in an accumulated manner and the measured WiFi signal intensity data into grids at corresponding positions according to the signal intensity data positions. Due to the adoption of the plane grid fingerprint library, the problem of positioning accuracy reduction caused by uneven distribution of fingerprint information can be effectively avoided. Meanwhile, Gaussian filtering is carried out on the original collected data, so that errors caused by data collection are reduced, and a solid foundation is laid for further improving indoor positioning accuracy subsequently.
Drawings
In order to illustrate the embodiments of the invention or the technical solutions in the prior art more clearly, the drawings that are needed in the description of the embodiments or the prior art will be briefly introduced below, it is obvious that the drawings in the description only show some embodiments of the invention and therefore should not be considered as limiting the scope, and for a person skilled in the art, other related drawings can also be obtained from these drawings without inventive effort.
Fig. 1 is a schematic diagram of a conventional WiFi data location fingerprint database establishment process.
FIG. 2 is a diagram of a prior art sample location fingerprint point distribution.
FIG. 3 is a schematic flow chart of a WiFi data-based location fingerprint database establishment and fitting method of the present invention.
FIG. 4 is a schematic diagram of a region to be divided.
Fig. 5 is a schematic diagram of grid cutting.
FIG. 6 is a schematic diagram of a blank fingerprint library.
FIG. 7 is a schematic diagram of a WiFi data grid fingerprint database establishment process of the present invention.
Fig. 8 schematic diagram of WiFi signal strength filling result (db).
FIG. 9 is a three-dimensional schematic of the fitting results.
Fig. 10 illustrates an inscribed sample selection process.
FIG. 11 is a schematic diagram of the results of a library of location fingerprints.
The implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
To more clearly illustrate the technical solution of the present invention, the following claims are further described in detail with reference to specific embodiments and the drawings, it is to be understood that the following drawings only show some embodiments of the present invention, and therefore should not be considered as limiting the scope, and any limited number of modifications made by anyone within the scope of the claims still fall within the scope of the claims.
It should be understood that in the description of the present invention, it is to be noted that the terms "upper", "lower", and the like refer to orientations or positional relationships based on the orientations or positional relationships shown in the drawings or orientations or positional relationships conventionally arranged in use of products of the present invention, and are used for convenience of description and simplicity of description, but do not refer to or imply that the devices or elements referred to must have a specific orientation, be constructed in a specific orientation, and be operated, and thus, should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and the like are used merely to distinguish one description from another, and are not to be construed as indicating or implying relative importance.
As shown in fig. 1, the present invention provides an embodiment of a method for establishing and fitting a location fingerprint database based on WiFi data.
The WiFi data-based location fingerprint database establishing and fitting method comprises the following steps,
a step of establishing a position fingerprint database and a step of fitting the fingerprint database, wherein the step of establishing the position fingerprint database comprises,
dividing a grid fingerprint database, namely dividing the latitude of a fingerprint database area to be divided into grid areas, and setting two points as the diagonal lines of a rectangular area to obtain blank grid fingerprints;
and (4) fingerprint database filling, namely filling the WiFi signal intensity data uploaded by the user in an accumulated manner and the measured WiFi signal intensity data into grids at corresponding positions according to the positions of the signal intensity data, namely matching the RSSI data into a well-divided blank grid fingerprint database according to the latitude and longitude areas to which the RSSI data belongs.
Specifically, the grid fingerprint database is established for two purposes, namely, the redundancy of sample points can be reduced, and the matching rate of points to be positioned and the sample points in the positioning process is improved. The specific division is described in more detail in formulas (1) to (3) and fig. (4) to (6) described in the following embodiments. The WiFi data-based location fingerprint database establishment comprises location fingerprint database establishment, and the location fingerprint database establishment comprises a fingerprint database dividing step.
In order to overcome the defect of nonuniform distribution of the position points of the fingerprint database, the fingerprint database adopts a planar rectangular grid fingerprint database. Therefore, before the planar rectangular grid fingerprint database is filled with data, the grid area needs to be divided according to the longitude and latitude. Specifically, according to the theory that two points can determine a rectangular range, the longitude and latitude area to be divided can be set as: ' 30.00, 120.00; 30.50, 120.50' and the two points are the diagonals of the rectangular area, the determined area is shown in fig. 4, namely, two points can determine a plane area. The divided areas can comprise indoor areas and outdoor areas, and the invention mainly solves the positioning state of the indoor areas.
Determining a region needing grid segmentation, and needing a position segmentation algorithm, wherein the position refers to a rectangular range determined by two diagonal points, and can be understood as a grid for segmenting a large-range region into small regions, so as to form a blank grid fingerprint database. The specific segmentation idea is as follows:
by setting the division step size to 10, the longitude direction to be m parts and the latitude direction to be n parts, the region can be divided into m x n parts
m-(max(longitude)-min(longitude))/10，m∈N (1)
n＝(max(latitude)-min(latitude))/10，n∈N (2)
After longitude division and latitude division, two basic lists list (longitude) and list (latitude) are obtained, and row-column splicing cannot be performed simultaneously in the grid fingerprint cutting algorithm. Taking the above flow as an example, the column splicing process may be affected by the row splicing. When the a-th row splicing operation is completed, the ordinal number of the column splicing is i
i＝long(list(logitude))+(long(list(longitude))*a-1) (3)
After the region splicing is completed, the grid fingerprint can be output, and the final grid fingerprint output of the scheme is shown in fig. 6. It should be noted that the grid is still in a blank state so far, and the fingerprint information needs to be filled to form a preliminary fingerprint library.
Fingerprint library filling step
The original data in the invention is WiFi data and partial measurement data uploaded by users in an accumulated way, and RSSI (received Signal Strength indication) data of WiFi is filled into a corresponding grid based on the planar grid segmentation result. Prior to direct grid fingerprint library filling, the RSSI data is gaussian filtered and averaged herein.
It has been explained above that the process of the terminal collecting WiFi data to directly build the location fingerprint database is as shown in fig. 1 above. Unlike fig. 1, the position point of the present scheme is determined by the grid point, and the position where the sample is acquired will also be one-to-one matched with the position of the grid. The grid fingerprint database construction process is shown in fig. 7.
In the scheme, in order to improve the subsequent position positioning efficiency, each grid point of the final fingerprint database can only correspond to one RSSI value. Due to the complexity of signal propagation, the variability between WiFi devices, and the instability of the propagation environment, such phenomena may cause data errors in data at both the transmitting end and the receiving end. Therefore, the scheme provides a noise point filtering and averaging mode, which can reduce the error on one hand and calculate the RSSI value of each grid of the final fingerprint database on the other hand.
Set all AP samples collected by the terminal as { RSSI1，RSSI2，...，RSSIn}. Aiming at the error existing in the data acquisition process, the scheme provides a Gaussian filtering mode to relieve the sampling error. Let all RSSI data samples obey (μ, s)2) Is a Gaussian distribution of which the probability density function is
Figure BDA0002107246490000091
Wherein
Figure BDA0002107246490000093
N is the total number of samples of RSSI data, RSSIiThe signal strength measured for a certain one of the total samples. For all RSSIs, the data in (μ -2s, μ +2s) is retained and the data outside the above range is deleted. The RSSI data is subjected to the Gaussian filtering operation, so that errors caused in the data acquisition process can be avoided to a certain extent.
Furthermore, the scheme adopts a method of RSSI data set mean value to calculate the RSSI value of the fingerprint database of each sample point, and the specific operation mode is
Where P is the number of APs received by the fingerprint point. The averaging method can overcome errors caused by differences between different WiFi equipment and acquisition equipment to a certain extent.
Table 1 location fingerprint database data correspondence:
Figure BDA0002107246490000102
taking rssi (received Signal Strength indication) value of the WiFi Signal as an example, the filling effect is shown in fig. 8.
4.2 fingerprint library fitting
It can be analyzed from fig. 8 that, when the currently acquired data is used for filling, a large amount of data is missing, which is caused by the difficulty of completely acquiring each grid. The amount of the sample data and the labor cost are in a direct proportion relationship, and if a large amount of samples are manually collected, the labor cost is increased. The fingerprint library fitting scheme provided by the invention can save human resources on one hand and can realize data filling of all fingerprint grid library points on the other hand.
4.2.1 polynomial function fitting
According to the scheme, a polynomial fitting model is firstly established to learn the existing data to obtain the fitting model. Specifically, on the local subdomains of the fitting region, the fitting grid signal intensity representation function f (x, y) is expressed as
Figure BDA0002107246490000103
Wherein a (x, y) ═ a1(x，y)，a2(x，y)，...，am(x，y)]TThe coefficient matrix to be solved is a function of coordinates (x, y), wherein x represents a longitude variable and y represents a latitude variable. p (x, y) ═ p1(x，y)，p2(x，y)，...，pm((x，y)]TIt is a perfect polynomial of k-th order, called polynomial basis function, and m is the number of terms of the basis function. For example, the problem addressed by the present invention is an approximate plane problem, with a linear basis p (x, y) — [1, x, y ]]The quadratic group p (x, y) ═ 1, x, y, x2，xy，y2]Introducing a weight function to construct a loss function:
Figure BDA0002107246490000111
where n is the number of nodes in the region of influence, f (xy) is a fitting function, z1Is x + x1Node x ofiWeight function w (x-x)i) The weight function is a Gaussian weight function. To ensureConstant coefficient a (x, y), equation (9) should take a minimum value, and equation (9) is derived for a:
Figure BDA0002107246490000112
a(x，y)＝A-1(x，y)B(x，y)z， (11)
wherein
Figure BDA0002107246490000113
The process of a (x, y) inversion is very complicated in calculation amount, and the inversion is generally performed by adopting elementary transformation, but the phenomenon that the inversion cannot be performed occurs in many scenes. In order to avoid that A (x, y) can not be inverted, the invention adopts pseudo-inversion instead, and the situation that the metamorphosis matrix can not be inverted can be avoided.
B＝[w(x-x1)p(x1)，w(x-x2)p(x2)，w(x-xn)p(xn)] (13)
zT＝[z1，z2，...，zn] (14)
zTFor the RSSI value dataset of the sample, the final fitting function can be obtained by substituting equation (11) for equation (8):
Figure BDA0002107246490000114
taking an indoor area as an example, the final polynomial fitting effect is shown in fig. 9. Where the vertical axis represents WiFi signal strength (RSSI) data for each grid, each grid is populated with WiFi signal strength RSSI data after the data is completely fit.
Careful inspection of the upper graph reveals that the fitting result is changed abruptly at the edge due to the drift of the data itself, which is frequent in practical application. A mutation in the fitting value may cause part of the fitting result to be out of the range of (μ -2s, μ +2s) described above, so that the fitting result deviates from the actual value. Analyzing the above situation means that the fitting of the fingerprint library cannot rely entirely on polynomial functions, and separate processing is required for the mutation outliers. So far, sample data are greatly enriched through polynomial function fitting, and the problem of insufficient sample points can not occur when an interpolation method is used for grid fitting. Therefore, in order to solve the problem of the abnormal value due to mutation, the invention provides an inverse distance weighting algorithm to fit the abnormal value region.
Inverse distance weighted fitting
The inverse distance weighting algorithm is an interpolation algorithm based on the distance between a point to be fitted and a known point, and can better solve the problem of sudden change of the fitting result of the polynomial function. And (3) removing fitting data points out of the range of (mu-2 s and mu +2s) aiming at the fitting result of 4.2.1, and calculating the RSSI value of the point to be fitted by using inverse distance weighting, wherein the specific operation method comprises the following steps:
a) all known grid points are known as known data, and the RSSI value is recorded as Z (X)i，Yj)，Xi、YjRepresenting the horizontal and vertical coordinates in the grid fingerprint;
b) sequentially selecting grid position information to be fitted as (X)0，Y0)；
c) Calculating the distance of the grid to be calibrated from all inscribed sample grids
Figure BDA0002107246490000121
For the concept of inscribed samples, the invention is defined as all known points contained in the circular area with the point to be fitted as the center point and r as the radius, the process is shown in fig. 10. Within this region, the total number of sample points is denoted as N.
d) Calculate the weight for each calibration grid: the weight is a function of the reciprocal of the distance between the point to be fitted and each sample point
Figure BDA0002107246490000122
e) Calculating the signal intensity difference of the grid to be calibrated:
Figure BDA0002107246490000131
and repeating the steps a-e until all the grid fingerprint points to be filled are calculated.
1) A certain indoor area is selected as a fitting result to be displayed, and the fitting result of the indoor fingerprint database based on the WiFi data is shown in fig. 11.
2) And establishing one of the key points and the protection points of the indoor position fingerprint database. The position fingerprint database establishment mainly relates to the segmentation of a plane grid and the filling processing of the data of the grid fingerprint database; (step 4.1)
3) The indoor fingerprint library fitting mode is another protection point of the invention. The method designs a polynomial function model, learns the rule of known sample points through the model, solves the polynomial function parameters, and finally calculates unknown raster data by using the polynomial. In order to solve the problem of sudden change in polynomial fitting, the scheme uses an inverse distance weighting algorithm as a supplement of a fitting algorithm. (step 4.2)
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: it is to be understood that modifications may be made to the above-described arrangements in the embodiments or equivalents may be substituted for some of the features of the embodiments described above, and such modifications or substitutions may be made without departing from the spirit and scope of the present invention in its aspects.

Claims (13)

1. The WiFi data-based location fingerprint database establishing and fitting method comprises the following steps,
the method comprises the steps of establishing a position fingerprint library and fitting the fingerprint library, wherein the establishing step of the position fingerprint library comprises the steps of dividing a grid fingerprint library, dividing a grid area into latitudes of fingerprint library areas to be divided, setting two points as diagonal lines of a rectangular area, and obtaining blank grid fingerprints;
and (4) fingerprint database filling, namely filling the WiFi signal intensity data uploaded by the user in an accumulated manner and the measured WiFi signal intensity data into grids at corresponding positions according to the signal intensity data positions.
2. The WiFi data based location fingerprint database creation and fitting method of claim 1, wherein the fingerprint database is a flat rectangular grid fingerprint database.
3. The WiFi data based location fingerprint database creation and fitting method of claim 1 or 2, wherein before the fingerprint database filling step further comprises filtering WiFi signal strength data but excluding singular APs.
4. The WiFi data-based location fingerprint database establishing and fitting method of claim 3, wherein filtering the WiFi signal strength data further comprises calculating filtered signal strength means of each raster data.
5. The WiFi data based location fingerprint database creation and fitting method of claim 1, characterized by that, the latitude and longitude area to be divided is set to (30.00, 120.00); (30.50, 120.50).
6. The WiFi data-based location fingerprint library creation and fitting method of claim 3, wherein the blank grid fingerprint obtaining step comprises,
setting a segmentation distance, and setting segmentation lengths in the longitude and latitude directions to obtain two basic lists;
and row-column splicing, namely splicing rows and columns after row splicing, or splicing rows and columns after column splicing, and outputting blank grid fingerprints.
7. The WiFi data based location fingerprint library creation and fitting method of claim 1, wherein each grid point of the fingerprint library corresponds to one signal strength data.
8. The WiFi data-based location fingerprint library creation and fitting method of claim 5, wherein the Gaussian filtered probability density function is,
Figure FDA0002107246480000021
wherein
Figure FDA0002107246480000022
Figure FDA0002107246480000023
N is the total number of samples of RSSI data, RSSIiData in the range of (μ -2 ∈, μ +2 ∈) for all RSSIs were retained for the measured signal strength of one of the total samples, μ being the mean of the sample set and S being the standard deviation.
9. The WiFi data-based location fingerprint library creation and fitting method of claim 9, wherein the signal strength data set mean value is calculated as,
Figure FDA0002107246480000024
wherein P is the number of APs received by the fingerprint point.
10. The WiFi data based location fingerprint database creation and fitting method of claim 1, wherein the fingerprint database fitting step includes
Establishing a polynomial fitting model to learn the existing data to obtain the fitting model, and setting a fitting grid signal intensity representation function f (x, y) on a local subdomain of the fitting region to be represented as
Figure FDA0002107246480000031
Wherein a (x, y) ═ a1(x，y)，a2(x，y)，...，am(x，y)]TThe coefficient matrix to be solved is a function of coordinates (x, y), wherein x represents a longitude variable, y represents a latitude variable, and the linear basis p (x, y) of the longitude and latitude coordinates is [1, x, y ]]The quadratic group p (x, y) ═ 1, x, y, x2，xy，y2]。
11. The WiFi data based location fingerprint library creation and fitting method of claim 10 wherein the linear basis of the latitude and longitude coordinates p (x, y) ═ 1, x, y]The quadratic group p (x, y) ═ 1, x, y, x2，xy，y2]The weight function constructs a loss function:
Figure FDA0002107246480000032
where n is the number of nodes in the region of influence, f (x, y) is the fitting function, Z1Is x + x1Node x of1Weight function w (x-x)i)。
12. The WiFi data-based location fingerprint library creation and fitting method of claim 11, wherein the coefficients a (x, y) are obtained,
Figure FDA0002107246480000033
a(x，y)＝Λ-1(x，y)B(x，y)z，
wherein the content of the first and second substances,
Figure FDA0002107246480000034
13. the WiFi data-based location fingerprint library creation and fitting method of claim 12, wherein a pseudo-inverse is substituted,
B＝[w(x-x1)p(x1)，w(x-x2)p(x2)，w(x-xn)p(xn)]
ZT＝[Z1，Z2，...，Zn]
zTfor the RSSI value data set of the sample, let formula a (x, y) be A-1(x, y) B (x, y) z, into formula
Figure FDA0002107246480000041
A final fit function is obtained that is,
Figure FDA0002107246480000042
CN201910557380.1A 2019-06-25 2019-06-25 WiFi data-based position fingerprint database establishing and fitting method Active CN110650436B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910557380.1A CN110650436B (en) 2019-06-25 2019-06-25 WiFi data-based position fingerprint database establishing and fitting method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910557380.1A CN110650436B (en) 2019-06-25 2019-06-25 WiFi data-based position fingerprint database establishing and fitting method

Publications (2)

Publication Number Publication Date
CN110650436A true CN110650436A (en) 2020-01-03
CN110650436B CN110650436B (en) 2023-02-24

Family

ID=68989756

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910557380.1A Active CN110650436B (en) 2019-06-25 2019-06-25 WiFi data-based position fingerprint database establishing and fitting method

Country Status (1)

Country Link
CN (1) CN110650436B (en)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110418283A (en) * 2019-06-25 2019-11-05 北京航空航天大学 Location fingerprint base establishing method based on WiFi data
CN111417066A (en) * 2020-03-05 2020-07-14 滴图(北京)科技有限公司 Positioning method, positioning device, electronic equipment and computer readable storage medium
CN111447580A (en) * 2020-03-06 2020-07-24 普联技术有限公司 Indoor positioning method, medium, network management system and terminal control method
CN112073968A (en) * 2020-08-19 2020-12-11 青岛大学 Full-model pseudo AP detection method and detection device based on phase error drift range
CN112218233A (en) * 2020-09-04 2021-01-12 北京爱笔科技有限公司 Position fingerprint generation method, position determination method, position fingerprint generation device and computer equipment
CN113271658A (en) * 2021-05-20 2021-08-17 北京智慧图科技有限责任公司 WiFi fingerprint automatic generation method in indoor positioning
CN116482609A (en) * 2023-06-21 2023-07-25 中国电信股份有限公司 Fingerprint data processing method, fingerprint data processing device, computer equipment, storage medium and product
CN116482608A (en) * 2023-06-21 2023-07-25 中国电信股份有限公司 Information base construction method, device, computer equipment, storage medium and product

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102480742A (en) * 2010-11-25 2012-05-30 中国移动通信集团广东有限公司 Method and device for detecting overshooting of wireless network
US20130004179A1 (en) * 2011-07-01 2013-01-03 Certusview Technologies, Llc Neighborhood node mapping methods and apparatus for ingress mitigation in cable communication systems
CN104853317A (en) * 2014-11-24 2015-08-19 北京航空航天大学 WiFi indoor positioning fingerprint database construction and update method
CN108243405A (en) * 2016-12-26 2018-07-03 中国移动通信集团广东有限公司 The localization method and device of a kind of method for building up of fingerprint base, measurement report MR
CN110418283A (en) * 2019-06-25 2019-11-05 北京航空航天大学 Location fingerprint base establishing method based on WiFi data

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102480742A (en) * 2010-11-25 2012-05-30 中国移动通信集团广东有限公司 Method and device for detecting overshooting of wireless network
US20130004179A1 (en) * 2011-07-01 2013-01-03 Certusview Technologies, Llc Neighborhood node mapping methods and apparatus for ingress mitigation in cable communication systems
CN104853317A (en) * 2014-11-24 2015-08-19 北京航空航天大学 WiFi indoor positioning fingerprint database construction and update method
CN108243405A (en) * 2016-12-26 2018-07-03 中国移动通信集团广东有限公司 The localization method and device of a kind of method for building up of fingerprint base, measurement report MR
CN110418283A (en) * 2019-06-25 2019-11-05 北京航空航天大学 Location fingerprint base establishing method based on WiFi data

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
卢恒惠等: "基于三角形与位置指纹识别算法的WiFi定位比较", 《移动通信》 *

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110418283A (en) * 2019-06-25 2019-11-05 北京航空航天大学 Location fingerprint base establishing method based on WiFi data
CN111417066A (en) * 2020-03-05 2020-07-14 滴图(北京)科技有限公司 Positioning method, positioning device, electronic equipment and computer readable storage medium
CN111417066B (en) * 2020-03-05 2021-01-05 滴图(北京)科技有限公司 Positioning method, positioning device, electronic equipment and computer readable storage medium
CN111447580A (en) * 2020-03-06 2020-07-24 普联技术有限公司 Indoor positioning method, medium, network management system and terminal control method
CN112073968A (en) * 2020-08-19 2020-12-11 青岛大学 Full-model pseudo AP detection method and detection device based on phase error drift range
CN112218233A (en) * 2020-09-04 2021-01-12 北京爱笔科技有限公司 Position fingerprint generation method, position determination method, position fingerprint generation device and computer equipment
CN113271658A (en) * 2021-05-20 2021-08-17 北京智慧图科技有限责任公司 WiFi fingerprint automatic generation method in indoor positioning
CN116482609A (en) * 2023-06-21 2023-07-25 中国电信股份有限公司 Fingerprint data processing method, fingerprint data processing device, computer equipment, storage medium and product
CN116482608A (en) * 2023-06-21 2023-07-25 中国电信股份有限公司 Information base construction method, device, computer equipment, storage medium and product
CN116482609B (en) * 2023-06-21 2023-09-29 中国电信股份有限公司 Fingerprint data processing method, fingerprint data processing device, computer equipment, storage medium and product
CN116482608B (en) * 2023-06-21 2023-09-29 中国电信股份有限公司 Information base construction method, device, computer equipment, storage medium and product

Also Published As

Publication number Publication date
CN110650436B (en) 2023-02-24

Similar Documents

Publication Publication Date Title
CN110650436B (en) WiFi data-based position fingerprint database establishing and fitting method
CN110418283B (en) WiFi data-based location fingerprint database creation method
CN108271165B (en) Method and system for predicting coverage state of base station network signal
CN111062466B (en) Method for predicting field intensity distribution of cell after antenna adjustment based on parameters and neural network
CN109526019B (en) Method for correcting wireless private network base station coverage propagation model based on measured data
CN110430578B (en) Method for realizing cell azimuth prediction based on mobile terminal data
CN114547553B (en) Inversion method, device and equipment for carbon dioxide emission and storage medium
CN108540987B (en) LTE network coverage state evaluation method and device
CN112381337B (en) Multi-source meteorological data fusion processing method, system, terminal and medium
CN109143157B (en) Dynamic parameter updating ranging method based on signal intensity value hybrid filtering
CN114490905B (en) Clear sky surface net long wave radiation integrated inversion method and system
CN108207005B (en) LTE wireless network evaluation method and server
JP2023535841A (en) Energy-saving processing method and processing device for base station
CN110727752B (en) Position fingerprint database processing method, device and computer readable storage medium
CN115100376A (en) Electromagnetic spectrum map implementation method based on improved inverse distance interpolation method
WO2021259372A1 (en) Wireless signal propagation prediction method and apparatus
CN106488555B (en) Indoor positioning method and device
CN105828342B (en) Method and device for confirming neighbor relation
CN105704730A (en) Calibration method for calibrating SPM channel propagation model
CN112799039B (en) Cognitive radio ranging system
CN114170427B (en) Wireless microwave rain attenuation model SSIM image similarity evaluation method based on rain cells
KR100949576B1 (en) A Position Estimation Method using Compensation
CN108738129A (en) A kind of base station positioning method based on drive test data
CN112243242B (en) Large-scale antenna beam configuration method and device
CN114819264A (en) Photovoltaic power station irradiance ultra-short term prediction method based on space-time dependence and storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant