CN109068267A - A kind of indoor orientation method based on LoRa SX1280 - Google Patents
A kind of indoor orientation method based on LoRa SX1280 Download PDFInfo
- Publication number
- CN109068267A CN109068267A CN201810876943.9A CN201810876943A CN109068267A CN 109068267 A CN109068267 A CN 109068267A CN 201810876943 A CN201810876943 A CN 201810876943A CN 109068267 A CN109068267 A CN 109068267A
- Authority
- CN
- China
- Prior art keywords
- fingerprint
- lora
- coordinate
- special
- value
- 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
- QVFWZNCVPCJQOP-UHFFFAOYSA-N chloralodol Chemical compound CC(O)(C)CC(C)OC(O)C(Cl)(Cl)Cl QVFWZNCVPCJQOP-UHFFFAOYSA-N 0.000 title claims abstract description 58
- 238000000034 method Methods 0.000 title claims abstract description 57
- 238000001914 filtration Methods 0.000 claims abstract description 12
- 238000004364 calculation method Methods 0.000 claims abstract description 4
- 238000012549 training Methods 0.000 claims description 19
- 238000004891 communication Methods 0.000 claims description 15
- 239000011159 matrix material Substances 0.000 claims description 4
- 238000010276 construction Methods 0.000 claims description 2
- 238000013507 mapping Methods 0.000 claims description 2
- 238000005457 optimization Methods 0.000 claims description 2
- 238000012706 support-vector machine Methods 0.000 claims description 2
- 230000004807 localization Effects 0.000 description 9
- 238000005259 measurement Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 230000005540 biological transmission Effects 0.000 description 3
- XCWPUUGSGHNIDZ-UHFFFAOYSA-N Oxypertine Chemical compound C1=2C=C(OC)C(OC)=CC=2NC(C)=C1CCN(CC1)CCN1C1=CC=CC=C1 XCWPUUGSGHNIDZ-UHFFFAOYSA-N 0.000 description 2
- HUTDUHSNJYTCAR-UHFFFAOYSA-N ancymidol Chemical compound C1=CC(OC)=CC=C1C(O)(C=1C=NC=NC=1)C1CC1 HUTDUHSNJYTCAR-UHFFFAOYSA-N 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 240000005373 Panax quinquefolius Species 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
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
-
- 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
- H04W4/30—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
-
- 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
Abstract
The invention discloses a kind of indoor orientation methods based on LoRa SX1280.It includes offline fingerprint collecting modelling phase and online real-time positioning stage.The offline fingerprint collecting modelling phase: being to need the special fingerprint characteristic based on LoRa SX1280 first, and special fingerprint characteristic includes ranging fingerprint and RSSI fingerprint；Then it when special fingerprint is entered into fingerprint base, needs to pre-process special fingerprint using gaussian filtering method, finally obtains location Calculation model using improvement support vector regression method.Online real-time positioning stage: first with mobile node device and anchor point, region is acquired t times indoors, forms the special fingerprint characteristic of t group；Secondly median filtering method is used to this special fingerprint characteristic of t group, obtains a unique fingerprint feature；Finally this unique fingerprint feature is updated in obtained computation model, the position coordinates of node to be positioned are calculated.The present invention eliminates fingerprint vulnerable to interference problem.
Description
Technical field
The present invention relates to internet of things field more particularly to a kind of indoor orientation methods based on LoRa SX1280.
Background technique
LoRa technology is that Semtech company uses and a kind of overlength distance based on spread spectrum promoted wirelessly passes
Transmission scheme belongs to one of low-power consumption wide area network (Low Power Wide Area Network, LPWAN) communication technology.
LoRa technology is no longer limited by transmission range and the compromise of power consumption considers that providing one kind for users can be realized transmission range
Far, low in energy consumption, multinode system, to be expanded into network.
As the audience of LoRa technology is more and more and the sustainable development of itself.And then Semtech company pushes away
SX1280 chip is gone out, it is meant that the LoRa's of 2.4GHz frequency range emerges.The built-in chip type ranging engine function, that is, be based on
Arrival time (Time Of Arrival, TOA) ranging engine of time-of-flight (TOF).The distance measurement function base of SX1280
The measurement of the turnaround time of wireless signal between SX1280 transceiver.The process uses LoRa modulation scheme, therefore
Benefit from all advantages for the remote and low power consumption operation that LoRa is authorized.
Continuous with Internet of Things application expands and gos deep into, and location based service, which has, to be more and more widely used.Room
Interior positioning has some particularity because of special environmental factor, such as lacks unified basic framework, indoor positioning technologies
It is difficult to agreeing property, universality, indoor environment variation also has serious interference effect.Currently, indoor orientation method is main
Including the localization method based on ranging and based on the localization method of fingerprint.Based on the localization method of ranging according to mode ranging
There is distinct methods, such as TOA, reaching time-difference (TDOA), received signal strength (RSSI) etc. again.Inside these methods, base
It is widely used in RSSI ranging because its hardware cost and computing cost are relatively low, but RSSI range accuracy is poor, causes
Its position error is also larger.Although SX1280 uses TOA ranging, show the ranging engine of SX1280 by actual test
Range error indoors is also bigger.Therefore, the positioning that location Calculation obtains is carried out only in accordance with the TOA distance measurement value of SX1280
Error is also bigger.Localization method based on fingerprint mainly utilizes the otherness of wireless signal spatial position, passes through location fingerprint
Relational database and matching algorithm search the corresponding physical location information of location fingerprint of destination node, to estimate node location.
Common matching algorithm has arest neighbors (NN) algorithm, neural network algorithm, vector machine algorithm, K arest neighbors (KNN) algorithm etc..It is existing
There is fingerprint positioning method mainly to use RSSI fingerprint or magnetic field fingerprint etc..In practical indoor environment, since multipath, shade are imitated
Answer, personnel walk about etc. influences, complicated time-varying statistical property is often shown from the RSSI value of each AP point on fixed position,
This can reduce system accuracy and increase system complexity.
Summary of the invention
The invention discloses a kind of indoor heating system and localization method based on LoRa SX1280.
It includes: mobile node device, anchor point, net that the present invention, which solves indoor locating system used by its technical problem,
It closes, location-server；Wherein mobile node device and all integrated LoRa SX1280 mould group of anchor point.
The LoRa SX1280 mould group supports LoRa communication pattern and LoRa ranging engine-model.
The mobile node device is configured on all movable equipments, and position is not fixed.Mobile node device
Using SX1280 mould group, stand-alone antenna, vibrating sensor, battery power supply.
The anchor point refers to the constant device in the position for being placed on known coordinate, is mainly used for assisting mobile node
The collection of device completion finger print information.Anchor point uses SX1280 mould group, stand-alone antenna, mains-supplied.
The gateway is mainly used for mobile node device and location-server communication.
The location-server is used for after the Location Request and relevant information for receiving mobile node device, is positioned
It calculates.
The technical solution adopted by the present invention to solve the technical problems includes: offline fingerprint collecting modelling phase and online reality
When positioning stage.
One, offline fingerprint collecting modelling phase, specific steps include:
Area to be targeted is carried out grid dividing by 1-1., and using mobile node device and anchor point to each mesh point
Region acquires the special fingerprint characteristic based on LoRa SX1280；
The special fingerprint characteristic is included the ranging fingerprint obtained based on LoRa SX1280 ranging engine-model and is based on
Received signal strength (RSSI) fingerprint that LoRa SX1280 communication pattern obtains；Ranging fingerprint and RSSI fingerprint are combined, increased
Add the dimension of characteristic fingerprint.
Further, the center position coordinates for remembering i-th of mesh point region are Pi=(xi,yi), remember i-th of mesh point area
The collected characteristic fingerprint matrix in domain is
ForIt is come from j-th when l=2j-1 (odd column) indicates that kth time is collected at i-th of mesh point region
The reception signal intensity RSSI fingerprint value that anchor point is obtained based on communication pattern, when l=2j (even column) is indicated i-th
The collected ranging fingerprint value obtained from j-th of anchor point based on ranging engine-model of kth time at a mesh point region,
Wherein i=1,2 ... N, j=1,2 ... M, l=1,2 ... 2M, k=1,2 ... T.RiColumns be characterized the dimension 2M of fingerprint vector.
In above, N is the mesh point number of area to be targeted, and M is anchor point number, and T is the times of collection in each mesh point.
1-2. is directed to the collected special fingerprint characteristic of each mesh point, using gaussian filtering method by Outliers mean value
Substitution, is then stored in finger print data for filtered special fingerprint characteristic and the corresponding two-dimensional coordinate position of the mesh point together
Library.
Then 1-3. is constructed using the fingerprint in fingerprint database as training set using improved support vector regression algorithm
Computation model between special fingerprint characteristic two-dimensional coordinate corresponding with the mesh point.
Two, online real-time positioning stage, specific steps include:
2-1. in t special fingerprint of its station acquisition, forms the special finger of t group using mobile node device (node to be positioned)
Line feature.
2-2. uses median filtering method to this special fingerprint characteristic of t group, obtains a unique fingerprint feature.
This unique fingerprint feature is updated in the obtained computation model of step 1-3 by 2-3., is calculated to be positioned
The position coordinates of node.
Further, meter described in the construction method and step 2-3 of support vector regression computation model is improved described in step 1-3
Calculating node location to be positioned and sitting calibration method all is the improvement support vector regression ISVR algorithm proposed using document [1].
The present invention has the beneficial effect that:
The present invention provides a kind of indoor locating system and method based on LoRa SX1280 chip, is substantially a kind of base
In the indoor locating system and method for fingerprint, it can solve and carry out high-precision indoor locating system using LoRa SX1280 chip
Design problem.With conventional fingerprint localization method the difference is that: fingerprint uses the fingerprint that merges of distance measurement value and RSSI value, eliminates
A kind of fingerprint is vulnerable to interference problem；All fingerprint is pre-processed in off-line phase and on-line stage, improves the finger of acquisition
The consistency of line and position；The ISVR proposed early period using inventor carries out location Calculation, improves fingerprint characteristic and position
The precision of computation model.Due to the improvement in terms of these three, so that the positioning of indoor locating system and method that the present invention announces
The positioning accuracy of ratio of precision conventional fingerprint positioning system and method wants high.
[1] Yao Yingbiao, Mao Weiyong, Yao Ruili, Yan Junrong, Feng Wei are calculated based on the indoor positioning for improving support vector regression
Method [J], Chinese journal of scientific instrument, 2017,38 (9): 84-91.
Detailed description of the invention
Fig. 1 (a) is indoor locating system device figure of the present invention；
Fig. 1 (b) is indoor locating system device figure of the present invention；
Fig. 2 is the special fingerprint characteristic of the present invention；
Fig. 3 is the schematic diagram proposed by the present invention that RSSI fingerprint is obtained using LoRa communication pattern；
Fig. 4 is the schematic diagram proposed by the present invention that ranging fingerprint is obtained using LoRa ranging engine-model；
Fig. 5 is the localization method block diagram of support vector regression of the invention；
Fig. 6 is one of specific embodiment of the invention indoor positioning environment schematic.
Specific embodiment
With reference to the accompanying drawings and examples, specific embodiments of the present invention will be described in further detail.Implement below
Example is not intended to limit the scope of the invention for illustrating the present invention.
Fig. 1 (a) and (b) are a kind of indoor locating system device figure based on LoRa SX1280 proposed by the present invention.Include
Removable node apparatus Fig. 1 (a) and fixed two kinds of devices of anchor point Fig. 1 (b).Removable node apparatus and fixed anchor point
Require to install the LoRa mould group containing SX1280 chip on device.The SX1280 chip of Sheng Te company of the U.S. possesses LoRa communication
Mode and LoRa ranging engine-model.
Fig. 2 is a kind of special fingerprint characteristic based on LoRa SX1280 proposed by the present invention.This fingerprint characteristic includes
Ranging fingerprint and RSSI fingerprint；As shown in Fig. 2, by received signal strength RSSI fingerprint value and being based on LoRa ranging engine-model
Obtained ranging fingerprint value combines, and increases the dimension of feature vector, improves positioning accuracy.
Fig. 3 is the schematic diagram proposed by the present invention that RSSI fingerprint is obtained using LoRa communication pattern.Need to use Fig. 1's
Acquisition device installs the LoRa mould group containing SX1280 chip on removable node apparatus and fixed anchor point device.
SX1280 chip itself possesses LoRa communication pattern and LoRa ranging engine-model.In a communication mode, removable node is utilized
Fixed anchor point (AP access point) around devices broadcasting paging, receives the RSSI value that each anchor point is read after anchor point is replied.
Fig. 4 is the schematic diagram proposed by the present invention that ranging fingerprint is obtained using LoRa ranging engine-model.It is obtained in previous step
After taking RSSI fingerprint, removable node apparatus enters ranging engine-model host mode, and each anchor point enters ranging engine
Mode slave mode, removable node apparatus and each anchor point carry out ranging one by one, and read the survey between each anchor point
Away from value.
Fig. 5 is a kind of localization method block diagram of the support vector regression based on LoRa SX1280 proposed by the present invention；
It include: the offline acquisition phase of S1；Area to be targeted is subjected to grid dividing, rationally the fixed anchor point of deployment.
And the above-mentioned special fingerprint characteristic comprising distance measurement value and RSSI value is acquired using removable node apparatus and anchor point.It is sharp again
The data filtering of initial fingerprint feature is carried out with gaussian filtering method, i.e., the collected fingerprint characteristic of each reference point locations is used
Gaussian filtering method substitutes Outliers with mean value.It then will be by the RSSI fingerprint and ranging fingerprint phase after gaussian filtering method
The feature samples of fusion and respective coordinates position are stored in fingerprint database.And it is constructed using support vector regression algorithm is improved
Off-line training model between position coordinates and reference point fingerprint characteristic.S2, online real-time positioning stage；Utilize removable node
Device acquires radio signal characteristics t times in real time in certain bit positions undetermined, and the data preprocessing method filtered using median,
The method for using each dimension to take median t fingerprint characteristic at this same position, only selection obtains only one fingerprint
Feature.This fingerprint characteristic is updated in the obtained off-line training model of S1, so that the position of node to be positioned be calculated
Set coordinate.
The localization method specific implementation step of support vector regression based on LoRa SX1280 is as follows:
Step 1:
Carry out grid dividing according to area in area to be targeted, and establish rectangular coordinate system, choose N number of reference point region and
M LoRa access point, each LoRa access point is fixed, and installs anchor point.
In the embodiment shown in fig. 6, region is carried out grid dividing with dotted line, often by the indoor environment of a 16m*12m
One grid is fingerprint reference point region, and black × number expression reference point region center, which is the reference point
Coordinate；Black triangle is selected LoRa access point, needs to install fixed anchor point, here N=48, M=4.
Step 2:
T fingerprint vector signal spy evenly distributedly is acquired with removable node apparatus at each reference point regional location
Sign is used as 1 group of data.The fingerprint vector signal characteristic includes removable node acquisition between surrounding fixed anchor point device
RSSI value fingerprint and ranging fingerprint based on LoRa ranging engine-model.Specifically, we will utilize LoRa ranging engine-model,
Therefore the SX1280 chip released using Liao Shengte company, the built-in chip type ranging engine function, i.e. arrival time (Time Of
Arrival, TOA) fusion ranging engine, use time-of-flight (TOF) distance measuring method.The distance measurement function base of SX1280
The measurement of turnaround time between a pair of of SX1280 transceiver (node to be positioned and fixed anchor point i.e. herein).Hold
Intelligibly, the feature vector for combining RSSI value and distance measurement value is trained in the location algorithm based on support vector regression, it can
To increase the dimension of feature vector, the accuracy of positioning is improved.
The center position coordinates for remembering i-th of mesh point region are Pi=(xi,yi), i-th of mesh point region of note collects
Characteristic fingerprint matrix be
ForIt is come from j-th when l=2j-1 (odd column) indicates that kth time is collected at i-th of mesh point region
The reception signal intensity RSSI fingerprint value that anchor point is obtained based on communication pattern, when l=2j (even column) is indicated i-th
The collected ranging fingerprint value obtained from j-th of anchor point based on ranging engine-model of kth time at a mesh point region,
Wherein i=1,2 ... N, j=1,2 ... M, l=1,2 ... 2M, k=1,2 ... T.RiColumns be characterized the dimension 2M of fingerprint vector.
In above, N is the mesh point number of area to be targeted, and M is anchor point number, and T is the times of collection in each mesh point.Actually answer
In, to obtain preferable locating effect, T is generally greater than 30.
Step 3:
It will evenly distributedly collected fingerprint vector signal characteristic carries out initial signal and locates in advance in each reference point region
Reason.Initial signal pretreatment uses gaussian filtering method, specifically according to following principle: to the fingerprint of formula (1) each reference point acquisition
Matrix RiFind out the average value mu of its each dimension datal(mean values of each column), wherein l=1,2 ..., 2M, indicate dimension.One is arranged again
A threshold probability λ, it indicates that signal value is distributed in section [μl-θl,μl+θl] probability, wherein θlCollected instruction can be passed through
Practice collection data and λ is acquired.Specifically, it is assumed that the equal Normal Distribution of each column data in collected training set data, note
Its probability density function is fl(x), then θlMeet
θ is found out according to formula (2)lAfterwards, R is obtainediIn every column element zone of reasonableness be [μl-θl,μl+θl].In this range
Interior signal value is considered reliable；, whereas if signal value is not in this range, then it is assumed that the signal value is unreliable.Finally,
For RiEach of elementIfRetainOtherwise,λ can basis
Actual environment is selected, such as can take λ=0.8.
Step 4:
Utilize the fingerprint matrices R of each reference point after step 3 pretreatmentiEvery a line and position coordinates Pi=
(Xi, Yi) establish fingerprint database D, Ri∈R2M, and the training sample set as SVR algorithm, then total sample number should be N*T.
Step 5:
Utilize the non-linear relation between the fingerprint vector and coordinate of SVR algorithm building reference point.It is non-by one first
Linear Mapping Φ: R2M→ F is by input space R2MBe mapped in the feature space F of higher-dimension, then in F construct position coordinates P with
The optimum regression function of fingerprint R:
P=WT·Φ(R)+b。 (3)
In formula: W is weight coefficient, and W ∈ F, b are biasing coefficient
In order to finally determine the parameter W and b in formula (5), following convex two can be solved according to structure risk minimum principle
Secondary planning problem.
Meet:
In formula: C is penalty constant, ξi,For slack variable.
The lagrange polynomial of the convex quadratic programming problem of formula (5) are as follows:
In formula:Lagrange multiplier, and need to meet
The condition according to existing for optimal solution enables L to W, b,Ask local derviation that can obtain for 0:
By solution primal-dual optimization problem, following formula can be obtained:
Meet:
0≤αi *,αi≤C
Wherein κ (Ri,Rj) it is kernel function, kernel function is typically chosen gaussian kernel function
Wherein xi,xjIndicate input vector, σ is Gaussian kernel width.
Thus, it is possible to obtain
Here optimal penalty constant C and kernel function width value σ can use support vector machines tool box LIBSVM, carry out
Grid search obtains.
Based on the positioning of SVR algorithm, here only consider indoor plane positioning, need to export user two-dimensional coordinate (x,
Y), that is, need to use the prediction of multi output.And the output of tradition SVR algorithm be all it is one-dimensional, at this time can be by multiple
Single output replaces multi output, to realize multi output SVR algorithm.But indoor coordinate system is two associated information, institute
The accuracy of training pattern can be reduced to a certain extent with the two-dimensional coordinate vector for individually training each coordinate to obtain.
In order to reduce the error of individually building x and y coordinates model, the pass between two-dimensional position information and signal strength is improved
Connection property can increase by one calibration coordinate z=x*y of training in the training stage, and z includes the information of x and y simultaneously.
Step 6:
It acquires the radio signal characteristics of node to be positioned in real time using removable node apparatus, and passes through data prediction.
Data prediction uses median filter method, specifically, in a certain bit positions undetermined t finger of removable node continuous acquisition
Line feature vector obtains t item record.This t item is denoted as
Wherein forWhen l=2j-1 (odd column) indicates the t times collected letter from j-th of anchor point
Number intensity RSSI fingerprint value, when l=2j (even column) indicates collected to come from j-th of anchor point and draw based on ranging for the t times
Hold up the ranging fingerprint value of mode acquisition.Wherein l=1,2...2M, j=1,2...M.The median for remembering each column is R "l, then Z
Median vector should be [R "1, R "2... R "l…R″2M-1, R "2M]。
Specifically, it is assumed that t item record is as follows, t=6:
Then the median vector of the Z is [- 55,8.35, -61,5.25, -65,4.58, -71,10.5].
Step 7:
The off-line training mould built in step 5 will be updated to by the median vector of the Z after data prediction
Type.
Due to constructing 3 training patterns respectively for x, y, z in the training stage.Therefore 3 are had in the prediction result obtained
Group coordinate value, is (x ', y '), (x ', z '/x '), (z '/y ', y ') respectively.After the 3 groups of coordinate values obtained by calibration coordinate, such as
The coordinate value that fruit obtains is identical, then can directly export result；, whereas if obtained coordinate value is different, then need this 3
One group of coordinate output is selected in group coordinate.
For this problem, selected using the method for weighting anti-k nearest neighbor (WIKNN): according to 3 groups of obtained prediction bits
Combinatorial coordinates are set, by calculating the Euclidean distance in fingerprint base between the coordinate and predicted position coordinate of each position, select Europe
Formula is apart from k nearest position coordinates, after being then multiplied by a weighting coefficient to the corresponding fingerprint characteristic of selected k coordinate
It sums again, obtains fingerprint characteristic vector corresponding to predicted position.
Specifically,
Step 7.1, it is assumed that the coordinate of selection is (x ', y '), calculates (x ', y ') and fingerprint database first with formula (14)
In coordinate (Xi, Yi) the distance between di, then according to diK reference point before sequential selection from small to large, and calculate the ginseng
The mean vector R of fingerprint characteristic vector of the examination point in fingerprint databasei, finally using formula (15) to this k feature vector into
Row weighting, obtains the corresponding fingerprint characteristic vector R of coordinate (x ', y ').
In formula:
Step 7.2, according to the fingerprint characteristic vector R ' collected at (x ', y '), it is special that fingerprint is calculated using formula (16)
Levy the Euclidean distance D (1) between vector R and R '.
Step 7.3, step 7.1 and 7.2 is repeated to (x ', z '/x '), (z '/y ', y ') then respectively, respectively obtain it is European away from
From D (2) and D (3).
Step 7.4, by Euclidean distance D (1) obtained above, D (2), D (3) sequence select the smallest D (i) of Euclidean distance
Corresponding coordinate, and using this coordinate as final position coordinates.To estimate the actual location of user.
Finally, being enlightenment with above-mentioned specific embodiment, it is not intended to limit the scope of the present invention.It is all of the invention
Within spirit and principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.
Claims (10)
1. a kind of indoor orientation method based on LoRa SX1280, it is characterised in that including the offline fingerprint collecting modelling phase and
Online real-time positioning stage, is implemented as follows:
One, offline fingerprint collecting modelling phase, specific steps include:
Area to be targeted is carried out grid dividing by 1-1., and using mobile node device and anchor point to each mesh point region
Acquire the special fingerprint characteristic based on LoRa SX1280；
The special fingerprint characteristic includes the ranging fingerprint obtained based on LoRa SX1280 ranging engine-model and based on LoRa
Received signal strength (RSSI) fingerprint that SX1280 communication pattern obtains；Ranging fingerprint and RSSI fingerprint are combined, increased special
Levy the dimension of fingerprint；
Further, the center position coordinates for remembering i-th of mesh point region are Pi=(xi, yi), i-th of mesh point region of note is adopted
The characteristic fingerprint matrix collected is
ForJ-th of anchor point is come from when l=2j-1 (odd column) indicates that kth time is collected at i-th of mesh point region
The reception signal intensity RSSI fingerprint value that device is obtained based on communication pattern, when l=2j (even column) is indicated in i-th of net
The collected ranging fingerprint value obtained from j-th of anchor point based on ranging engine-model of kth time at lattice point region, wherein
I=1,2...N, j=1,2...M, l=1,2...2M, k=1,2...T；RiColumns be characterized the dimension 2M of fingerprint vector；
In above, N is the mesh point number of area to be targeted, and M is anchor point number, and T is the times of collection in each mesh point；
1-2. is directed to the collected special fingerprint characteristic of each mesh point, is replaced Outliers with mean value using gaussian filtering method
Then filtered special fingerprint characteristic and the corresponding two-dimensional coordinate position of the mesh point are stored in fingerprint database by generation together；
For 1-3. using the fingerprint in fingerprint database as training set, it is special then to be constructed using improved support vector regression algorithm
Computation model between fingerprint characteristic two-dimensional coordinate corresponding with the mesh point；
Two, online real-time positioning stage, specific steps include:
2-1., in t special fingerprint of its station acquisition, forms the special fingerprint characteristic of t group using mobile node device；
2-2. uses median filtering method to this special fingerprint characteristic of t group, obtains a unique fingerprint feature；
This unique fingerprint feature is updated in the obtained computation model of step 1-3 by 2-3., and node to be positioned is calculated
Position coordinates.
2. a kind of indoor orientation method based on LoRa SX1280 according to claim 1, it is characterised in that step 1-3
Node location coordinate to be positioned is calculated described in the construction method and step 2-3 for improving support vector regression computation model
Method is all using improvement support vector regression ISVR algorithm.
3. a kind of indoor orientation method based on LoRa SX1280 according to claim 2, it is characterised in that this method is real
The equipment now needed includes mobile node device, anchor point, gateway, location-server；Wherein mobile node device and anchor point
The all integrated LoRa SX1280 mould group of device, and the LoRa SX1280 mould group supports LoRa communication pattern and LoRa ranging to draw
Hold up mode.
4. a kind of indoor orientation method based on LoRa SX1280 according to claim 3, it is characterised in that the shifting
Dynamic node apparatus is configured on all movable equipments；Mobile node device is passed using SX1280 mould group, stand-alone antenna, vibration
Sensor, battery power supply.
5. a kind of indoor orientation method based on LoRa SX1280 according to claim 4, it is characterised in that the anchor
Point device refers to the device fixedly secured in appointed place, and the communication for being responsible between mobile node device, auxiliary is completed
The location related information of mobile node device is collected；Anchor point uses SX1280 mould group, stand-alone antenna, mains-supplied.
6. a kind of indoor orientation method based on LoRa SX1280 according to claim 5, it is characterised in that the net
It closes and is communicated for mobile node device and location-server；Location-server in the positioning for receiving mobile node device for asking
After asking, location Calculation is carried out.
7. a kind of indoor orientation method based on LoRa SX1280 according to claim 6, it is characterised in that use Gauss
Filter method substitutes Outliers with mean value, is implemented as follows:
To the fingerprint matrices R of formula (1) each reference point acquisitioniFind out the average value mu of its each dimension datal, wherein l=1,2...,
2M indicates dimension；One threshold probability λ is set, and threshold probability λ indicates that signal value is distributed in section [μl-θl, μl+θl] it is general
Rate, wherein θlIt is acquired by collected training set data and λ；
If the equal Normal Distribution of each column data in collected training set data, remember that its probability density function is fl(x),
Then θlMeet
θ is found out according to formula (2)lAfterwards, R is obtainediIn every column element zone of reasonableness be [μl-θl, μl+θl]；Within this range
Signal value is considered reliable；, whereas if signal value is not in this range, then it is assumed that the signal value is unreliable；
For RiEach of elementIfRetainOtherwise,λ
It is selected according to actual environment.
8. a kind of indoor orientation method based on LoRa SX1280 according to claim 7, it is characterised in that described adopts
With support vector regression ISVR algorithm is improved, it is implemented as follows:
Construct the non-linear relation between the fingerprint vector and coordinate of reference point:
Pass through Nonlinear Mapping Φ: R first2M→ F is by input space R2MIt is mapped in the feature space F of higher-dimension, then in F
The optimum regression function of middle building position coordinates P and fingerprint R:
P=WT.Φ(R)+b； (3)
In formula: W is weight coefficient, and W ∈ F, b are biasing coefficient
In order to which the parameter W and b that finally determine in formula (5) solve following convex quadratic programming according to structure risk minimum principle
Problem；
Meet:
In formula: C is penalty constant, ξi, ξi *For slack variable；
The lagrange polynomial of the convex quadratic programming problem of formula (5) are as follows:
In formula: ηi,αi,Lagrange multiplier, and need to meet
The condition according to existing for optimal solution enables L to W, b, ξi *Ask local derviation that can obtain for 0:
By solution primal-dual optimization problem, following formula can be obtained:
Meet:
Wherein, κ (Ri, Rj) it is kernel function, and Selection of kernel function gaussian kernel function κ (xi, xj)=exp (- | | xi-xj||2/2σ2)
(10)
Wherein xi, xjIndicate input vector, σ is Gaussian kernel width；
It obtains as a result,
Here optimal penalty constant C and kernel function width value σ utilizes support vector machines tool box LIBSVM, carries out grid search
It obtains；
In order to reduce the error of individually building x and y coordinates model, being associated between two-dimensional position information and signal strength is improved
Property, increase by one calibration coordinate z=x*y of training in the training stage, z includes the information of x and y simultaneously.
9. a kind of indoor orientation method based on LoRa SX1280 according to claim 8, it is characterised in that the use
Median filtering method obtains a unique fingerprint feature, is implemented as follows:
In the removable t fingerprint characteristic vector of node continuous acquisition of a certain bit positions undetermined, t item record Z is obtained:
Wherein forWhen l=2j-1 (odd column) indicates the t times collected signal strength from j-th of anchor point
RSSI fingerprint value, when l=2j (even column) indicates collected to come from j-th of anchor point based on ranging engine-model the t time
The ranging fingerprint value of acquisition；Wherein l=1,2...2M, j=1,2...M；The median for remembering each column is R "l, then the middle position of Z
Number vector should be [R "1, R "2... R "l…R″2M-1, R "2M]。
10. a kind of indoor orientation method based on LoRa SX1280 according to claim 9, it is characterised in that the step
Rapid 2-3 is implemented as follows:
Due to constructing 3 training patterns respectively for x, y, z in the training stage；Therefore 3 groups of seats are had in the prediction result obtained
Scale value is (x ', y '), (x ', z '/x '), (z '/y ', y ') respectively；The 3 groups of coordinate values obtained by calibration coordinate, if obtained
Coordinate value it is identical, then can directly export result；, whereas if obtained coordinate value is different, then need in this 3 groups of coordinates
Using the one group of coordinate output of method choice for weighting anti-k nearest neighbor (WIKNN)；
Specific step is as follows:
1. assuming that the coordinate of selection is (x ', y '), the coordinate in (x ', y ') and fingerprint database is calculated first with formula (14)
(Xi, Yi) the distance between di, then according to diK reference point before sequential selection from small to large, and calculate the reference point and referring to
The mean vector R of fingerprint characteristic vector in line databasei, finally this k feature vector is weighted using formula (15), is obtained
To the corresponding fingerprint characteristic vector R of coordinate (x ', y ').
In formula:
2. calculating fingerprint characteristic vector R and R ' using formula (16) according to the fingerprint characteristic vector R ' collected at (x ', y ')
Between Euclidean distance D (1)；
1. and 2. 3. repeating previous step to (x ', z '/x '), (z '/y ', y ') respectively, Euclidean distance D (2) and D is respectively obtained
(3)；
4. by Euclidean distance D (1) obtained above, D (2), D (3) sequence select seat corresponding to the smallest D (i) of Euclidean distance
Mark, and using this coordinate as final position coordinates.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810876943.9A CN109068267B (en) | 2018-08-03 | 2018-08-03 | Indoor positioning method based on LoRa SX1280 |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810876943.9A CN109068267B (en) | 2018-08-03 | 2018-08-03 | Indoor positioning method based on LoRa SX1280 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109068267A true CN109068267A (en) | 2018-12-21 |
| CN109068267B CN109068267B (en) | 2020-06-23 |
Family
ID=64833168
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810876943.9A Active CN109068267B (en) | 2018-08-03 | 2018-08-03 | Indoor positioning method based on LoRa SX1280 |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109068267B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109581286A (en) * | 2018-12-28 | 2019-04-05 | 皖西学院 | A kind of two dimension fingerprint distribution indoor positioning method and structure |
| CN110826524A (en) * | 2019-11-18 | 2020-02-21 | 广东博智林机器人有限公司 | Fingerprint library generation method and device, electronic equipment and storage medium |
| CN110996387A (en) * | 2019-12-02 | 2020-04-10 | 重庆邮电大学 | LoRa positioning method based on TOF and position fingerprint fusion |
| CN111065046A (en) * | 2019-11-21 | 2020-04-24 | 东南大学 | LoRa-based outdoor unmanned aerial vehicle positioning method and system |
| US10761200B1 (en) | 2019-02-27 | 2020-09-01 | Osram Gmbh | Method for evaluating positioning parameters and system |
| WO2020177539A1 (en) * | 2019-03-01 | 2020-09-10 | 阿里巴巴集团控股有限公司 | Positioning, seeking and tracking methods and apparatuses, and system |
| CN113051976A (en) * | 2019-12-27 | 2021-06-29 | 广东博智林机器人有限公司 | Fingerprint positioning method and device, electronic equipment and storage medium |
| CN117872269A (en) * | 2024-03-13 | 2024-04-12 | 电子科技大学 | High-precision positioning method for self-adaptive data processing |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102625447A (en) * | 2012-03-14 | 2012-08-01 | 东南大学 | Heuristic node localization method in wireless sensor network |
| CN105530702A (en) * | 2016-01-25 | 2016-04-27 | 杭州电子科技大学 | Wireless sensing network mobile node positioning method based on self-organizing mapping |
| CN107124698A (en) * | 2017-04-28 | 2017-09-01 | 深圳市亿兆互联技术有限公司 | A kind of vertical space positioner based on LoRa technologies, method and system |
| CN107333238A (en) * | 2017-07-03 | 2017-11-07 | 杭州电子科技大学 | A kind of indoor fingerprint method for rapidly positioning based on support vector regression |
| KR20180064682A (en) * | 2016-12-06 | 2018-06-15 | 에이티온 주식회사 | SYSTEM MEASURING POSITION WITH SMALL POWER USING LoRa NETWORK AND METHOD THEREOF |
| CN108318860A (en) * | 2017-12-29 | 2018-07-24 | 谷米科技有限公司 | LORA method of locating terminal, system and LORA Internet of Things management platforms |
-
2018
- 2018-08-03 CN CN201810876943.9A patent/CN109068267B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102625447A (en) * | 2012-03-14 | 2012-08-01 | 东南大学 | Heuristic node localization method in wireless sensor network |
| CN105530702A (en) * | 2016-01-25 | 2016-04-27 | 杭州电子科技大学 | Wireless sensing network mobile node positioning method based on self-organizing mapping |
| KR20180064682A (en) * | 2016-12-06 | 2018-06-15 | 에이티온 주식회사 | SYSTEM MEASURING POSITION WITH SMALL POWER USING LoRa NETWORK AND METHOD THEREOF |
| CN107124698A (en) * | 2017-04-28 | 2017-09-01 | 深圳市亿兆互联技术有限公司 | A kind of vertical space positioner based on LoRa technologies, method and system |
| CN107333238A (en) * | 2017-07-03 | 2017-11-07 | 杭州电子科技大学 | A kind of indoor fingerprint method for rapidly positioning based on support vector regression |
| CN108318860A (en) * | 2017-12-29 | 2018-07-24 | 谷米科技有限公司 | LORA method of locating terminal, system and LORA Internet of Things management platforms |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109581286A (en) * | 2018-12-28 | 2019-04-05 | 皖西学院 | A kind of two dimension fingerprint distribution indoor positioning method and structure |
| CN109581286B (en) * | 2018-12-28 | 2020-12-04 | 皖西学院 | Two-dimensional fingerprint distribution indoor positioning method and structure |
| US10761200B1 (en) | 2019-02-27 | 2020-09-01 | Osram Gmbh | Method for evaluating positioning parameters and system |
| WO2020177539A1 (en) * | 2019-03-01 | 2020-09-10 | 阿里巴巴集团控股有限公司 | Positioning, seeking and tracking methods and apparatuses, and system |
| CN110826524A (en) * | 2019-11-18 | 2020-02-21 | 广东博智林机器人有限公司 | Fingerprint library generation method and device, electronic equipment and storage medium |
| CN110826524B (en) * | 2019-11-18 | 2022-08-16 | 广东博智林机器人有限公司 | Fingerprint library generation method and device, electronic equipment and storage medium |
| CN111065046A (en) * | 2019-11-21 | 2020-04-24 | 东南大学 | LoRa-based outdoor unmanned aerial vehicle positioning method and system |
| CN111065046B (en) * | 2019-11-21 | 2020-12-11 | 东南大学 | LoRa-based outdoor unmanned aerial vehicle positioning method and system |
| CN110996387A (en) * | 2019-12-02 | 2020-04-10 | 重庆邮电大学 | LoRa positioning method based on TOF and position fingerprint fusion |
| CN113051976A (en) * | 2019-12-27 | 2021-06-29 | 广东博智林机器人有限公司 | Fingerprint positioning method and device, electronic equipment and storage medium |
| CN117872269A (en) * | 2024-03-13 | 2024-04-12 | 电子科技大学 | High-precision positioning method for self-adaptive data processing |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109068267B (en) | 2020-06-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109068267A (en) | A kind of indoor orientation method based on LoRa SX1280 | |
| Chen et al. | Distributed and collaborative localization for swarming UAVs | |
| Wang et al. | Information-theoretic approaches for sensor selection and placement in sensor networks for target localization and tracking | |
| CN103476118B (en) | A kind of WLAN indoor location fingerprint positioning method for monitoring in real time | |
| CN108696932A (en) | It is a kind of using CSI multipaths and the outdoor fingerprint positioning method of machine learning | |
| CN103068035B (en) | A kind of wireless network localization method, Apparatus and system | |
| CN105636201B (en) | Indoor orientation method based on sparse signal fingerprint database | |
| KR102116824B1 (en) | Positioning system based on deep learnin and construction method thereof | |
| CN108989976A (en) | Fingerprint positioning method and system in a kind of wisdom classroom | |
| CN104038901B (en) | Indoor positioning method for reducing fingerprint data acquisition workload | |
| CN106793082A (en) | A kind of positioning of mobile equipment method in WLAN/ bluetooth heterogeneous network environments | |
| CN103747419B (en) | A kind of indoor orientation method based on signal strength difference and dynamic linear interpolation | |
| CN112218330B (en) | Positioning method and communication device | |
| CN101352061A (en) | Method and apparatus for determining the location of a mobile object | |
| CN103826298B (en) | Wireless sensor network positioning and computing method for collaborative iterative optimization | |
| CN108650628A (en) | The indoor orientation method that ranging and fingerprint based on Wi-Fi network are combined | |
| CN105973246A (en) | Drawing method and apparatus of geomagnetic map, and robot | |
| Xu et al. | An efficient indoor localization method based on the long short-term memory recurrent neuron network | |
| CN107727095A (en) | 3D indoor orientation methods based on spectral clustering and weighting reverse transmittance nerve network | |
| CN106970379B (en) | Based on Taylor series expansion to the distance-measuring and positioning method of indoor objects | |
| Adege et al. | Applying Deep Neural Network (DNN) for large-scale indoor localization using feed-forward neural network (FFNN) algorithm | |
| CN107743296A (en) | A kind of RSSI area segmentation formula localization methods based on compressed sensing | |
| CN106231549A (en) | A kind of 60GHz pulse indoor orientation method based on restructing algorithm | |
| CN102231911A (en) | Method for carrying out multidirectional scaling positioning on wireless sensor network by distance sensing | |
| CN103249144A (en) | C-type-based wireless sensor network node location method |
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 | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20201216 Address after: 314500 Tongxiang, Jiaxing, Zhejiang, Wutong Street East Road (East) 55, Tongxiang chamber of Commerce Building 1 unit 1702, 1703 room -A-227 Patentee after: Jiaxing Tongzhi Electromechanical Equipment Co.,Ltd. Address before: 310016 Tongxiang, Hangzhou, Zhejiang, Wutong Street East Road (East) 55, Tongxiang chamber of Commerce Building 1 unit 1702, 1703 room -A-227 Patentee before: Zhejiang Zhiduo Network Technology Co.,Ltd. Effective date of registration: 20201216 Address after: 310016 Tongxiang, Hangzhou, Zhejiang, Wutong Street East Road (East) 55, Tongxiang chamber of Commerce Building 1 unit 1702, 1703 room -A-227 Patentee after: Zhejiang Zhiduo Network Technology Co.,Ltd. Address before: 310018 No. 2 street, Xiasha Higher Education Zone, Hangzhou, Zhejiang Patentee before: HANGZHOU DIANZI University |
|
| TR01 | Transfer of patent right |
Effective date of registration: 20210205 Address after: 310051 rooms 103 and 105, building 4, No.1 Weiye Road, Puyan street, Binjiang District, Hangzhou City, Zhejiang Province Patentee after: HANGZHOU CCRFID MICROELECTRONICS Co.,Ltd. Address before: 314500 Tongxiang, Jiaxing, Zhejiang, Wutong Street East Road (East) 55, Tongxiang chamber of Commerce Building 1 unit 1702, 1703 room -A-227 Patentee before: Jiaxing Tongzhi Electromechanical Equipment Co.,Ltd. |
|
| TR01 | Transfer of patent right | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: An indoor location method based on Lora sx1280 Effective date of registration: 20210604 Granted publication date: 20200623 Pledgee: China Minsheng Banking Corp Hangzhou branch Pledgor: HANGZHOU CCRFID MICROELECTRONICS Co.,Ltd. Registration number: Y2021330000513 |
|
| PE01 | Entry into force of the registration of the contract for pledge of patent right | ||
| PC01 | Cancellation of the registration of the contract for pledge of patent right |
Date of cancellation: 20220315 Granted publication date: 20200623 Pledgee: China Minsheng Banking Corp Hangzhou branch Pledgor: HANGZHOU CCRFID MICROELECTRONICS Co.,Ltd. Registration number: Y2021330000513 |
|
| PC01 | Cancellation of the registration of the contract for pledge of patent right |