CN107948930A - Indoor positioning optimization method based on location fingerprint algorithm - Google Patents
Indoor positioning optimization method based on location fingerprint algorithm Download PDFInfo
- Publication number
- CN107948930A CN107948930A CN201711495301.6A CN201711495301A CN107948930A CN 107948930 A CN107948930 A CN 107948930A CN 201711495301 A CN201711495301 A CN 201711495301A CN 107948930 A CN107948930 A CN 107948930A
- Authority
- CN
- China
- Prior art keywords
- point
- rssi
- signal strength
- location fingerprint
- neighborhood
- 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
Links
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
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S11/00—Systems for determining distance or velocity not using reflection or reradiation
- G01S11/02—Systems for determining distance or velocity not using reflection or reradiation using radio waves
- G01S11/06—Systems for determining distance or velocity not using reflection or reradiation using radio waves using intensity measurements
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
- G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
- G01S5/0252—Radio frequency fingerprinting
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
- G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
- G01S5/12—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves by co-ordinating position lines of different shape, e.g. hyperbolic, circular, elliptical or radial
-
- 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
- H04W4/023—Services making use of location information using mutual or relative location information between multiple location based services [LBS] targets or of distance thresholds
-
- 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
- H04W4/025—Services making use of location information using location based information parameters
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Position Fixing By Use Of Radio Waves (AREA)
Abstract
The invention discloses a kind of indoor positioning optimization method based on location fingerprint algorithm.The present invention has carried out corresponding optimization processing, so as to fulfill the precision improvement and improved efficiency to position fixing process on the basis of existing location fingerprint localization method to signal strength collection, fingerprint base density, fingerprint base precision, location efficiency, equipment error.The present invention is in positioning process, and without the support of additional hardware, therefore it is with relatively low positioning cost.It is accurate in signal strength measurement, and the precision in location fingerprint storehouse and density it is higher in the case of, the present invention can provide higher positioning accuracy.
Description
Technology neighborhood
The invention belongs to indoor positioning technologies field, and in particular to a kind of indoor positioning side based on location fingerprint algorithm
Method.
Background technology
Fingerprinting localization algorithm passes through RSSI (Received Signal Strength Indication, received signal strength
Indicate) correlation between information completes positioning, by by the RSSI information at anchor point and the RSSI information of location fingerprint into
Row is matched to realize the estimation of the elements of a fix.The positioning principle of fingerprinting localization algorithm is as shown in Figure 1, its position fixing process is divided into two
Stage：First, off-line training step：Select n point is as a reference point to carry out location fingerprint collection in the zone, so as to complete position
The structure of fingerprint base, wherein, location fingerprint by reference point position coordinates and RSSI information structures；2nd, the tuning on-line stage：
The RSSI information at anchor point is gathered, and the RSSI information collected is matched with the location fingerprint in location fingerprint storehouse,
Filter out the highest preceding K location fingerprint of matching degree and realize position estimation.Location fingerprint, which positions common algorithm, includes arest neighbors
Method, k-nearest neighbor (KNN), weighting k-nearest neighbor (WKNN).
Nearest neighbor algorithm is positioned according to the similarity between sample, its core concept is chosen in the fingerprint base of position
Coordinate of the coordinate of most like reference point as positioning node with RSSI vectors at anchor point, wherein, between RSSI vectors
Similarity judges that Euclidean distance is smaller by the Euclidean distance between RSSI vectors, represents the phase between two RSSI vectors
It is higher like spending.Euclidean distance between RSSI vectorsWherein rssiiRepresent to save in positioning
The signal strength of i-th of wireless aps (access node) at point, rssik,iRepresent in location fingerprint storehouse the at k-th of reference point
The signal strength of i wireless aps, m represent the number of wireless aps.
Improvement of the k nearest neighbor algorithm as nearest neighbor algorithm, it is also the similarity between foundation node to be positioned.
During positioning, k nearest neighbor algorithm can select the highest K reference point of similarity between positioning node in the fingerprint base of position, and count
The barycenter for the polygon that K reference point is formed is calculated, the position where barycenter is the position of positioning node：Wherein, (xi,yi) represent reference point coordinate.
Weighting k nearest neighbor algorithm is the improvement to k nearest neighbor algorithm, reduce further the error in position fixing process.Compared to K
Nearest neighbor algorithm, weighting k nearest neighbor algorithm are not K reference point groups of direct calculating after the highest reference point of K similarity is obtained
Into polygon barycenter, but according to the similarity between each reference point and positioning node to the coordinate position of reference point into
Row weighting, the result of weighting summation is final positioning result：Wherein, diBetween expression node
The Euclidean distance of RSSI vectors, (xi,yi) represent reference point coordinate.Weighting k nearest neighbor algorithm controls reference point by weighting
The influence of anchor point, makes the result of positioning more accurate.
But existing location fingerprint localization method, during fingerprint collecting, simply the simple RSSI information that carries out is adopted
Collect to build location fingerprint storehouse, its fingerprint collecting precision is deposited not enough；And when building location fingerprint storehouse, simply simply by people
The mode of work collection is built, once the density requirements in location fingerprint storehouse improve, its construction work amount will roll up；And
And for noise present in location fingerprint storehouse, it is not removed by the way of any, the fingerprint essence in its location fingerprint storehouse
Accuracy needs to be further improved；Meanwhile in matched position fingerprint, the mode used to match one by one, its matching treatment when
Between complexity it is of a relatively high, in addition, do not carry out special processing for the position error that equipment otherness is brought, positioning is tied
Fruit can bring certain position error.
The content of the invention
The goal of the invention of the present invention is：For above-mentioned problem, there is provided a kind of effectively lifting setting accuracy is simultaneously
The positioning and optimizing method of fingerprint matching processing time complexity.
The indoor positioning optimization method based on location fingerprint algorithm of the present invention comprises the following steps：
Build location fingerprint storehouse step：
101：Dispose m radio reception device in area to be targeted, n reference point of setting (such as to indoor plane figure into
Row mesh generation, mesh point is as a reference point), and record the positional information of each reference point；
102：The received signal strength value rssi of each radio reception device of each reference point of multi collect, is being adopted every time
During collection, if the rssi currently collected meets | rssi- μ | 1.96 σ of ＞, reject the rssi currently collected, and wherein σ, μ distinguishes
Represent the standard deviation and average of the Gaussian Profile of received signal strength；
And mean filter smoothing processing is carried out to multi collect result, obtain each radio reception device of each reference point
Received signal strength value, be denoted as rssik,i, wherein k is that i is radio reception device identifier with reference to point identifier；By same
The m group received signal strength values of reference point form the signal strength vector of the reference point, are denoted as RSSIk=(rssik,1,
rssik,2,…,rssik,n), by the signal strength vector sum location information sets synthesising position fingerprint of each reference point and preserve in place
Put in fingerprint base；
103：Position fingerprint base is handled into line density enhancing：
Increase the location fingerprint in database by way of many-sided curve fit models；
Neighborhood Filtering method screening location fingerprint storehouse noise is carried out to the database after increase again：Travel through in location fingerprint storehouse
Each reference point, using current reference point as central point c, and the P neighborhoods based on central point c, calculate central point c and each neighborhood point
xpBetween average similarityWherein P is default Neighborhood Number, and subscript p is neighborhood point xpPoint identifier,Represent central point c and neighborhood point xpBetween signal strength vector Euclidean distance；And by each neighborhood point xpP neighborhoods
In central point c reject, obtain each neighborhood point xp'sA neighborhood pointSubscript j is neighborhood pointPoint identifier, then calculate
Each neighborhood point xpAverage similarityWhereinRepresent point xpWith pointBetween signal strength vector
Euclidean distance, so as to obtain the average of the average similarity of all neighborhood points of central point cIf
Current central point c is then denoted as noise z；
The received signal strength value rssi of the received signal strength value renewal noise z of P neighborhoods based on each noise zz,iFor：WhereinRepresent the neighborhood point of noise zTo the received signal strength of i-th of radio reception device
Value, wherein, j represents neighborhood pointPoint identifier,Represent neighborhood pointWeights, andLz,jTable
Show the neighborhood point of noise zEuclidean distance between two coordinate points of noise z, so as to obtain the reference point corresponding to noise z
New signal strength vector；
104：Clustering processing is carried out to position fingerprint base using K mean cluster, and judges whether group's number of all kinds of clusters does not surpass
Group's number threshold value is crossed, if so, then preserving current class cluster, otherwise continues to carry out K mean cluster processing to obtained class cluster, until obtaining
Group's number of class cluster be no more than group's number threshold value so that the location fingerprint storehouse built；
Real-time position matching step：
201：The m group received signal strength values at point to be determined are gathered, obtain the signal strength vector of point to be determined
RSSIt；
202：Weighting k-nearest neighbor based on difference obtains the matching result of point to be determined in the fingerprint base of position：
By the signal strength vector RSSI of point to be determinedtIn each component subtract RSSItIn minimum component, obtain difference
Vectorial Δ RSSIt；
By RSSItOr Δ RSSItSimilarity mode is carried out (between signal strength vector with each cluster centre of the class cluster of preservation
The smaller then similarity of Euclidean distance it is higher), search the highest class cluster of similarity；
Based on difference value vector Δ RSSIt, the highest preceding K position of lookup similarity refers in the highest class cluster of similarity
Line, is weighted addition to the corresponding position coordinates of K location fingerprint using weighting k nearest neighbor algorithm, obtains the positioning of anchor point
Coordinate；
203：Latitude and longitude coordinates conversion is carried out to positioning relative coordinate, obtains the positional information of point to be determined.
In conclusion by adopting the above-described technical solution, the beneficial effects of the invention are as follows：
(1) fingerprint collecting precision improvement：Existing method is simply simple to carry out RSSI information during fingerprint collecting
Gather to build location fingerprint storehouse；Gaussian filtering and mean filter have been carried out to RSSI information in the present invention, has effectively increased and adopts
The accuracy of the RSSI information of collection.
(2) location fingerprint storehouse density strengthens：Existing method is when building location fingerprint storehouse, simply simply by artificial
The mode of collection is built, once the density requirements in location fingerprint storehouse improve, its construction work amount will roll up；This hair
It is bright by the way of many-sided curve fit models, in the case of original location fingerprint storehouse density is less, pass through data be fitted side
Fingerprint quantity in formula increase database, achievees the purpose that to increase location fingerprint storehouse density, and then improves location fingerprint positioning and calculate
The positioning accuracy of method.
(3) noise in location fingerprint storehouse is effectively removed, makes the fingerprint precision higher in location fingerprint storehouse, so as to be fixed
The tuning on-line stage of position algorithm provides high-precision data and supports.
(4) location efficiency is lifted：In existing method, when matching for location fingerprint storehouse matches one by one, its
It is O (n) from the time complexity of location fingerprint storehouse matching process；And in the present invention, position is referred to by K-means clustering algorithms
After line storehouse carries out cluster analysis, by first matching cluster centre, then the mode of class cluster is matched, can be complicated by the time of position fixing process
Degree is reduced to O (logn).
(5) equipment error is effectively eliminated：By the difference uniformity of RSSI between distinct device to fingerprint matching process
Algorithm is improved, and significantly reduces the position error that equipment otherness is brought.
Brief description of the drawings
Fig. 1 is location fingerprint positioning schematic；
Fig. 2 is 8 neighborhood administrative division maps；
Fig. 3 is localizing environment plan；
Fig. 4 builds flow chart for location fingerprint storehouse；
Fig. 5 is tuning on-line flow chart；
Fig. 6 is RSSI information gathering figures；
Fig. 7 is k-nearest neighbor (KNN) error map；
Fig. 8 is weighting k-nearest neighbor (WKNN) error map；
Fig. 9 is the error map of the method for the present invention；
Figure 10 is position error comparison diagram；
Figure 11 is cumulative errors distribution map.
Embodiment
To make the object, technical solutions and advantages of the present invention clearer, with reference to embodiment and attached drawing, to this hair
It is bright to be described in further detail.
The present invention passes through signal strength optimization of collection, the optimization of location fingerprint storehouse density, location fingerprint storehouse precision optimizing, positioning
Efficiency optimization, equipment error optimization are imitated to lift the positioning accuracy of the indoor positioning of existing location fingerprint algorithm and localization process
Rate.
In the building process of position fingerprint base, because the signal strength information collected has time variation, therefore, this hair
The bright method that filtering can be used in specific implementation handles collection RSSI information, to filter control information therein, carries
The accuracy of height collection information.Because the distribution of the signal strength measured in same position meets Gaussian distribution model.Therefore
For the RSSI vector RSSI=(rssi at any reference point1,rssi2,rssi3,…,rssim), its any component (rssii,i
=1 ..., m, m represent AP numbers) distribution meet Gaussian distribution model, wherein m represents the number of wireless access point, i.e. RSSI
The number of components of vector.
In statistics, confidential interval represents the estimation interval of the population parameter value constructed to sample, confidential interval
Confidence level is determined by confidence level.For the Gauss of signal strength at reference point (the signal strength values rssi of the single AP of collection)
It is distributed, the relation of its confidential interval and confidence level is：P{X1＜ rssi ＜ X2}=1- α, wherein, X1Represent under confidential interval
Boundary, X2Represent the upper bound of confidential interval, 1- α represent that rssi is in confidential interval (X1, X2) in confidence level.
Sample at the confidential interval expression reference point of signal strength at reference point is on some section of population parameter
Estimation, select rational confidential interval can real reaction signal intensity excursion.For fiducial interval range with
Outer signal strength, then it is assumed that the event for the signal strength occur is a small probability event, meanwhile, for the letter at reference point
For the collection of number intensity, sample beyond confidential interval beyond normal error range, can seriously affect sampling as a result,
Therefore, the present invention rejects the sample beyond confidential interval.In present embodiment, the bound of confidential interval is set
(μ -1.96 σ, μ+1.96 σ) is set to, the sample not in fiducial interval range is then considered gross error, and is rejected.Root
Set according to the border, for the signal strength values rssi of the single AP of collection, when | rssi- μ | during 1.96 σ of ＞, then reject this
The signal strength values of collection, wherein σ represent the standard deviation of Gaussian Profile, and μ represents the average of Gaussian Profile.
Then the signal strength after filtering is smoothed by the way of average value processing again, at average
Reason mode is smoothed, and obtains smoothing processing resultWherein T is any reference
The hits to same wireless aps is put, by the letter of obtained signal strength place's location fingerprint as a reference point after smoothing processing
Number intensity is added in location fingerprint storehouse, so as to improve the precision of signal strength, ensures location fingerprint storehouse construction process middle position
Put the reliability of fingerprint.
In the fingerprinting localization algorithm of position, the density of location fingerprint is bigger in location fingerprint storehouse, the positioning accurate of location algorithm
Degree is higher.The fingerprint density in location fingerprint storehouse determines that sampled point number is more, fingerprint by the sampled point number in unit area
Density is bigger, therefore, the fingerprint density in location fingerprint storehouse can be improved by the sampled point number increased in location fingerprint storehouse, from
And realize the optimization to position fingerprinting localization algorithm, still, with the lifting of location fingerprint density in location fingerprint storehouse, position refers to
Sampled point number in line storehouse will roll up, its construction work amount will be exponentially increased.
On the basis of location fingerprint storehouse construction work amount is not increased, the present invention increases position by the way of data fitting
The fingerprint density of fingerprint base, new location fingerprint information is fitted according to known location fingerprint information, and is added to position
Put in fingerprint base, so as to achieve the purpose that to increase location fingerprint storehouse density.Because the change of wireless signal strength in space is in company
Continuous property, therefore, during this specific implementation is let go, according to the corresponding signal strength values of known location fingerprint, is intended using polynomial surface
The mode of conjunction is fitted the distribution surface of wireless signal strength in space, is calculated according to the surface equation fitted new
The signal strength values of sampled point, with reference to the coordinate and signal strength values of new sample point, complete one group of new location fingerprint
Construction.Substantial amounts of new location fingerprint can be calculated by the signal intensity profile surface equation fitted, and added
Enter into position fingerprint base, so that the fingerprint density of raised position fingerprint base.
When improving the density in location fingerprint storehouse by the way of many-sided curve fit models because the result fitted with it is true
It there may exist certain difference between real value, it can cause to introduce some noises in location fingerprint storehouse, therefore, it is necessary to carrying
The location fingerprint storehouse for having risen density carries out denoising, is location fingerprint to improve the accuracy of location fingerprint in location fingerprint
Location algorithm provides accurately data and supports.Because the distribution of signal strength in space is in continuity and neighbouring similitude, therefore,
The present invention carries out denoising using neighbour average filtering algorithm to position fingerprint base.The core of neighbour average filtering algorithm is thought
Want to replace noise using the average of the point of noise adjacent domain, so as to achieve the purpose that to remove noise.Neighbour average filtering
It would generally be related to the select permeability of neighborhood in algorithm, it is common including 4 neighborhoods, 8 neighborhoods etc., to position fingerprint base
During reason, preferably 8 neighborhoods of the invention.When carrying out location fingerprint storehouse denoising using neighbour average filtering algorithm, the judgement of noise is
The key of Denoising Algorithm.The continuity and neighbouring similitude being distributed according to wireless signal strength in space, the present invention is according to neighbouring
The similarity of location fingerprint judges noise between point.Location fingerprint in space between adjacent point should have very high similar
Degree, when, there are during noise spot, the similarity between its location fingerprint will decrease between consecutive points, wherein, consecutive points it
Between the similarity of location fingerprint weighed by the Euclidean distance between RSSI vectors, Euclidean distance is shorter, represents location fingerprint
Between similarity it is higher.
In neighbour average filtering algorithm, the rule that noise judges is：Average phase between central point and its neighborhood point
When being more than the average similarity between neighborhood point and its corresponding neighborhood point like degree, then it is assumed that central point is a noise spot, such as
Shown in Fig. 2, point c centered on pending point, 8 neighborhoods of central point are by x1~x8Corresponding point expression, central point and its neighborhood
Point between average similarity be：Wherein, P represent c neighborhood point number (i.e. for 8 fields, then P=
8),Represent c points and xiThe Euclidean distance of RSSI vectors between point.
The average similarity between neighborhood point corresponding to center neighborhood of a point point and neighborhood point is：
Wherein,Represent xiThe corresponding neighborhood point of pointNumber, wherein neighborhood pointIn do not include point c,Represent xiPoint and
The Euclidean distance of RSSI vectors between point.
Average similarity between the corresponding neighborhood point of all neighborhood points for trying to achieve central point and then to obtaining
All results carry out average value processing, obtain the average of the average similarity of all neighborhood points of c：
WhenWhen, then it is assumed that c points are noise, are denoted as noise z, and calculate noise using inverse distance weighted interpolation method
The signal strength values at place：The neighborhood point weights of wherein noise zI=
1 ..., m, Lz,jRepresent the neighborhood point x of noise zz,jEuclidean distance between two coordinate points of noise z, so as to obtain the noise
The new RSSI vectors RSSI at placez=(rssiz,1,rssiz,2,…,rssiz,m)。
Inverse distance weighted interpolation method reflects not in weighted fashion according to the position relationship between central point and neighborhood point
Influence of the same neighborhood point to central point, so as to more accurately calculate the interpolated value of central point.At common average
Reason, inverse distance weighted interpolation method are precisely controlled influence of the neighborhood point to center interpolation point by way of weighting, tie calculating
Fruit is more accurate.After denoising, location fingerprint ensure that with very high accuracy fixed online in location fingerprint storehouse
Position stage, location fingerprint storehouse provide accurately data supporting to position it.
When carrying out tuning on-line using location fingerprint location algorithm, it is necessary to which the RSSI vectors that measurement is obtained refer to position
RSSI vectors in line storehouse corresponding to location fingerprint are matched one by one.Wherein, the similarity between RSSI vectors from RSSI to
Euclidean distance between amount is weighed.Because the location fingerprint density in location fingerprint storehouse is bigger, the positional accuracy of location algorithm is got over
It is high.Therefore, in order to improve the positioning accuracy of location fingerprint location algorithm, usually being stored in location fingerprint storehouse has substantial amounts of position to refer to
Line information, seriatim matches the overlong time that position fixing process can be caused to spend, so that impact position refers to for location fingerprint
The location efficiency of line location algorithm.For a large amount of finger print informations in location fingerprint storehouse, present invention introduces the thought contraposition of cluster
Put fingerprint base and carry out clustering, to improve the matching efficiency of location fingerprint location algorithm, so as to improve the positioning of location algorithm
Efficiency.In off-line training step, location fingerprint all in location fingerprint storehouse is divided into N number of class cluster by clustering algorithm, for drawing
The class cluster separated, as still there is a large amount of location fingerprint data in fruit cluster, then again further clusters such cluster, Zhi Daoxi
Location fingerprint number in the class cluster separated is relatively fewer, that is, is no more than default amount threshold；The fingerprinting localization algorithm in position
In the tuning on-line stage, the RSSI vectors of anchor point are matched with the location fingerprint at each class cluster center, are selected and anchor point
The highest class cluster of similarity.For the class cluster filtered out, if such cluster has carried out further clustering, using above-mentioned side
Method continues to match the cluster centre that such cluster marks off, until not segmented further in the class cluster filtered out, then right
Location fingerprint in the class cluster filtered out is matched one by one, the highest preceding K location fingerprint of wherein matching degree is selected, by it
Respective coordinates are weighted the result for trying to achieve positioning., can be by online after being optimized using the thought of cluster to position fingerprint base
The positioning time complexity of positioning stage is reduced to o (logn) by o (n), so as to significantly increase the efficiency of location algorithm.Will be poly-
After the thought of class introduces location fingerprint location algorithm, although the cluster process in location fingerprint storehouse can consume substantial amounts of time resource and
Computing resource, but cluster process by location fingerprint place server off-line training step complete, therefore, cluster process is not
The location efficiency in tuning on-line stage can be impacted.
In clustering algorithm, the clustering algorithm based on division include K-means clustering algorithms, K-medoids clustering algorithms,
Clarans clustering algorithms etc..In present embodiment, using K-means clustering algorithms, its core concept clustered is to pass through
Iteration is not stopped to K clustering scheme so that corresponding Different categories of samples is represented with the average of this K clustering scheme
When, the global error of gained is minimum.In the fingerprinting localization algorithm of position, using K-means clustering algorithms to position fingerprint base into
Row classification, can efficiently reduce the matching times between RSSI vectors in position fixing process, accelerate matching process, reach raising positioning
The purpose of efficiency.After collection, fitting and the denoising work in location fingerprint storehouse is completed, can use K-means clustering algorithms into
Row clustering：
(1) initial center point of the K sample as K cluster is randomly selected in the fingerprint base of position.
(2) Euclidean distance in the fingerprint base of calculation position between each sample and K cluster centre point, sample is divided into
With in the shortest class cluster of its Euclidean distance, Euclidean distance of the Euclidean distance between RSSI vectors in the algorithm.
(3) according to the location fingerprint sample being referred in class cluster, the K central point clustered is recalculated using averaging method.
(4) (1) (2) (3) step is repeated, (center position is not changed or iterated to specified time until iteration convergence
Number etc.), so as to complete the K divisions to position fingerprint base.
Off-line training step and tuning on-line stage of the location fingerprint storehouse in position fingerprinting localization algorithm, can all be related to making
RSSI information is gathered with terminal, but intelligent terminal used in different phase and different user is all there may be difference, and foreign peoples
The RSSI information that the wireless network card of terminal is gathered in same position there may be very big difference, meanwhile, different radio network interface card
Antenna gain can also allow measurement result to produce difference.Therefore, when being positioned using different types of intelligent terminal, the knot of positioning
Fruit can change because of the difference of intelligent terminal, so as to influence the accuracy of positioning.For the technical problem, this specific embodiment party
In formula, positioned using the weighting k nearest neighbor algorithm based on RSSI differences, compared to common weighting location algorithm, the algorithm
It is capable of the position error that the otherness of abatement apparatus is brought, effectively improves the accuracy and reliability of location algorithm.
Weighting k nearest neighbor algorithm based on RSSI differences is the improvement to weighting k nearest neighbor algorithm, its main distinction is embodied in pair
In the processing of RSSI vectors, i.e., difference processing is carried out to the RSSI of terminal collection, each component in RSSI vectors is subtracted into RSSI
Minimum component rssi in vectormin, obtain Δ RSSI=(rssi1-rssimin,rssi2-rssimin,rssi3-rssimin,…,
rssin-rssimin), because in the RSSI of Heterogeneous Terminal collection, the difference between each component is consistent, and therefore, foreign peoples is whole
The Δ RSSI obtained in same positioning node is held to be consistent, carrying out location fingerprint positioning based on Δ RSSI can be effectively
Eliminate the error that terminal otherness is brought.When carrying out location fingerprint matching using the weighting k nearest neighbor algorithm based on RSSI differences,
Matched using Δ RSSI vectors, filtering out the highest preceding K of similarity in location fingerprint storehouse, (in the present invention, K-means is clustered
The value of K in algorithm has no incidence relation with K herein, only refers to preset constant) a location fingerprint, it is near using weighting K
Adjacent algorithm is weighted addition to the corresponding position coordinates of K location fingerprint, finally obtains the elements of a fix of anchor point.
Embodiment
The present embodiment is in an about 84m2Indoor environment in carry out, wherein, 4 AP are deployed with indoor environment, and
It is as a reference point come to build the distance between location fingerprint storehouse, neighboring reference point be 2m that 25 points are have chosen in environment indoors,
Referring to Fig. 3, the deployed position of 4 AP is as shown in figure 3, including two TP_LINK TL-WR886N routers, a TP_
LINK TL-WR740N routers and a Tenda 811R router.Location fingerprint storehouse
In order to realize the real-time localization process to any anchor point, location fingerprint storehouse, ginseng are built first under off-line state
See Fig. 4, its detailed process is as follows：
(1) location fingerprint acquisition scheme is set.
Acquisition terminal of the wireless notepad as signal strength equipped with network adapter is used in the present embodiment,
And RSSI information gatherings and data processing are carried out using softwares such as WirelessMon, SPSS Statistics, MATLAB, its
In, WirelessMon is used for the RSSI information for gathering 4 AP, and SPSS Statistics are used to complete the K to position fingerprint base
Mean cluster, MATLAB are used to carry out many-sided curve fit models to the distribution surface of signal strength.
(2) n group RSSI information is gathered, every group of signal strength values for including m difference AP, that is, obtain n groups RSSI vectors.
By the RSSI information (n is default reference point quantity) of 4 AP nodes in WirelessMon software collections room, Fig. 6
The process that RSSI information gatherings are carried out using WirelessMon softwares is illustrated, X-axis represents the time in figure, and Y-axis represents that signal is strong
Degree, unit dBm.
(3) gaussian filtering removes gross error.
It will be appreciated from fig. 6 that the rssi value collected fluctuation up and down between -60dBm and -80dBm, it can be seen that signal strength
With time variation, therefore gaussian filtering process is carried out to it：When the single rssi value collected meets | rssi- μ | during 1.96 σ of ＞,
Then reject the signal strength values of this collection.
(4) the smooth RSSI information of mean filter.
The RSSI information obtained to step (3) carries out mean value smoothing processing.
(5) by RSSI information and location information sets synthesising position fingerprint：(xk,yk):(rssik,1,rssik,2,…,
rssik,n)；And location fingerprint is added into position fingerprint base.
(6) (2)~(5) are repeated until completing reference point collecting work.
(7) data fitting increase location fingerprint storehouse density.
Each component in the RSSI vectors in the fingerprint base of position is carried out using the Curve Fitting tool boxes of MATLAB
Many-sided curve fit models, for example, by the coefficient of determination (R side) of fitting surface be 0.85.Each component in RSSI vectors are fitted
After signal intensity profile surface chart, the surface equation of signal intensity profile can be obtained, can be calculated according to surface equation
The corresponding RSSI information of diverse location in embodiment, the coordinate of diverse location and corresponding RSSI information are combined, just intended
New location fingerprint data are closed out, can be complete by building multigroup location fingerprint data and being added in location fingerprint storehouse
The increase work of fingerprint quantity into location fingerprint storehouse, so that the fingerprint density of raised position fingerprint base.In the present embodiment, pass through
After data fit procedure, the location fingerprint quantity in location fingerprint storehouse increases 79 groups by original 25 groups, location fingerprint storehouse
Fingerprint density improve more than 3 times.
(8) Neighborhood Filtering method screening location fingerprint storehouse noise.
8 neighborhood of neighborhood mode is set, each reference point in location fingerprint storehouse is traveled through, using current reference point as central point
C, and calculate the average similarity between central point c and its neighborhood pointAnd all neighborhood points of central point c is average similar
The average of degreeIfCurrent central point c is then denoted as noise z.
(9) the RSSI information of inverse distance weighted interpolation method renewal noise.
(10) K mean distances classify position fingerprint base, so that the location fingerprint storehouse built.
K mean cluster is carried out to 79 groups of location fingerprints in the fingerprint base of position as instrument using SPSS Statistics,
All location fingerprints are divided into 8 class clusters, wherein the cluster membership in each cluster is as shown in table 1, each class cluster
Cluster centre it is as shown in table 2.
Table 1
| Cluster | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
| Number | 18 | 10 | 7 | 15 | 9 | 6 | 6 | 8 |
Table 2
After structure completes location fingerprint storehouse, then the RSSI information that can be gathered based on any anchor point is calculated and exported in real time
The location coordinate information of the anchor point, referring to Fig. 5, it includes following process：
(1) the m group rssi values at point to be determined are gathered, obtain the RSSI vectors of point to be determined.
In the present embodiment, 51 points are arbitrarily have chosen in area to be targeted as point to be determined (test point), to this hair
Bright positioning performance is tested.
First, the relative coordinate between 51 points to be determined and location fingerprint storehouse datum mark is surveyed using ranging instrument
After amount, the true relative coordinate of 51 points to be determined has been obtained.Then, it is undetermined to 51 using WirelessMon sampling instruments
RSSI information at site is acquired, and obtains the RSSI vectors of each point to be determined.
(2) the RSSI vectors of point to be determined are smoothed using mean filter method.
(3) the weighting k-nearest neighbor matched position fingerprint base based on difference.
(4) relative position coordinates of point to be determined are obtained.
(5) latitude and longitude coordinates conversion is carried out to positioning relative coordinate, obtains the positional information of point to be determined.
In order to verify that the positioning of the present invention is thought rather, by the way that the relative position coordinates being calculated are opposite with what is actually measured
Position coordinates is contrasted, so as to complete the precision analysis to positioning and optimizing scheme.
(1) position error distribution surface map analysis.
Fig. 7,8 and 9 be respectively be not optimised k nearest neighbor algorithm, be not optimised weighting k nearest neighbor algorithm and the present invention optimization after
Locating scheme error distribution surface figure in the present embodiment, wherein position error weighed using absolute error.In figure,
X-axis and Y-axis represent position coordinates, unit cm, and Z axis represents the size of error distance, unit cm.Fig. 7,8 and 9 are shown respectively
The distribution situation of error, is drawn by carrying out analysis to three figures in different localization methods：Positioning side after the optimization of the present invention
The overall position error fluctuation of case is small, and global error maintains a relatively low level, is calculated compared to the k nearest neighbor being not optimised
Method and the weighting k nearest neighbor algorithm being not optimised, the locating scheme positioning accuracy after optimization are more excellent.
(2) position error contrast map analysis.
Positioning side after Figure 10 is the k nearest neighbor algorithm being not optimised, is not optimised weighting k nearest neighbor algorithm and the optimization of the present invention
Error comparison diagram between case, wherein, X-axis represents the numbering of the anchor point for testing, and Y-axis represents the corresponding positioning of anchor point
Error.Being analyzed from figure to obtain, and the worst error of the locating scheme after optimization is less than two kinds of location algorithms being not optimised.Meanwhile
The position error data progress average value processing of three kinds of positioning methods is drawn, the average localization error for the k nearest neighbor algorithm being not optimised
For 252.86 centimetres, the average localization error for the weighting k nearest neighbor algorithm being not optimised is 251.67 centimetres, the locating scheme after optimization
Average localization error be 196.57 centimetres, it is excellent compared to the k nearest neighbor algorithm that is not optimised and the weighting k nearest neighbor algorithm being not optimised
The positioning accuracy of locating scheme after change improves 22.26% and 21.89% respectively.
(3) cumulative errors profiling analysis
The weighting k nearest neighbor algorithm and the optimum position scheme of the present invention that Figure 11 is the k nearest neighbor algorithm being not optimised, is not optimised
Between deviation accumulation distribution map, wherein, X-axis represent error distance, Y-axis represent error distance corresponding to accumulated probability.From
Analysis is drawn in figure, and in the positioning result of locating scheme after optimization, the result for having 49.02% falls error at 200 centimetres
In the range of, the result for having 88.24% falls in 300 centimetres of error range, and the result for having 100% falls error at 400 centimetres
In the range of, compared to the locating scheme after optimization, this data for the k nearest neighbor algorithm being not optimised are 33.34%, 68.63% and
84.31, this data for the weighting k nearest neighbor algorithm being not optimised are 31.38%, 64.71% and 86.27%, and analysis data can obtain,
The overall position error of locating scheme after optimization is better than two kinds of location algorithms being not optimised.
(4) location efficiency is analyzed
Location efficiency of the location algorithm before and after position fingerprint base carries out clustering is contrasted in table 3, is analyzed
Data in table can obtain, and after carrying out clustering to position fingerprint base, the location efficiency of location algorithm is far above before clustering
Location efficiency, in the case where position fingerprint base scale is bigger, the advantage of location efficiency is more obvious.
To sum up, the present invention from signal strength collection, location fingerprint storehouse density, location fingerprint storehouse precision, location efficiency and determines
Error five aspects in position are started with, and are realized the optimization to position fingerprinting localization algorithm, are completed to positioning accuracy and location efficiency
Double optimization.Test result indicates that the average localization error of the locating scheme after optimization is 196.57 centimetres, compared to common
K nearest neighbor algorithm and weighting k nearest neighbor algorithm, locating scheme after optimization 22.26% He is improved in terms of positioning accuracy
21.89%；The location fingerprint location algorithm not clustered compared to location fingerprint storehouse, locating scheme after optimization is by position fixing process
Time complexity is reduced to O (logn) by O (n), will in terms of its location efficiency in the case that in position, fingerprint base is on a grand scale
With very big advantage.
The above description is merely a specific embodiment, any feature disclosed in this specification, except non-specifically
Narration, can be replaced by other alternative features that are equivalent or have similar purpose；Disclosed all features or all sides
Method or during the step of, in addition to mutually exclusive feature and/or step, can be combined in any way.
Claims (3)
1. the indoor positioning optimization method based on location fingerprint algorithm, it is characterised in that comprise the following steps：
Build location fingerprint storehouse step：
101：M radio reception device is disposed in area to be targeted, n reference point is set, and records the position letter of each reference point
Breath；
102：The received signal strength value rssi of each radio reception device of each reference point of multi collect, is gathering every time
When, if the rssi currently collected meets | rssi- μ | 1.96 σ of ＞, reject the rssi currently collected, and wherein σ, μ distinguishes table
Show the standard deviation and average of the Gaussian Profile of received signal strength；
And mean filter smoothing processing is carried out to multi collect result, obtain connecing for each radio reception device of each reference point
Signal strength values are received, are denoted as rssik,i, wherein k is that i is radio reception device identifier with reference to point identifier；By same reference
The m group received signal strength values of point form the signal strength vector of the reference point, are denoted as RSSIk=(rssik,1,rssik,2,…,
rssik,n), by the signal strength vector sum location information sets synthesising position fingerprint of each reference point and it is saved in location fingerprint storehouse；
103：Position fingerprint base is handled into line density enhancing：
Increase the location fingerprint in database by way of many-sided curve fit models；
Neighborhood Filtering method screening location fingerprint storehouse noise is carried out to the database after increase again：Travel through each in location fingerprint storehouse
Reference point, using current reference point as central point c, and the P neighborhoods based on central point c, calculate central point c and each neighborhood point xpIt
Between average similarityWherein P is default Neighborhood Number, and subscript p is neighborhood point xpPoint identifier,
Represent central point c and neighborhood point xpBetween signal strength vector Euclidean distance；And by each neighborhood point xpP neighborhoods in
Central point c is rejected, and obtains each neighborhood point xp'sA neighborhood pointSubscript j is neighborhood pointPoint identifier, then calculate each neighbour
Domain point xpAverage similarityWhereinRepresent point xpWith pointBetween signal strength vector it is European
Distance, so as to obtain the average of the average similarity of all neighborhood points of central point cIfThen will
Current central point c is denoted as noise z；
The received signal strength value rssi of the received signal strength value renewal noise z of P neighborhoods based on each noise zz,iFor：WhereinRepresent the neighborhood point of noise zTo the received signal strength of i-th of radio reception device
Value, wherein, j represents neighborhood pointPoint identifier,Represent neighborhood pointWeights, andLz,jTable
Show the neighborhood point of noise zEuclidean distance between two coordinate points of noise z, so as to obtain the reference point corresponding to noise z
New signal strength vector；
104：Clustering processing is carried out to position fingerprint base using K mean cluster, and judges whether group's number of all kinds of clusters is no more than group
Number threshold value, if so, then preserving current class cluster, otherwise continues to carry out K mean cluster processing to obtained class cluster, until obtained class
Group's number of cluster is no more than group's number threshold value, so that the location fingerprint storehouse built；
Real-time position matching step：
201：The m group received signal strength values at point to be determined are gathered, obtain the signal strength vector RSSI of point to be determinedt；
202：Weighting k-nearest neighbor based on difference obtains the matching result of point to be determined in the fingerprint base of position：
By the signal strength vector RSSI of point to be determinedtIn each component subtract RSSItIn minimum component, obtain difference value vector
ΔRSSIt；
By RSSItOr Δ RSSItSimilarity mode is carried out with each cluster centre of the class cluster of preservation, searches the highest class of similarity
Cluster；
Wherein similarity mode is：The smaller then similarity of Euclidean distance between signal strength vector is higher；
Based on difference value vector Δ RSSIt, the highest preceding K location fingerprint of similarity is searched in the highest class cluster of similarity, is used
Weighting k nearest neighbor algorithm is weighted addition to the corresponding position coordinates of K location fingerprint, obtains the elements of a fix of anchor point；
203：Latitude and longitude coordinates conversion is carried out to positioning relative coordinate, obtains the positional information of point to be determined.
2. the method as described in claim 1, it is characterised in that neighborhood mode is preferably 8 neighborhoods.
3. method as claimed in claim 1 or 2, it is characterised in that the reference point is：Grid is carried out to indoor plane figure to draw
Point, mesh point is as a reference point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711495301.6A CN107948930B (en) | 2017-12-31 | 2017-12-31 | Indoor positioning optimization method based on position fingerprint algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711495301.6A CN107948930B (en) | 2017-12-31 | 2017-12-31 | Indoor positioning optimization method based on position fingerprint algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107948930A true CN107948930A (en) | 2018-04-20 |
| CN107948930B CN107948930B (en) | 2020-07-17 |
Family
ID=61938193
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711495301.6A Active CN107948930B (en) | 2017-12-31 | 2017-12-31 | Indoor positioning optimization method based on position fingerprint algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107948930B (en) |
Cited By (27)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108596271A (en) * | 2018-05-09 | 2018-09-28 | 中国平安人寿保险股份有限公司 | Appraisal procedure, device, storage medium and the terminal of fingerprint developing algorithm |
| CN108650626A (en) * | 2018-05-18 | 2018-10-12 | 华南师范大学 | A kind of fingerprinting localization algorithm based on Thiessen polygon |
| CN108983254A (en) * | 2018-08-08 | 2018-12-11 | 中国科学院电子学研究所 | Two-dimentional range unit and distance measuring method based on light stream sensor |
| CN109001674A (en) * | 2018-05-31 | 2018-12-14 | 中国矿业大学 | A kind of WiFi finger print information Quick Acquisition and localization method based on continuous videos sequence |
| CN109059919A (en) * | 2018-06-08 | 2018-12-21 | 华中科技大学 | A kind of indoor orientation method based on crowdsourcing sample weighting surface fitting |
| CN109116343A (en) * | 2018-09-11 | 2019-01-01 | 中北大学 | A kind of filtering method of mobile terminal received signal strength |
| CN109587627A (en) * | 2018-12-12 | 2019-04-05 | 嘉兴学院 | The indoor positioning algorithms of terminal heterogeneity are improved based on RSSI |
| CN109739830A (en) * | 2019-02-28 | 2019-05-10 | 电子科技大学 | A kind of location fingerprint database fast construction method based on crowdsourcing data |
| CN109738863A (en) * | 2019-04-08 | 2019-05-10 | 江西师范大学 | A kind of WiFi fingerprint indoor positioning algorithms and system merging confidence level |
| CN110049447A (en) * | 2019-04-12 | 2019-07-23 | 桂林电子科技大学 | A kind of partnership analysis method based on location information |
| CN110166930A (en) * | 2019-04-03 | 2019-08-23 | 华中科技大学 | A kind of indoor orientation method and system based on WiFi signal |
| CN110300372A (en) * | 2019-07-11 | 2019-10-01 | 桂林电子科技大学 | A kind of WIFI indoor orientation method based on location fingerprint |
| CN110381580A (en) * | 2019-07-25 | 2019-10-25 | 上海开域信息科技有限公司 | A kind of WiFi localization method based on ratio optimization |
| WO2020022953A1 (en) * | 2018-07-26 | 2020-01-30 | Singapore Telecommunications Limited | System and method for identifying an internet of things (iot) device based on a distributed fingerprinting solution |
| CN111198365A (en) * | 2020-01-16 | 2020-05-26 | 东方红卫星移动通信有限公司 | Indoor positioning method based on radio frequency signal |
| CN111405461A (en) * | 2020-03-16 | 2020-07-10 | 南京邮电大学 | Wireless indoor positioning method for optimizing equal-interval fingerprint sampling number |
| CN111654843A (en) * | 2019-03-04 | 2020-09-11 | 深圳光启空间技术有限公司 | Method and system for automatically updating fingerprint database and wifi positioning method and system |
| CN111953937A (en) * | 2020-07-31 | 2020-11-17 | 云洲(盐城)创新科技有限公司 | Drowning person lifesaving system and drowning person lifesaving method |
| CN112040397A (en) * | 2020-08-13 | 2020-12-04 | 西北师范大学 | CSI indoor fingerprint positioning method based on adaptive Kalman filtering |
| CN112055308A (en) * | 2020-08-21 | 2020-12-08 | 中通服咨询设计研究院有限公司 | Multi-layer high-robustness fingerprint positioning method |
| CN112261578A (en) * | 2020-10-21 | 2021-01-22 | 南京工业大学 | Indoor fingerprint positioning method based on mode filtering |
| CN112437485A (en) * | 2020-10-29 | 2021-03-02 | 北京邮电大学 | Positioning method and device of fingerprint space interpolation method based on neural network |
| CN112511985A (en) * | 2020-12-21 | 2021-03-16 | 迪爱斯信息技术股份有限公司 | Alarm position positioning method, system, computer equipment and storage medium |
| CN113079466A (en) * | 2020-10-21 | 2021-07-06 | 中移(上海)信息通信科技有限公司 | Fingerprint database construction method, device, equipment and computer storage medium |
| CN113075648A (en) * | 2021-03-19 | 2021-07-06 | 中国舰船研究设计中心 | Clustering and filtering method for unmanned cluster target positioning information |
| CN114630274A (en) * | 2022-03-31 | 2022-06-14 | 大连理工大学 | Precision estimation method for fingerprint positioning |
| CN114845388A (en) * | 2022-05-17 | 2022-08-02 | 电子科技大学 | Indoor positioning method for position fingerprint of sub-direction entropy weighting WKNN |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060089153A1 (en) * | 2004-10-27 | 2006-04-27 | Leonid Sheynblat | Location-sensitive calibration data |
| CN105960021A (en) * | 2016-07-07 | 2016-09-21 | 济南东朔微电子有限公司 | Improved position fingerprint indoor positioning method |
| CN106093852A (en) * | 2016-05-27 | 2016-11-09 | 东华大学 | A kind of method improving WiFi fingerprint location precision and efficiency |
| US20170067982A1 (en) * | 2013-12-26 | 2017-03-09 | Lntel Corporation | Method and apparatus for cross device automatic calibration |
| CN106646338A (en) * | 2016-12-07 | 2017-05-10 | 华南理工大学 | Rapidly accurate indoor location method |
-
2017
- 2017-12-31 CN CN201711495301.6A patent/CN107948930B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060089153A1 (en) * | 2004-10-27 | 2006-04-27 | Leonid Sheynblat | Location-sensitive calibration data |
| US20170067982A1 (en) * | 2013-12-26 | 2017-03-09 | Lntel Corporation | Method and apparatus for cross device automatic calibration |
| CN106093852A (en) * | 2016-05-27 | 2016-11-09 | 东华大学 | A kind of method improving WiFi fingerprint location precision and efficiency |
| CN105960021A (en) * | 2016-07-07 | 2016-09-21 | 济南东朔微电子有限公司 | Improved position fingerprint indoor positioning method |
| CN106646338A (en) * | 2016-12-07 | 2017-05-10 | 华南理工大学 | Rapidly accurate indoor location method |
Non-Patent Citations (3)
| Title |
|---|
| 刘志建 等: "基于克里金空间插值的位置指纹数据库建立算法", 《计算机应用研究》 * |
| 毛勤: "基于WiFi位置指纹的室内定位算法的研究与优化", 《万方学术》 * |
| 郭昕刚 等: "基于边界过滤和邻域均值滤波的室内定位算法", 《长春工业大学学报》 * |
Cited By (40)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108596271B (en) * | 2018-05-09 | 2023-05-26 | 中国平安人寿保险股份有限公司 | Evaluation method and device of fingerprint construction algorithm, storage medium and terminal |
| CN108596271A (en) * | 2018-05-09 | 2018-09-28 | 中国平安人寿保险股份有限公司 | Appraisal procedure, device, storage medium and the terminal of fingerprint developing algorithm |
| CN108650626A (en) * | 2018-05-18 | 2018-10-12 | 华南师范大学 | A kind of fingerprinting localization algorithm based on Thiessen polygon |
| CN109001674A (en) * | 2018-05-31 | 2018-12-14 | 中国矿业大学 | A kind of WiFi finger print information Quick Acquisition and localization method based on continuous videos sequence |
| CN109001674B (en) * | 2018-05-31 | 2022-09-02 | 中国矿业大学 | WiFi fingerprint information rapid acquisition and positioning method based on continuous video sequence |
| CN109059919A (en) * | 2018-06-08 | 2018-12-21 | 华中科技大学 | A kind of indoor orientation method based on crowdsourcing sample weighting surface fitting |
| WO2020022953A1 (en) * | 2018-07-26 | 2020-01-30 | Singapore Telecommunications Limited | System and method for identifying an internet of things (iot) device based on a distributed fingerprinting solution |
| CN108983254A (en) * | 2018-08-08 | 2018-12-11 | 中国科学院电子学研究所 | Two-dimentional range unit and distance measuring method based on light stream sensor |
| CN109116343A (en) * | 2018-09-11 | 2019-01-01 | 中北大学 | A kind of filtering method of mobile terminal received signal strength |
| CN109116343B (en) * | 2018-09-11 | 2022-06-03 | 中北大学 | Filtering method for mobile terminal receiving signal intensity |
| CN109587627A (en) * | 2018-12-12 | 2019-04-05 | 嘉兴学院 | The indoor positioning algorithms of terminal heterogeneity are improved based on RSSI |
| CN109739830A (en) * | 2019-02-28 | 2019-05-10 | 电子科技大学 | A kind of location fingerprint database fast construction method based on crowdsourcing data |
| CN111654843B (en) * | 2019-03-04 | 2024-04-30 | 深圳光启空间技术有限公司 | Method and system for automatically updating fingerprint database, wifi positioning method and system |
| CN111654843A (en) * | 2019-03-04 | 2020-09-11 | 深圳光启空间技术有限公司 | Method and system for automatically updating fingerprint database and wifi positioning method and system |
| CN110166930A (en) * | 2019-04-03 | 2019-08-23 | 华中科技大学 | A kind of indoor orientation method and system based on WiFi signal |
| CN109738863A (en) * | 2019-04-08 | 2019-05-10 | 江西师范大学 | A kind of WiFi fingerprint indoor positioning algorithms and system merging confidence level |
| CN110049447A (en) * | 2019-04-12 | 2019-07-23 | 桂林电子科技大学 | A kind of partnership analysis method based on location information |
| CN110300372A (en) * | 2019-07-11 | 2019-10-01 | 桂林电子科技大学 | A kind of WIFI indoor orientation method based on location fingerprint |
| CN110381580A (en) * | 2019-07-25 | 2019-10-25 | 上海开域信息科技有限公司 | A kind of WiFi localization method based on ratio optimization |
| CN111198365A (en) * | 2020-01-16 | 2020-05-26 | 东方红卫星移动通信有限公司 | Indoor positioning method based on radio frequency signal |
| CN111405461A (en) * | 2020-03-16 | 2020-07-10 | 南京邮电大学 | Wireless indoor positioning method for optimizing equal-interval fingerprint sampling number |
| CN111405461B (en) * | 2020-03-16 | 2021-10-08 | 南京邮电大学 | Wireless indoor positioning method for optimizing equal-interval fingerprint sampling number |
| CN111953937A (en) * | 2020-07-31 | 2020-11-17 | 云洲(盐城)创新科技有限公司 | Drowning person lifesaving system and drowning person lifesaving method |
| CN111953937B (en) * | 2020-07-31 | 2022-11-08 | 云洲(盐城)创新科技有限公司 | Drowning person lifesaving system and drowning person lifesaving method |
| CN112040397A (en) * | 2020-08-13 | 2020-12-04 | 西北师范大学 | CSI indoor fingerprint positioning method based on adaptive Kalman filtering |
| CN112040397B (en) * | 2020-08-13 | 2023-01-24 | 西北师范大学 | CSI indoor fingerprint positioning method based on adaptive Kalman filtering |
| CN112055308A (en) * | 2020-08-21 | 2020-12-08 | 中通服咨询设计研究院有限公司 | Multi-layer high-robustness fingerprint positioning method |
| CN112055308B (en) * | 2020-08-21 | 2024-02-27 | 中通服咨询设计研究院有限公司 | Multi-layer high-robustness fingerprint positioning method |
| CN112261578A (en) * | 2020-10-21 | 2021-01-22 | 南京工业大学 | Indoor fingerprint positioning method based on mode filtering |
| CN113079466B (en) * | 2020-10-21 | 2022-04-26 | 中移(上海)信息通信科技有限公司 | Fingerprint database construction method, device, equipment and computer storage medium |
| CN113079466A (en) * | 2020-10-21 | 2021-07-06 | 中移(上海)信息通信科技有限公司 | Fingerprint database construction method, device, equipment and computer storage medium |
| CN112437485B (en) * | 2020-10-29 | 2021-11-12 | 北京邮电大学 | Positioning method and device of fingerprint space interpolation method based on neural network |
| CN112437485A (en) * | 2020-10-29 | 2021-03-02 | 北京邮电大学 | Positioning method and device of fingerprint space interpolation method based on neural network |
| CN112511985A (en) * | 2020-12-21 | 2021-03-16 | 迪爱斯信息技术股份有限公司 | Alarm position positioning method, system, computer equipment and storage medium |
| CN113075648A (en) * | 2021-03-19 | 2021-07-06 | 中国舰船研究设计中心 | Clustering and filtering method for unmanned cluster target positioning information |
| CN113075648B (en) * | 2021-03-19 | 2024-05-17 | 中国舰船研究设计中心 | Clustering and filtering method for unmanned cluster target positioning information |
| CN114630274A (en) * | 2022-03-31 | 2022-06-14 | 大连理工大学 | Precision estimation method for fingerprint positioning |
| CN114630274B (en) * | 2022-03-31 | 2023-03-14 | 大连理工大学 | Precision estimation method for fingerprint positioning |
| CN114845388A (en) * | 2022-05-17 | 2022-08-02 | 电子科技大学 | Indoor positioning method for position fingerprint of sub-direction entropy weighting WKNN |
| CN114845388B (en) * | 2022-05-17 | 2023-02-28 | 电子科技大学 | Indoor positioning method for position fingerprint of sub-direction entropy weighting WKNN |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107948930B (en) | 2020-07-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107948930A (en) | Indoor positioning optimization method based on location fingerprint algorithm | |
| CN106646338B (en) | A kind of quickly accurate indoor orientation method | |
| CN106131959B (en) | A kind of dual-positioning method divided based on Wi-Fi signal space | |
| CN103747419B (en) | A kind of indoor orientation method based on signal strength difference and dynamic linear interpolation | |
| CN104038901B (en) | Indoor positioning method for reducing fingerprint data acquisition workload | |
| CN107071743A (en) | WiFi localization methods in a kind of quick KNN rooms based on random forest | |
| CN110557716A (en) | Indoor positioning method based on lognormal model | |
| CN110320495A (en) | A kind of indoor orientation method based on Wi-Fi, bluetooth and PDR fusion positioning | |
| CN107677989B (en) | A kind of indoor location localization method carrying out RSSI removal noise based on RSSI maximum value | |
| CN106714109A (en) | WiFi fingerprint database updating method based on crowdsourcing data | |
| CN104105106A (en) | Wireless communication network intelligent-antenna-covered scene automatic classification and recognition method | |
| CN109041206A (en) | A kind of indoor positioning floor method of discrimination based on improvement fuzzy kernel clustering | |
| CN105407529B (en) | Localization Algorithm for Wireless Sensor Networks based on fuzzy C-means clustering | |
| CN110351660B (en) | Bluetooth indoor positioning method based on double-step fingerprint matching architecture | |
| CN107509171A (en) | Indoor orientation method and device | |
| CN110536256A (en) | A kind of indoor orientation method based on double layer grid | |
| CN111901749A (en) | High-precision three-dimensional indoor positioning method based on multi-source fusion | |
| CN108009485A (en) | Wireless fingerprint storehouse update method based on crowdsourcing data | |
| CN112135248A (en) | WIFI fingerprint positioning method based on K-means optimal estimation | |
| CN102831431A (en) | Detector training method based on hierarchical clustering | |
| CN110213710A (en) | A kind of high-performance indoor orientation method, indoor locating system based on random forest | |
| CN110366244A (en) | A kind of WiFi fingerprint indoor orientation method | |
| CN106686720A (en) | Wireless fingerprint positioning method and system based on time dimension | |
| CN109541537B (en) | Universal indoor positioning method based on ranging | |
| CN111757257B (en) | Dynamic fuzzy matching indoor positioning method for overcoming equipment difference |
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 |