CN109041206A - A kind of indoor positioning floor method of discrimination based on improvement fuzzy kernel clustering - Google Patents
A kind of indoor positioning floor method of discrimination based on improvement fuzzy kernel clustering Download PDFInfo
- Publication number
- CN109041206A CN109041206A CN201810710813.8A CN201810710813A CN109041206A CN 109041206 A CN109041206 A CN 109041206A CN 201810710813 A CN201810710813 A CN 201810710813A CN 109041206 A CN109041206 A CN 109041206A
- Authority
- CN
- China
- Prior art keywords
- floor
- cluster centre
- reference point
- rssi
- fingerprint database
- 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.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W64/00—Locating users or terminals or network equipment for network management purposes, e.g. mobility management
-
- 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
-
- 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 present invention discloses a kind of based on the indoor positioning floor method of discrimination for improving fuzzy kernel clustering, the following steps are included: 1, in each floor of localization region, acquire visible WiFi access points (Access Point in the only floor, AP received signal strength (RSSI)) establishes the corresponding home position-fingerprint database of each floor and all AP in the signal intensity profile range data library of floor gap;2, using improved fuzzy kernel clustering algorithm, home position-fingerprint database is clustered, obtains position-fingerprint database with classification marker;3, according to the WiFi signal intensity data acquired in real time, floor where judging tested point using rough sort and cluster centre.The present invention is based on the thought of rough sort and cluster, can position indoors in quickly and effectively judge floor where tested point.
Description
Technical field
It is the invention belongs to indoor positioning technologies field, in particular to a kind of based on the indoor positioning building for improving fuzzy kernel clustering
Layer method of discrimination.
Background technique
Indoor positioning has a good application prospect, and has a large amount of scholars to probe into it, but in previous correlative study
In, it focuses mostly in the plane positioning precision and performance of same floor, elevation information is often ignored or is set using additional hardware
Standby instruction, or even need artificial specified.And in large-scale more floor indoor spaces, elevation information is the important of location-aware services
Component part, traditional location-aware services based on two-dimensional surface are not able to satisfy user demand.Therefore, the interior of more floors
Locating scheme is gradually concerned by people.
WiFi signal because have wide coverage, network deployment it is at low cost, can be used for the advantages that communicating, indoors positioning and
Tracking is concerned in field.The position based on signal strength is mostly used to refer to greatly currently based on the indoor orientation method of WiFi signal
Line method, and when WiFi signal often passes through a floor, can all occur acutely decaying, this leads to the signal of the different same AP of floor gap
Intensity has apparent difference, can use this phenomenon and carries out floor differentiation, realizes more floor indoor positionings.
But be guarantee network in general building without dead angle covering and communication quality, arrange sufficient amount
Enterprise-level AP, signal penetration power is stronger, therefore the signal of the same AP possibly is present at multiple floors, and in contiguous floors
Between difference be not particularly evident.If all AP selected in localization region carry out floor differentiation, there can be serious letter
Redundancy phenomena is ceased, and may cause floor and differentiate that result is not accurate enough.
Summary of the invention
Goal of the invention: aiming at the problems existing in the prior art, the present invention provides a kind of based on improving fuzzy kernel clustering
Indoor positioning floor method of discrimination must can quick and precisely realize that the floor in indoor positioning differentiates.
Technical solution: in order to solve the above technical problems, the present invention provide it is a kind of fixed based on the interior for improving fuzzy kernel clustering
Position floor method of discrimination, includes the following steps:
(1) in the indoor positioning region of p floor, q visible AP are distributed with, choose location reference point in each floor,
Home position-the fingerprint database for establishing each floor after the RSSI of the visible AP of each floor is acquired, after acquisition, analyzes and establishes
The distribution database of each all AP of floor gap;
(2) home position-fingerprint database to each floor established in step (1), using optimization fast search and
It was found that density peaks, that is, CFSFDP method, determines the clusters number and cluster centre of each floor, by the cluster of each floor of gained
The input parameter of number and cluster centre as fuzzy kernel clustering algorithm, and using the cluster result of fuzzy kernel clustering, in each building
Classification marker is added in home position-fingerprint database of layer, forms the fingerprint database for having classification marker;
(3) one group of RSSI value from all visible AP, the first distribution model with all AP of floor gap are acquired on tested point
It encloses database to be compared, obtains rough sort as a result, directly determining floor where tested point if result is unique, otherwise, utilize
Cluster centre determines floor where tested point.
Further, the home position-fingerprint database and all AP of each floor gap of each floor are established in the step (1)
Distribution database specific step is as follows:
(1.1) firstly, the q visible AP to localization region are numbered according to MAC Address from 1~q, then, to each
Floor definition set(f=1,2 ..., p), QfIndicate the volume of the visible AP of each floor
Number set, wherein p be total number of floor levels, f is floor number, qfFor the number of visible AP in f floor, set element nums's
Value range is 1≤nums≤ q, s are positive integer;The RSSI value that k WiFi is acquired at each reference point obtains a k × qf
Rank matrix, the number that comes from that the t row s column of matrix indicate that the t times receives is numsVisible AP RSSI value;K, qf, t,
S is positive integer, 1≤t≤k, 1≤s≤qf;
(1.2) k × q that will be obtained at each reference pointfThe all elements of rank matrix column vector are added
After be averaged, obtain the finger print information of the reference point, the finger print information be a qfTie up row vectorThe s column element expression of the row vector carries out k times at reference point nf, i
The n-th um is come from after samplingsThe RSSI average value of a AP;Then by the two-dimensional position coordinate l of the reference pointf,i(xf,i,yf,i) with
Where finger print information is stored in the reference point in home position-fingerprint database of floor;After acquisition process
Form home position-fingerprint database of each floor;If can't detect certain AP in reference point, default value -110dBm is used
Instead of;
(1.3) for the q visible AP of localization region, according to home position-fingerprint database information of each floor,
The RSSI range of j-th of AP of f floor is1≤j≤q, if some AP of the floor is invisible,
Its RSSI range is set as (- 110dBm, -110dBm), thus constructs all AP in the distribution database D R of each floor,
Wherein DR is the space of a p × 2q, and odd column stores the minimum value of each AP RSSI range, and even column stores each AP
The maximum value of RSSI range.
Further, home position-finger print data of each floor is clustered in the step (2), by raw bits
Set-fingerprint database in be added classification marker specific step is as follows:
(2.1) using the finger print data in home position-fingerprint database of f floor as the input of CFSFDP algorithm,
So that it is determined that clusters number and cluster centre, wherein by reference point nf,iOffice as the reference point of the sum of K nearest neighbor distance
Portion's density pi, i.e.,R ∈ KNN (i) indicates reference point nf,iThe reference point serial number of K neighbour, d (nf,i,
nf,r) indicate reference point nf,iAnd nf,rBetween Euclidean distance, specific formula for calculation is
,
It is optimized in the way of this to be suitable for reference point and choose non-uniform positioning scene;
(2.2) using obtained clusters number and cluster centre as the clusters number of fuzzy kernel clustering algorithm and initial clustering
Center obtains final cluster centre after updating iteration to degree of membership and cluster centre, and Kernel Function is chosen for Gaussian kernel letter
Number;
(2.3) other reference points for f floor in addition to cluster centre are calculated European with each cluster centre
Distance, and it is classified as same class with apart from nearest cluster centre, and add in original-location fingerprint database of the floor
Add category label, generates the position-fingerprint database for having classification marker.
Further, the specific step of floor where determining tested point using rough sort and cluster centre in the step (3)
It is rapid as follows:
(3.1) tested point acquires one group of RSSI data from all AP;
(3.2) it is compared with database D R, by judging that the affiliated range of each AP signal strength carries out rough sort;
(3.3) judge whether classification results unique, if result uniquely if using rough sort result as floor differentiate as a result,
(3.4) are entered step if result is not unique;
(3.4) nearest cluster centre is calculated, differentiates result for floor number where the cluster centre as floor.
Further, the detailed step of floor where determining tested point using rough sort and cluster centre in the step (3)
It is rapid as follows:
(3.1) it acquires one group of RSSI data from all AP in real time at tested point, constitutes online fingerprint vector
(3.2) by comparing with database D R, judge collected signal strengthAffiliated range, if belonging to
The range, i.e., for a certain floor f,Then floor score+1, by highest scoring
Floor number differentiates result as floor rough sort;
(3.3) if occurring, two or more floor scores are identical, the cluster result that is obtained according to step (2) and it is European away from
From calculation formula, cluster centre nearest therewith is calculated, returns to floor number where this cluster centre;In finding nearest cluster
When the heart, the RSSI vector dimension acquired in real time is q, and the dimension of the sub- location fingerprint database of each floor is qf, the two is different
It causes, the set Q defined according to step (1.1)f, accordingly lower target element is taken in fingerprint vector S, by its dimensionality reduction to calculate
The consistent dimension of floor, then calculate Euclidean distance between the two.
Compared with the prior art, the advantages of the present invention are as follows: the present invention utilizes existing hardware portion in capable of positioning indoors
Administration's facility fast and accurately differentiates floor, realizes that the floor in more floor indoor positionings differentiates, and improve floor differentiation
Accuracy.
Detailed description of the invention
Fig. 1 is overview flow chart of the invention;
Fig. 2 is the floor map in 1 building, indoor positioning region in specific embodiment;
Fig. 3 is the floor map in 2 building, indoor positioning region in specific embodiment;
Fig. 4 is the floor map in 3 building, indoor positioning region in specific embodiment;
Fig. 5 is the flow chart of floor where determining tested point using rough sort and cluster centre in Fig. 1.
Specific embodiment
With reference to the accompanying drawings and detailed description, the present invention is furture elucidated.Embodiments described herein are only
A part of the embodiments of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, ordinary skill people
Member's obtained other embodiments without making creative work belong to the range that the present invention is protected.
It is a kind of based on the indoor positioning floor method of discrimination for improving fuzzy kernel clustering as shown in Figure 1:, comprising the following steps:
Step 1: in the indoor positioning region of p floor, q visible AP are distributed with, wherein the visible AP of f floor
Number is qf(1≤f≤p), chooses n in the floorfA reference point, disease acquire k times in each reference point and come from qfA AP's
RSSI value establishes home position-fingerprint database of the floor after carrying out data processing, and establishes point of all AP of each floor gap
Cloth range data library;The detailed process of the step are as follows:
Step 1.1: firstly, the q visible AP to localization region are numbered according to MAC Address from 1~q, then, to every
One floor definition set(f=1,2 ..., p), QfIndicate the visible AP's of each floor
Number set, wherein p is total number of floor levels, and f is floor number, qfFor the number of visible AP in f floor, set element nums
Value range be 1≤nums≤ q, s are positive integer.The RSSI value that k WiFi is acquired at each reference point obtains a k
×qfRank matrix, the number that comes from that the t row s column of matrix indicate that the t times receives is numsVisible AP RSSI value;K,
qf, t, s are positive integer, 1≤t≤k, 1≤s≤qf;
Step 1.2: the k × q that will be obtained at each reference pointfThe all elements of rank matrix column vector take after being added
Average value obtains the finger print information of the reference point, which is a qfTie up row vectorThe s column element of the row vector is indicated in reference point nf,iPlace carries out k times
The n-th um is come from after samplingsThe RSSI average value of a AP.Then by the two-dimensional position coordinate l of the reference pointf,i(xf,i,yf,i) with
Where finger print information is stored in the reference point in home position-fingerprint database of floor.After acquisition process
Form home position-fingerprint database of each floor.If can't detect certain AP in reference point, default value -110dBm is used
Instead of;
Step 1.3: for q visible AP of localization region, according to home position-fingerprint database letter of each floor
Breath, the RSSI range of j-th of AP of f floor are1≤j≤q, if some AP of the floor can not
See, its RSSI range is set as (- 110dBm, -110dBm), thus constructs all AP in the distribution database of each floor
DR, wherein DR is the space of a p × 2q, and odd column stores the minimum value of each AP RSSI range, and even column stores each AP
The maximum value of RSSI range.This implementation steps provides location fingerprint database sample for subsequent cluster and position matching, and builds
The distribution fingerprint base DR of all AP of floor gap needed for floor method of discrimination is found.
Step 2: the poly- of each floor is determined using CFSFDP method to home position-fingerprint database of each floor
Class number and cluster centre are joined using the clusters number of each floor of gained and cluster centre as the input of fuzzy kernel clustering algorithm
Number, and using the cluster result of fuzzy kernel clustering, classification marker is added in home position-fingerprint database of each floor;It should
The detailed process of step are as follows:
Step 2.1: using the finger print data in home position-fingerprint database of f floor as the defeated of CFSFDP algorithm
Enter, so that it is determined that clusters number and cluster centre, wherein by reference point nf,iThe sum of K nearest neighbor distance as the reference point
Local density ρi, i.e.,R ∈ KNN (i) indicates reference point nf,iThe reference point serial number of K neighbour, d
(nf,i,nf,r) indicate reference point nf,iAnd nf,rBetween Euclidean distance, specific formula for calculation is
,
It is optimized in the way of this to be suitable for reference point and choose non-uniform positioning scene;
Step 2.2: using obtained clusters number and cluster centre as the clusters number of fuzzy kernel clustering algorithm and initially
Cluster centre obtains final cluster centre after updating iteration to degree of membership and cluster centre, and Kernel Function is chosen for Gauss
Kernel function.
Step 2.3: for other reference points of the f floor in addition to cluster centre, calculating and each cluster centre
Euclidean distance, and it is classified as same class with apart from nearest cluster centre, and in original-location fingerprint database of the floor
Middle addition category label generates the position-fingerprint database for having classification marker.
This implementation steps carries out effectively accurately classification to the location fingerprint data of localization region, and can be suitably used for reference point
The case where choosing Density inhomogeneity, can improve the precision of cluster result to a certain extent, be that tested point classification is drawn in step 3
Divide and lays the foundation.
Step 3: choosing a test point in positioning scene at random, and one group is acquired on tested point from all visible AP
RSSI value, be first compared with the distribution database of all AP of floor gap, obtain rough sort as a result, if result is unique,
Floor where then directly determining tested point, otherwise, floor where determining tested point using cluster centre;The detailed process of the step
Are as follows:
Step 3.1: acquiring one group of RSSI data from all AP in real time at tested point, constitute online fingerprint vector
Step 3.2: by comparing with database D R, judging collected signal strengthAffiliated range, if belong to
In the range, i.e., for a certain floor f,Then floor score+1, by highest scoring
Floor number as floor rough sort differentiate result;
Step 3.3: if occurring, two or more floor scores are identical, the cluster result that is obtained according to step 2 and European
Distance calculation formula calculates cluster centre nearest therewith, returns to floor number where this cluster centre.Finding nearest cluster
When center, the RSSI vector dimension acquired in real time is q, and the dimension of the sub- location fingerprint database of each floor is qf, the two is not
Unanimously, the set Q defined according to step 1.1f, accordingly lower target element is taken in fingerprint vector S, by its dimensionality reduction to calculate
The consistent dimension of floor, then calculate Euclidean distance between the two.
Example:
In a preferred embodiments of the invention, tested in the indoor positioning scene shown in Fig. 2-4, dotted line in figure
Indicate that reference point acquires path, each reference point spacing distance is 1m;A total of 3 floors of the scene, the area 50m of every floor
× 20m, mainly there is office, meeting room, toilet, several layouts of stair, and wireless access point unifies portion in corridor using school
The SSID of administration is the access point of " seu-wlan ", and installation position is unknown, and 1,2,3 building visible AP number is respectively 48,51,42, three
A total of 82 visible AP of floor.Red dotted line indicates that reference point acquires path, between neighboring reference point between be divided into 1m, acquire altogether
171 reference points, wherein 1,2,3 building reference point number is respectively 50,31,90.
Using Huawei's honor 8 (youth version) smart phone, LIPS software is write, acquires to come in the reference point of each floor
From the RSSI value of the visible AP of the only floor, sampling period 1s, each reference point acquires 60 groups of RSSI values and correlation AP
Information.After the finger print information of all reference points is averaged according to step 1.2, it is stored into the floor jointly with position coordinates
Position-fingerprint database, as shown in table 1, wherein APsIndicate s-th of AP, by MAC information to AP all in positioning scene into
Row number.According to the RSSI information of each floor reference point, all AP are established in the distribution number of each floor gap according to step 1.3
According to library, as shown in table 2.Using home position-fingerprint database data of each floor as the defeated of improvement fuzzy kernel clustering algorithm
Enter data, carries out data clusters, establish location fingerprint database of each floor with classification marker.
In experimental situation of the invention, 45 test points for randomly selecting different floors are tested, each test point
RSSI value 30 times from 82 AP are acquired, the online fingerprint vector of RSSI of 45 × 30=1350 tested point is always obtained, presses
Floor differentiation is carried out according to step 3.By statistics, floor differentiates that accuracy rate can achieve 97.8%.The present invention is by rough sort and changes
Cluster centre into fuzzy kernel clustering is used as the floor differentiation in indoor positioning, and floor differentiates that accuracy rate is higher, and can fit
Non-uniform positioning scene is chosen for reference point.
The foregoing is merely presently preferred embodiments of the present invention, is merely illustrative and not restrictive for the invention.
It modifies within all principles defined by the claims in the present invention to it, change or equivalent replacement should be included in this hair
Within bright protection scope.The content that the present invention is not elaborated, which belongs to, has skill well known to this professional domain technical staff
Art.
Table 1 is that each reference point carries out the location fingerprint data (RSSI signal unit: dBm) after 60 acquisitions averagely
Floor1:
Floor2:
x | y | AP1 | AP49 | AP3 | AP4 | AP5 | … | AP44 | AP45 | AP46 | AP47 | AP48 |
10 | 20 | -95 | -66 | -72 | -49 | -110 | … | -75 | -110 | -81 | -110 | -110 |
10 | 21 | -90 | -70 | -110 | -44 | -110 | … | -74 | -110 | -110 | -110 | -110 |
10 | 22 | -93 | -61 | -70 | -56 | -110 | … | -74 | -110 | -110 | -110 | -81 |
… | … | … | … | … | … | … | … | … | … | … | … | … |
Floor3:
x | y | AP49 | AP7 | AP65 | AP66 | AP67 | … | AP42 | AP81 | AP44 | AP46 | AP82 |
18 | 17 | -65 | -83 | -110 | -110 | -110 | … | -110 | -68 | -70 | -110 | -110 |
17 | 17 | -71 | -85 | -110 | -110 | -110 | … | -110 | -67 | -65 | -110 | -110 |
16 | 17 | -61 | -88 | -110 | -110 | -110 | … | -110 | -64 | -68 | -110 | -110 |
… | … | … | … | … | … | … | … | … | … | … | … | .. |
Table 2 is the distribution data (RSSI signal unit: dBm) of all AP of floor gap
Floor | rssiminAP1 | rssimaxAP1 | rssiminAP2 | rssimaxAP2 | … | rssiminAP82 | rssimaxAP82 |
1 | -110 | -90 | -79 | -63.73 | … | -110 | -110 |
2 | -110 | -79 | -110 | -110 | … | -110 | -110 |
3 | -110 | -110 | -110 | -110 | … | -110 | -39.27 |
Claims (5)
1. a kind of based on the indoor positioning floor method of discrimination for improving fuzzy kernel clustering, which comprises the steps of:
(1) in the indoor positioning region of p floor, q visible AP are distributed with, choose location reference point, acquisition in each floor
Home position-fingerprint database that each floor is established after the RSSI of the visible AP of each floor, after acquisition, analyzes and establishes each building
The distribution database of all AP of interlayer;
(2) home position-fingerprint database to each floor established in step (1), fast search and discovery using optimization
Density peaks, that is, CFSFDP method determine the clusters number and cluster centre of each floor, by the clusters number of each floor of gained
Input parameter with cluster centre as fuzzy kernel clustering algorithm, and using the cluster result of fuzzy kernel clustering, in each floor
Classification marker is added in home position-fingerprint database, forms the fingerprint database for having classification marker;
(3) one group of RSSI value from all visible AP, the first distribution number with all AP of floor gap are acquired on tested point
It is compared according to library, obtains rough sort as a result, if result is unique otherwise floor where directly determining tested point utilizes cluster
Center determines floor where tested point.
2. according to claim 1 a kind of based on the indoor positioning floor method of discrimination for improving fuzzy kernel clustering, feature
It is, home position-fingerprint database of each floor and the distribution number of all AP of each floor gap is established in the step (1)
According to library, specific step is as follows:
(1.1) firstly, the q visible AP to localization region are numbered according to MAC Address from 1~q, then, to each floor
Definition setQfIndicate the number set of the visible AP of each floor,
Wherein p is total number of floor levels, and f is floor number, qfFor the number of visible AP in f floor, set element numsValue range
For 1≤nums≤ q, s are positive integer;The RSSI value that k WiFi is acquired at each reference point obtains a k × qfRank matrix,
The number that comes from that the t row s column of matrix indicate that the t times receives is numsVisible AP RSSI value;K, qf, t, s are positive
Integer, 1≤t≤k, 1≤s≤qf;
(1.2) k × q that will be obtained at each reference pointfThe all elements of rank matrix column vector are averaged after being added, and are obtained
The finger print information of the reference point, the finger print information are a qfTie up row vector
The s column element expression of the row vector carries out coming from the n-th um after k sampling at reference point nf, isThe RSSI of a AP is average
Value;Then by the two-dimensional position coordinate l of the reference pointf,i(xf,i,yf,i) original of reference point place floor is stored in finger print information
In beginning position-fingerprint database;Home position-fingerprint database of each floor is formed after all referring to acquisition process;
If can't detect certain AP in reference point, replaced with default value -110dBm;
(1.3) for q visible AP of localization region, according to home position-fingerprint database information of each floor, f
The RSSI range of j-th of AP of floor is If some AP of the floor is invisible,
Its RSSI range is set as (- 110dBm, -110dBm), thus constructs all AP in the distribution database D R of each floor,
Wherein DR is the space of a p × 2q, and odd column stores the minimum value of each AP RSSI range, and even column stores each AP
The maximum value of RSSI range.
3. according to claim 1 a kind of based on the indoor positioning floor method of discrimination for improving fuzzy kernel clustering, feature
It is, home position-finger print data of each floor is clustered in the step (2), by home position-fingerprint database
Specific step is as follows for middle addition classification marker:
(2.1) using the finger print data in home position-fingerprint database of f floor as the input of CFSFDP algorithm, thus
Determine clusters number and cluster centre, wherein by reference point nf,iThe sum of K nearest neighbor distance it is close as the part of the reference point
Spend ρi, i.e.,R ∈ KNN (i) indicates reference point nf,iThe reference point serial number of K neighbour, d (nf,i,nf,r)
Indicate reference point nf,iAnd nf,rBetween Euclidean distance, specific formula for calculation is
,
It is optimized in the way of this to be suitable for reference point and choose non-uniform positioning scene;
(2.2) using obtained clusters number and cluster centre as in the clusters number and initial clustering of fuzzy kernel clustering algorithm
The heart obtains final cluster centre after updating iteration to degree of membership and cluster centre, and Kernel Function is chosen for gaussian kernel function;
(2.3) other reference points for f floor in addition to cluster centre, calculate with each cluster centre it is European away from
From, and it is classified as same class with apart from nearest cluster centre, and add in original-location fingerprint database of the floor
Category label generates the position-fingerprint database for having classification marker.
4. according to claim 1 a kind of based on the indoor positioning floor method of discrimination for improving fuzzy kernel clustering, feature
It is, specific step is as follows for floor where determining tested point using rough sort and cluster centre in the step (3):
(3.1) tested point acquires one group of RSSI data from all AP;
(3.2) it is compared with database D R, by judging that the affiliated range of each AP signal strength carries out rough sort;
(3.3) judge whether classification results unique, if result uniquely if using rough sort result as floor differentiate as a result, if
As a result (3.4) are not uniquely entered step then;
(3.4) nearest cluster centre is calculated, differentiates result for floor number where the cluster centre as floor.
5. according to claim 1 a kind of based on the indoor positioning floor method of discrimination for improving fuzzy kernel clustering, feature
It is, the detailed step of floor is as follows where determining tested point using rough sort and cluster centre in the step (3):
(3.1) it acquires one group of RSSI data from all AP in real time at tested point, constitutes online fingerprint vector
(3.2) by comparing with database D R, judge collected signal strengthAffiliated range, if belonging to the model
It encloses, i.e., for a certain floor f,Then floor score+1, by the floor of highest scoring
Number as floor rough sort differentiate result;
(3.3) if occurring, two or more floor scores are identical, the cluster result and Euclidean distance meter obtained according to step (2)
Formula is calculated, cluster centre nearest therewith is calculated, returns to floor number where this cluster centre;Finding nearest cluster centre
When, the RSSI vector dimension acquired in real time is q, and the dimension of the sub- location fingerprint database of each floor is qf, the two is inconsistent,
The set Q defined according to step (1.1)f, accordingly lower target element is taken in fingerprint vector S, by its dimensionality reduction to building to be calculated
The consistent dimension of layer, then calculate Euclidean distance between the two.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810710813.8A CN109041206A (en) | 2018-07-03 | 2018-07-03 | A kind of indoor positioning floor method of discrimination based on improvement fuzzy kernel clustering |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810710813.8A CN109041206A (en) | 2018-07-03 | 2018-07-03 | A kind of indoor positioning floor method of discrimination based on improvement fuzzy kernel clustering |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109041206A true CN109041206A (en) | 2018-12-18 |
Family
ID=65522315
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810710813.8A Pending CN109041206A (en) | 2018-07-03 | 2018-07-03 | A kind of indoor positioning floor method of discrimination based on improvement fuzzy kernel clustering |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109041206A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110231593A (en) * | 2019-04-11 | 2019-09-13 | 深圳市城市交通规划设计研究中心有限公司 | Indoor orientation method, device, computer readable storage medium and terminal device |
CN110958584A (en) * | 2019-12-05 | 2020-04-03 | 江南大学 | Hierarchical classification indoor positioning method based on received signal strength |
CN111639712A (en) * | 2020-05-29 | 2020-09-08 | 北斗数睿(北京)科技有限公司 | Positioning method and system based on density peak clustering and gradient lifting algorithm |
WO2020232869A1 (en) * | 2019-05-21 | 2020-11-26 | 平安科技(深圳)有限公司 | Indoor positioning method, apparatus, system and device, and storage medium |
CN112738714A (en) * | 2020-12-29 | 2021-04-30 | 北京邮电大学 | Floor recognition method for building, electronic device and computer readable 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 |
CN113079468A (en) * | 2021-03-23 | 2021-07-06 | 西安交通大学 | Indoor positioning method and system based on Wifi signal RSSI (received Signal Strength indicator) characteristics |
CN113905326A (en) * | 2021-11-15 | 2022-01-07 | 广州瀚信通信科技股份有限公司 | Indoor user position positioning method and device based on 5G network |
CN114286282A (en) * | 2021-11-18 | 2022-04-05 | 中国科学院空天信息创新研究院 | Fingerprint positioning method based on WiFi RSS data of mobile phone and dimension reduction algorithm |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102056293A (en) * | 2010-12-10 | 2011-05-11 | 北京星网锐捷网络技术有限公司 | Radio frequency fingerprint positioning method and system as well as access controller (AC) and positioning server |
CN102135429A (en) * | 2010-12-29 | 2011-07-27 | 东南大学 | Robot indoor positioning and navigating method based on vision |
CN103916821A (en) * | 2014-04-11 | 2014-07-09 | 北京工业大学 | Floor distinguishing method based on RSSI difference between floors |
CN108151744A (en) * | 2017-12-20 | 2018-06-12 | 四川长虹电器股份有限公司 | Indoor navigation system based on A-STAR algorithms |
CN108156580A (en) * | 2017-12-19 | 2018-06-12 | 江西师范大学 | A kind of indoor floor location method based on smart mobile phone |
-
2018
- 2018-07-03 CN CN201810710813.8A patent/CN109041206A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102056293A (en) * | 2010-12-10 | 2011-05-11 | 北京星网锐捷网络技术有限公司 | Radio frequency fingerprint positioning method and system as well as access controller (AC) and positioning server |
CN102135429A (en) * | 2010-12-29 | 2011-07-27 | 东南大学 | Robot indoor positioning and navigating method based on vision |
CN103916821A (en) * | 2014-04-11 | 2014-07-09 | 北京工业大学 | Floor distinguishing method based on RSSI difference between floors |
CN108156580A (en) * | 2017-12-19 | 2018-06-12 | 江西师范大学 | A kind of indoor floor location method based on smart mobile phone |
CN108151744A (en) * | 2017-12-20 | 2018-06-12 | 四川长虹电器股份有限公司 | Indoor navigation system based on A-STAR algorithms |
Non-Patent Citations (2)
Title |
---|
杜凯颖 等: "基于改进模糊核聚类的室内定位方法研究", 《测控技术》 * |
邓中亮 等: "一种基于K-means算法的WLAN室内定位楼层判别方法", 《软件》 * |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110231593A (en) * | 2019-04-11 | 2019-09-13 | 深圳市城市交通规划设计研究中心有限公司 | Indoor orientation method, device, computer readable storage medium and terminal device |
WO2020232869A1 (en) * | 2019-05-21 | 2020-11-26 | 平安科技(深圳)有限公司 | Indoor positioning method, apparatus, system and device, and storage medium |
CN110958584A (en) * | 2019-12-05 | 2020-04-03 | 江南大学 | Hierarchical classification indoor positioning method based on received signal strength |
CN110958584B (en) * | 2019-12-05 | 2021-06-25 | 江南大学 | Hierarchical classification indoor positioning method based on received signal strength |
CN111639712A (en) * | 2020-05-29 | 2020-09-08 | 北斗数睿(北京)科技有限公司 | Positioning method and system based on density peak clustering and gradient lifting algorithm |
CN112738714B (en) * | 2020-12-29 | 2021-11-30 | 北京邮电大学 | Floor recognition method for building, electronic device and computer readable storage medium |
CN112738714A (en) * | 2020-12-29 | 2021-04-30 | 北京邮电大学 | Floor recognition method for building, electronic device and computer readable 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 |
CN113079468A (en) * | 2021-03-23 | 2021-07-06 | 西安交通大学 | Indoor positioning method and system based on Wifi signal RSSI (received Signal Strength indicator) characteristics |
CN113079468B (en) * | 2021-03-23 | 2022-12-09 | 西安交通大学 | Indoor positioning method and system based on Wifi signal RSSI (received Signal Strength indicator) characteristics |
CN113905326A (en) * | 2021-11-15 | 2022-01-07 | 广州瀚信通信科技股份有限公司 | Indoor user position positioning method and device based on 5G network |
CN113905326B (en) * | 2021-11-15 | 2023-07-14 | 广州瀚信通信科技股份有限公司 | Indoor user position locating method and device based on 5G network |
CN114286282A (en) * | 2021-11-18 | 2022-04-05 | 中国科学院空天信息创新研究院 | Fingerprint positioning method based on WiFi RSS data of mobile phone and dimension reduction algorithm |
CN114286282B (en) * | 2021-11-18 | 2023-12-26 | 中国科学院空天信息创新研究院 | Fingerprint positioning method based on WiFi RSS data and dimension reduction algorithm of mobile phone |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109041206A (en) | A kind of indoor positioning floor method of discrimination based on improvement fuzzy kernel clustering | |
CN107948930A (en) | Indoor positioning optimization method based on location fingerprint algorithm | |
CN107241700B (en) | Indoor positioning method based on CSI space-frequency characteristic and reference point position clustering algorithm | |
CN103747524B (en) | A kind of Android terminal indoor orientation method based on cloud platform | |
CN106131959B (en) | A kind of dual-positioning method divided based on Wi-Fi signal space | |
CN105223546B (en) | Indoor orientation method based on received signal strength and reference point locations double focusing class | |
CN110557716B (en) | Indoor positioning method based on lognormal model | |
He et al. | Tilejunction: Mitigating signal noise for fingerprint-based indoor localization | |
CN104302000B (en) | The indoor orientation method of correlation is indicated based on signal receiving strength | |
CN108696932A (en) | It is a kind of using CSI multipaths and the outdoor fingerprint positioning method of machine learning | |
CN106093852A (en) | A kind of method improving WiFi fingerprint location precision and efficiency | |
CN105636201B (en) | Indoor orientation method based on sparse signal fingerprint database | |
CN104540221B (en) | WLAN indoor orientation methods based on semi-supervised SDE algorithms | |
CN105717483B (en) | A kind of location determining method and device based on multi-source positioning method | |
CN104038901B (en) | Indoor positioning method for reducing fingerprint data acquisition workload | |
Ding et al. | Efficient indoor fingerprinting localization technique using regional propagation model | |
CN107182036A (en) | The adaptive location fingerprint positioning method merged based on multidimensional characteristic | |
Liu et al. | Localization algorithm of indoor Wi-Fi access points based on signal strength relative relationship and region division. | |
CN109511085B (en) | UWB fingerprint positioning method based on MeanShift and weighted k nearest neighbor algorithm | |
CN107677989B (en) | A kind of indoor location localization method carrying out RSSI removal noise based on RSSI maximum value | |
CN104581945B (en) | The WLAN indoor orientation methods of semi-supervised APC clustering algorithms based on distance restraint | |
Meng et al. | Optimized access points deployment for WLAN indoor positioning system | |
CN109059919B (en) | Indoor positioning method based on crowdsourcing sample weighted surface fitting | |
CN109945865A (en) | The indoor orientation method merged based on WiFi with earth magnetism | |
CN109348416B (en) | Fingerprint indoor positioning method based on binary k-means |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181218 |