CN106792510B - A kind of prediction type fingerprint image searching method in fingerprint location - Google Patents
A kind of prediction type fingerprint image searching method in fingerprint location Download PDFInfo
- Publication number
- CN106792510B CN106792510B CN201611061169.3A CN201611061169A CN106792510B CN 106792510 B CN106792510 B CN 106792510B CN 201611061169 A CN201611061169 A CN 201611061169A CN 106792510 B CN106792510 B CN 106792510B
- Authority
- CN
- China
- Prior art keywords
- reference point
- search
- positioning
- euclidean distance
- fingerprint
- 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.)
- Active
Links
Classifications
-
- 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
-
- 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
- H04W4/029—Location-based management or tracking services
-
- H04W4/04—
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention belongs to radionavigation field of locating technology, the prediction type fingerprint image searching method in a kind of fingerprint location is provided.The present invention using the location information predicted using filtering algorithm as prior information, centered on prior information, is scanned for by way of to ambient radiation first.K reference point RP nearest away from signal characteristic TP in signal space Euclidean distance is constantly updated in search process.When continuous search L times and K RP does not change, then assert that this K RP is finally to select, for the reference point of location estimation, and bring into KNN algorithm and resolved, the present invention can replace traditional to fingerprint image progress sequential search method.Beneficial effects of the present invention are that can be improved positioning accuracy, compared with the conventional method of not sub-clustering, search speed is faster compared with the conventional method positioned after sub-clustering.
Description
Technical field
The invention belongs to radionavigation field of locating technology, are related to fingerprint location technology, are related specifically to a kind of finger
Prediction type fingerprint image searching method in line positioning.
Background technique
Fingerprint location technology is widely used in indoor positioning field, which is divided into two stages:
First stage is off-line phase, both in given area, by means such as point-to-point measurements, by physical spatial location with
Mapping relations are established between signal space feature, user can estimate self-position by comparing the mapping relations.Specifically: it is empty
Between establish mapping relations, stored by way of database, and store database be referred to as fingerprint image;If fingerprint image is by trepang
Examination point (RP) is constituted, and each RP stores its physical address and signal characteristic vector information.
Second stage is the tuning on-line stage, and system carries out the RP in the signal characteristic (TP) and fingerprint image that acquire side face
The calculating of space Euclidean distance obtains its similarity, to record K nearest RP of the Euclidean distance between TP.It is calculated by KNN
K RP value of gained is estimated that obtained position is position location by method.Positioning system is caused to mitigate the resolving time
Positioning result delay, filtering algorithm can be introduced toward contact in system, according to the preceding result positioned several times to user current location
It is predicted, to improve positioning accuracy.
The common fingerprint image problems of too of fingerprint location, excessive fingerprint image causes volumes of searches huge when both searching for every time, makes
It obtains locating speed and standard is not achieved.Conventional solutions are sub-clustering: the vector with similar signal feature being classified as cluster, is positioned
When first pass through the Euclidean distance calculated with cluster head, find out the cluster where itself, then carry out on the fingerprint image of the cluster traversal and search
Rope, to reach search classification, reduces workload to position.
But there are limitations for common cluster-dividing method: sub-clustering excessively causes sub-clustering to lose meaning, and sub-clustering is excessively few to be then not achieved drop
The effect of low volumes of searches, while bring cluster edge blurring problem, so that sub-clustering positioning easily causes error.Even if preparatory when sub-clustering
This situation can not be avoided completely by drawing crossover region.In addition, if occurring to will lead to biggish positioning if cluster head identification mistake
Error.Current sub-clustering means include the artificial sub-clustering of Subjective progress according to designer, and are carried out according to mathematical algorithm
Automatic sub-clustering.The former is likely to occur the disunity in cluster on signal characteristic domain, and the latter may cause on locational space domain not
It is unified.Therefore the two is likely to introduce error when cluster head identifies, to reduce positioning accuracy.
Summary of the invention
For, since sub-clustering brings accuracy decline problem, the present invention provides one kind more present in existing fingerprint location technology
Add the method for efficient fingerprint graph search.
The technical solution of the present invention is as follows:
A kind of prediction type fingerprint image searching method in fingerprint location, the prediction type fingerprint image searching method first will
The location information predicted using filtering algorithm is as prior information, centered on prior information, by the side of ambient radiation
Formula scans for.K reference point RP nearest away from signal characteristic TP in signal space Euclidean distance is constantly updated in search process.
When continuous search L times and K RP does not change, then assert that this K RP is finally selected, the reference for location estimation
Point, and bring into KNN algorithm and resolved, replace traditional to fingerprint image progress sequential search method.
The present invention specifically includes the following steps:
Step 1: preset search upper limit L;When positioning for the first time, user is positioned using conventional fingerprint positioning method, is obtained
To first positioning result;It is non-when positioning for the first time, the positioning result obtained before is calculated by filtering algorithm, is obtained pre-
Measured value (X 'i, Y 'i);The conventional fingerprint positioning method is traditional sub-clustering calculation method or not sub-clustering calculation method.
Step 2: calculating result (first positioning result or the predicted value (X ' obtained with step 1i, Y 'i)) nearest reference
Point RP(a, b), a and b indicate the number for the neighboring reference point that positioning result is nearest in step 1.
Step 3: calculating RP(a, b)Signal space Euclidean distance between TP;The TP is user in position fixing process
The signal characteristic of itself measurement.
Step 4: with RP(a, b)Centered on, search and RP(a, b)Adjacent reference point RP calculates each reference point RP respectively
Signal space Euclidean distance between TP.
Step 5: the smallest K reference point in 4 signal space Euclidean distance of recording step enables search counter Nc=0.
Step 6: extending to the outside a circle search range, and calculate the reference point RP in the search range, and calculate respectively every
Signal space Euclidean distance between one reference point RP and TP;The search range is all references to search for before
Centered on point RP, the reference point RP adjacent with the center is searched for.
Step 7: the smallest K reference point in signal space Euclidean distance in recording step 6.
Step 8: whether two groups of K selected reference point RP of 7 front and back of observation of steps change: if changing, returning
Step 6, Nc=0;If not changing, Nc=Nc+ 1, and judge the size relation of Nc Yu preset search upper limit L, if NcIt is not less than
L then enters step 9, if NcLess than L, then return step 6.
Step 9: the position of K reference point in step 7 being brought into KNN algorithm and is positioned, this positioning knot is obtained
Fruit (Xi, Yi);All steps are repeated, are positioned next time.
The invention has the advantages that improving fingerprint location system compared to the conventional method positioned after sub-clustering
Positioning accuracy, and faster than the traditional algorithm search speed of not sub-clustering.
Detailed description of the invention
Fig. 1 is work flow diagram of the present invention.
Fig. 2 is the positioning accuracy simulation result diagram of simulating, verifying.
Specific embodiment
A specific embodiment of the invention is described in detail below in conjunction with technical solution (and attached drawing).
Input/output variable of the present invention is respectively as follows:
Input: fingerprint image reference point signal characteristic database;Fingerprint image reference point spatial position data library;The signal of TP is special
Sign;The k value of KNN algorithm;The X-coordinate of previous positioning result;The Y-coordinate of previous positioning result;Search for upper limit L.
Output: positioning result;The reference point number searched for.
In specific implementation the present invention the following steps are included:
Embodiment 1:
Step 1: preset search upper limit L;When positioning for the first time, user is positioned using conventional fingerprint positioning method, is obtained
To first positioning result;It is non-when positioning for the first time, the positioning result obtained before is calculated by filtering algorithm, is obtained pre-
Measured value (X 'i, Y 'i)。
Step 2: calculating result (first positioning result or the predicted value (X ' obtained with step 1i, Y 'i)) nearest reference
Point RP(a, b), a and b indicate the number for the neighboring reference point that positioning result is nearest in step 1.
Step 3: calculating RP(a, b)Signal space Euclidean distance between TP;The TP is user in position fixing process
The signal characteristic of itself measurement.
Step 4: with RP(a, b)Centered on, search and RP(a, b)Adjacent reference point RP calculates each reference point RP respectively
Signal space Euclidean distance between TP.
Step 5: the smallest 4 reference points in 4 signal space Euclidean distance of recording step enable search counter Nc=0.
Step 6: extending to the outside a circle search range, and calculate the reference point RP in the search range, and calculate respectively every
Signal space Euclidean distance between one reference point RP and TP;The search range is all references to search for before
Centered on point RP, the reference point RP adjacent with the center is searched for.
Step 7: the smallest 4 reference points in signal space Euclidean distance in recording step 6.
Step 8: whether two groups 4 selected reference point RP of observation change: if changing, return step 6, Nc=0;
If not changing, Nc=Nc+ 1, and judge the size relation of Nc Yu preset search upper limit L, if NcNot less than L, then enter step
9, if NcLess than L, then return step 6.
Step 9: the position of 4 reference points in step 7 being brought into KNN algorithm and is positioned, this positioning knot is obtained
Fruit (Xi, Yi);All steps are repeated, are positioned next time.
Embodiment 2:
It is 4 that preset search upper limit L, which is 5, K value, in the present embodiment, and other steps are same as Example 1.
Embodiment 3:
It is 4 that preset search upper limit L, which is 10, K value, in the present embodiment, and other steps are same as Example 1.
Fig. 2 is conventional fingerprint localization method and the positioning accuracy simulation result diagram obtained using localization method of the present invention, table 1
For the locating speed simulation result of simulating, verifying.From figure 2 it can be seen that the positioning accuracy after sub-clustering is minimum, variance is maximum;
And positioning accuracy of the invention is then suitable with the conventional method of not sub-clustering.When one-parameter selection is reasonable when L=5 () in such as figure,
Positioning accuracy of the invention is even higher than the conventional fingerprint location algorithm of not sub-clustering.Since the speed of fingerprint location is mainly by each
The quantity for the reference point for needing to search in positioning determines, from table 1 it follows that the searching for reference point quantity of cluster-dividing method is most
It is few, but combine the simulation result in Fig. 2 it is found that its positioning accuracy is minimum.Again it can be seen that reference point of the invention in table 1
Number of searches is lower than the conventional fingerprint positioning means of not sub-clustering, that is, locating speed is higher, and combines the analysis in Fig. 2, this hair
Bright positioning accuracy is higher.
The locating speed simulation result of 1 simulating, verifying of table
Claims (1)
1. the prediction type fingerprint image searching method in a kind of fingerprint location, it is characterised in that following steps:
Step 1: preset search upper limit L;When positioning for the first time, user is positioned using conventional fingerprint positioning method, obtains
One positioning result;It is non-when positioning for the first time, the positioning result obtained before is calculated by filtering algorithm, obtains predicted value
(X′i, Y 'i);
Step 2: calculating the nearest reference point RP of the result obtained with step 1(a, b), a and b indicate that positioning result is nearest in step 1
Neighboring reference point number;
Step 3: calculating RP(a, b)Signal space Euclidean distance between TP;The TP be user in position fixing process itself
The signal characteristic of measurement;
Step 4: with RP(a, b)Centered on, search and RP(a, b)Adjacent reference point RP, calculates each reference point RP and TP respectively
Between signal space Euclidean distance;
Step 5: the smallest K reference point in 4 signal space Euclidean distance of recording step enables search counter Nc=0;
Step 6: extending to the outside a circle search range, and calculate the reference point RP in the search range, and calculate each respectively
Signal space Euclidean distance between reference point RP and TP;The search range is all reference point RP to search for before
Centered on, search for the reference point RP adjacent with the center;
Step 7: the smallest K reference point in signal space Euclidean distance in recording step 6;
Step 8: whether two groups of K selected reference point RP of 7 front and back of observation of steps change: if changing, return step
6, Nc=0;If not changing, Nc=Nc+ 1, and judge the size relation of Nc Yu preset search upper limit L, if NcNot less than L, then
9 are entered step, if NcLess than L, then return step 6;
Step 9: the position of K reference point in step 7 being brought into KNN algorithm and is positioned, this positioning result (X is obtainedi,
Yi);All steps are repeated, are positioned next time.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611061169.3A CN106792510B (en) | 2016-11-25 | 2016-11-25 | A kind of prediction type fingerprint image searching method in fingerprint location |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611061169.3A CN106792510B (en) | 2016-11-25 | 2016-11-25 | A kind of prediction type fingerprint image searching method in fingerprint location |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106792510A CN106792510A (en) | 2017-05-31 |
CN106792510B true CN106792510B (en) | 2019-05-24 |
Family
ID=58913407
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611061169.3A Active CN106792510B (en) | 2016-11-25 | 2016-11-25 | A kind of prediction type fingerprint image searching method in fingerprint location |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106792510B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108650625A (en) * | 2018-05-17 | 2018-10-12 | 重庆邮电大学 | Distributing antenna system localization method based on fast run-up library |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103068035A (en) * | 2011-10-21 | 2013-04-24 | 中国移动通信集团公司 | Wireless network location method, device and system |
CN103442432A (en) * | 2013-08-09 | 2013-12-11 | 京信通信系统(中国)有限公司 | Fingerprint locating method and server |
CN105578414A (en) * | 2016-01-08 | 2016-05-11 | 南京网河智能科技有限公司 | Terminal and positioning method and apparatus for same |
CN105974361A (en) * | 2016-05-06 | 2016-09-28 | 南开大学 | Indoor positioning method based on fingerprint section indexes and WiFi-FM fusion fingerprints |
-
2016
- 2016-11-25 CN CN201611061169.3A patent/CN106792510B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103068035A (en) * | 2011-10-21 | 2013-04-24 | 中国移动通信集团公司 | Wireless network location method, device and system |
CN103442432A (en) * | 2013-08-09 | 2013-12-11 | 京信通信系统(中国)有限公司 | Fingerprint locating method and server |
CN105578414A (en) * | 2016-01-08 | 2016-05-11 | 南京网河智能科技有限公司 | Terminal and positioning method and apparatus for same |
CN105974361A (en) * | 2016-05-06 | 2016-09-28 | 南开大学 | Indoor positioning method based on fingerprint section indexes and WiFi-FM fusion fingerprints |
Also Published As
Publication number | Publication date |
---|---|
CN106792510A (en) | 2017-05-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107810522B (en) | Real-time, model-based object detection and pose estimation | |
CN112847343B (en) | Dynamic target tracking and positioning method, device, equipment and storage medium | |
CN105493078B (en) | Colored sketches picture search | |
CN105913028B (en) | Face + + platform-based face tracking method and device | |
CN108022262A (en) | A kind of point cloud registration method based on neighborhood of a point center of gravity vector characteristics | |
Aldoma et al. | Automation of “ground truth” annotation for multi-view RGB-D object instance recognition datasets | |
US20190066311A1 (en) | Object tracking | |
CN103488972A (en) | Method for detection fingertips based on depth information | |
US11972578B2 (en) | Method and system for object tracking using online training | |
Huang et al. | A coarse-to-fine algorithm for registration in 3D street-view cross-source point clouds | |
CN111611936B (en) | Automatic identification system for similar vector diagrams in CAD drawings | |
Ni et al. | An improved adaptive ORB-SLAM method for monocular vision robot under dynamic environments | |
CN109463003A (en) | Object identifying | |
CN111738319A (en) | Clustering result evaluation method and device based on large-scale samples | |
CN109699003B (en) | Position determination method and device | |
CN106792510B (en) | A kind of prediction type fingerprint image searching method in fingerprint location | |
Park et al. | Extended keyframe detection with stable tracking for multiple 3d object tracking | |
US20190087684A1 (en) | Fast joint template matching | |
He et al. | Variable scale learning for visual object tracking | |
CN106792509B (en) | A kind of diffusion type fingerprint image searching method in fingerprint location | |
KR20140079639A (en) | Method for selecting similar users for collaborative filtering based on earth mover´s distance | |
CN113592906B (en) | Long video target tracking method and system based on annotation frame feature fusion | |
CN106597367B (en) | The fingerprint image searching method of steepest descending manner in a kind of fingerprint location | |
Ma et al. | Depth-guided progressive network for object detection | |
CN112633078A (en) | Target tracking self-correcting method, system, medium, equipment, terminal and application |
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 |