CN108668234A - A kind of parking lot reverse car seeking method based on WIFI - Google Patents

A kind of parking lot reverse car seeking method based on WIFI Download PDF

Info

Publication number
CN108668234A
CN108668234A CN201810198678.3A CN201810198678A CN108668234A CN 108668234 A CN108668234 A CN 108668234A CN 201810198678 A CN201810198678 A CN 201810198678A CN 108668234 A CN108668234 A CN 108668234A
Authority
CN
China
Prior art keywords
data
formula
follows
signal strength
parking lot
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201810198678.3A
Other languages
Chinese (zh)
Other versions
CN108668234B (en
Inventor
许磊
周渝曦
余星
朱奇武
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chongqing College of Electronic Engineering
Original Assignee
Chongqing College of Electronic Engineering
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Chongqing College of Electronic Engineering filed Critical Chongqing College of Electronic Engineering
Priority to CN201810198678.3A priority Critical patent/CN108668234B/en
Publication of CN108668234A publication Critical patent/CN108668234A/en
Application granted granted Critical
Publication of CN108668234B publication Critical patent/CN108668234B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02—Services making use of location information
    • H04W4/029—Location-based management or tracking services
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/14—Traffic control systems for road vehicles indicating individual free spaces in parking areas
    • G08G1/141—Traffic control systems for road vehicles indicating individual free spaces in parking areas with means giving the indication of available parking spaces
    • G08G1/144—Traffic control systems for road vehicles indicating individual free spaces in parking areas with means giving the indication of available parking spaces on portable or mobile units, e.g. personal digital assistant [PDA]
    • 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
    • H04W4/024—Guidance services
    • 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
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/80—Services using short range communication, e.g. near-field communication [NFC], radio-frequency identification [RFID] or low energy communication
    • 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
    • H04W64/006—Locating users or terminals or network equipment for network management purposes, e.g. mobility management with additional information processing, e.g. for direction or speed determination

Abstract

The invention discloses a kind of parking lot reverse car seeking methods based on Wi Fi, mainly include the following steps that：1) car park areas is determined.The parking lot is divided into m domain block.2) the wireless Wi Fi networks are based on, parking lot Indoor Locating Model is established.3) target vehicle to be positioned is determined.4) according to the parking lot Indoor Locating Model, the target vehicle position is obtained, realizes reverse car seeking.Therefore the present invention realizes the reverse car seeking positioning of car owner, while improving reverse car seeking efficiency, has been greatly saved cost using the Wi Fi networks of parking lot covering.

Description

A kind of parking lot reverse car seeking method based on WIFI
Technical field
The present invention relates to parking lot car searching field, specifically a kind of parking lot reverse car seeking method based on WIFI.
Background technology
With the continuous development of social progress and urbanization, personal wealth constantly increases, and personal private car possesses Amount is also continuous to be increased, and large-scale parking lot also occurs therewith.Currently, most of area of parking lots are big, park cars more.It is in Modern large parking lot, in the vehicle of numerous parkings, car owner, which seeks the vehicle oneself stopped also, to be become to expend the time further, is become It is more difficult.
And current reverse car seeking method is that equipment is added in parking lot mostly, cause establish parking lot cost it is big It is big to increase.
Invention content
Present invention aim to address problems of the prior art.
To realize the present invention purpose and the technical solution adopted is that such, a kind of parking based on Wi-Fi is field-reversed to seek Vehicle method, mainly includes the following steps that：
1) car park areas is determined.The parking lot is divided into m domain block.M is natural number.
The parking lot is covered with wireless Wi-Fi network.The wireless Wi-Fi network is mainly by n Wi-Fi wireless access Point APxComposition.X=1,2 ..., n.
2) it is based on the wireless Wi-Fi network, establishes parking lot Indoor Locating Model.Key step is as follows：
2.1) the signal sequence D and received signal strength RSS data sample set R for parking lot indoor positioning are established.
The key step for establishing signal sequence D and received signal strength RSS data sample set R is as follows：
2.1.1 the regional center of m domain block) is denoted as sampled point P respectivelyy.Y=1,2 ... m.The sampled point PyHave Unique position coordinates.
2.1.2) in the 0-T periods, m sampled point P is obtained respectivelyySignal strength, to build signal sequence D.
In formula, dfgRepresent sampled point PfIn the signal strength at g moment.F=1,2 ..., m.G=1,2 ..., T.
2.1.3) in the 0-T periods, the sampled point P is obtained respectivelyyReceive n Wi-Fi wireless access point APxSignal Intensity RSS data, to constitute the received signal strength RSS data sample set R for indoor positioning.The set of data samples R It indicates as follows：
R={ R1, R2..., Rt..., RT}。 (2)
In formula, R1, R2..., Rt..., RTRespectively the 1st moment, the 2nd moment ..., t moment ..., described in the T moment Sampled point PyReceive n Wi-Fi wireless access point APxSignal strength RSS data.
Wherein, sampled point P described in t momentyReceive n Wi-Fi wireless access point APxSignal strength RSS data RtTable Show as follows：
In formula,For sampled point PhWi-Fi wireless access point APs are received in t momentuSignal strength RSS data.H= 1,2 ..., m.U=1,2 ..., n.T is the time.
2.2) data in the received signal strength data sample set R are pre-processed, after obtaining pretreatment Received signal strength data sample set R*.Data in the signal sequence D are pre-processed, after obtaining pretreatment Signal sequence D*.The pretreatment includes mainly denoising and normalization.
Further, pretreated signal sequence D*As follows：
In formula,Represent pretreatment post-sampling point PfIn the signal strength at g moment.F=1,2 ..., m.G=1,2 ..., T。
Pretreated received signal strength data sample set R*As follows：
R={ R1 *, R2 *..., Rt *..., RT *}。 (5)
In formula, R1 *, R2 *..., Rt *..., RT *1st moment after respectively pre-processing, the 2nd moment ..., t moment ..., the Sampled point P described in the T momentyReceive n Wi-Fi wireless access point APxSignal strength RSS data.
Wherein, sampled point P described in t moment after pretreatmentyReceive n Wi-Fi wireless access point APxSignal strength RSS Data Rt *It indicates as follows：
In formula,To pre-process post-sampling point PhWi-Fi wireless access point APs are received in t momentuSignal strength RSS Data.H=1,2 ..., m.U=1,2 ..., n.T is the time.
2.3) all places that vehicle is parked in the parking lot are detected, stop C is denoted ase.E=1,2 ..., s.S is The sum to park cars in the parking lot.
Further, the key step for detecting all places that vehicle is parked in the parking lot is as follows：
2.3.1 morther wavelet ψ) is setab(t) it is：
In formula, a is scale factor, and b is shift factor.T is the time.
2.3.2 signal x (t)) is set in wavelet basis function ψa,b(t) projection on, i.e. parameter wavelet are WTx(a,b).Parameter Small echo WTx(a, b) is as follows：
In formula, a is scale factor, and b is shift factor.a0=2.b0=1.ψa,b(t) it is morther wavelet.X (t) is signal sequence D*In signal.
2.3.3) by morther wavelet ψa,b(t) and parameter wavelet WTx(a, b) discretization.
After discrete, scale factor a and shift factor b are as follows：
In formula, a0=2.b0=1.Z is set of integers, k, j ∈ Z.K and j is discrete conversion coefficient.
Discrete parameter small echoAs follows：
In formula, a0=2.b0=1.Z is set of integers, k, j ∈ Z.K and j is discrete conversion coefficient.ψa,b(t) it is morther wavelet.x (t) it is signal sequence D*In signal.
2.3.4) by the data set D*First row, i.e. signal sequence D1 *As process object.
2.3.5) to signal sequence D1 *The fitting for carrying out Gaussian function, obtains discrete signal sequence
By the discrete signal sequenceIt substitutes into formula 9, obtains signal sequence D1 *Signal sequence after decomposition
By signal sequencePreceding L data be fitted to Gaussian Profile, wherein probability density function f (x) is as follows：
In formula, x is signal sequenceArbitrary data.μ is mean value.σ2For variance.
2.3.6) judge signal sequenceRear m-L data whether meet the Gaussian Profile.If not meeting, count According to corresponding sampled point PyFor stop.Extract the signal strength RSS to park cars in the stop.
2.3.7) respectively by the data set D*Secondary series, third row ... T row as process object.Repeat step 5 With step 6, all stops are detected, and extract the signal strength RSS to park cars in each stop.
2.4) extraction stop CeThe signal characteristic to park cars.
2.5) according to stop CeWith the signal characteristic to park cars, positioned in real time using regression tree structure parking lot is piled up Model.
Further, the key step for building real-time location model is as follows：
2.5.1) by n column data sample sets R*Respectively as input sample.Determine the initial weight distribution D1 of all samples, Initial weight is distributed D2 ..., and initial weight is distributed Dn。
In formula, m is sampled point PySum.ω1vFor set of data samples R1 *The weight of v-th of data.
2.5.2) to weight DnThe data set R at place* nStudy, obtains basic classification device.Basic classification device is as follows：
In formula, v is sampled point PySequence number.X is set of data samples R*In data.
2.4.3) calculate Rn(x) in data set R* nOn error in classification rate en.Error in classification rate enAs follows：
In formula, yvFor v-th of sampled point.Pro is probability.I (*) is characterized function.ωnvFor set of data samples Rn *V-th The weight of data.M is sampled point PySum.
Wherein, Rn(x) factor alphanAs follows：
In formula, enFor error in classification rate.
2.5.4 data set R) is calculated*Weight distribution Dn+1.Weight distribution Dn+1As follows：
Dn+1=(ωn+1,1,···,ωn+1,v,···,ωn+1,m)。 (16)
In formula, m is sampled point PySum.ωn+1,mFor set of data samples Rn *The weight of than the m-th data.
Wherein, weights omegan+1,vAs follows：
In formula, AnFor standardizing factor.αnFor Rn(x) coefficient.yvFor v-th of sampled point.ωn,vFor weight.
Wherein, standardizing factor AnAs follows：
In formula, αnFor Rn(x) coefficient.yvFor v-th of sampled point.M is total number of sample points.xvFor Rn(x) v-th of number in According to.
2.5.5 the linear combination of basic classification device) is built：
In formula, n is Wi-Fi wireless access point APsxSum.αβFor Rβ(x) coefficient.Rβ(x) it is basic classification device.
2.5.6 final classification device G (x)) is obtained.Final classification device G (x) is as follows：
In formula, sign (*) is sign function.αnFor Rn(x) coefficient.
3) target vehicle to be positioned is determined.
4) according to the parking lot Indoor Locating Model, the target vehicle position is obtained, realizes reverse car seeking.
Further, the key step that corresponding indoor location is obtained by Indoor Locating Model is as follows：
4.1) real-time parking data, that is, signal strength RSS and the reverse car seeking data, that is, signal strength of car owner at this time are acquired RSS, and respectively constitute matrix T and matrix F.
4.2) using the data in matrix T as the signal in the input of real-time location model and the real-time location model Intensity data is matched, to obtain the position of the target vehicle.
4.3) the wireless Wi-Fi network, intelligent terminal is utilized to detect and record the current location of car owner.
4.4) according to the position of the current location of car owner and the target vehicle, intelligent terminal generates navigation routine, realizes Reverse car seeking.
The solution have the advantages that unquestionable.The present invention utilizes a kind of data based on Wi-Fi location fingerprints library Parking lot reverse car seeking method is excavated, auxiliary car owner quickly finds the position of oneself vehicle.The present invention is using fingerprint base side On the basis of method, the initial data of acquisition is pre-processed, fingerprint database is then built, realizes that the reverse car seeking of car owner is fixed Position.
Wi-Fi is at low cost, and the frequency separation where Wi-Fi has lower signal interference, is applied on a large scale. Therefore the present invention realizes the reverse car seeking positioning of car owner, is improving reverse car seeking efficiency using the Wi-Fi network of parking lot covering While, it has been greatly saved cost.
Description of the drawings
Fig. 1 is the abnormal state point of Wi-Fi access points AP1 detections；
Fig. 2 is the abnormal state point of Wi-Fi access points AP2 detections；
Fig. 3 is the abnormal state point of Wi-Fi access points AP3 detections；
Fig. 4 is the abnormal state point of Wi-Fi access points AP4 detections；
Fig. 5 is practical parking stall label and the prediction quasi- label graphic in parking stall；
Fig. 6 is experiment scene schematic diagram；
Fig. 7 is reverse car seeking structural schematic diagram；
Fig. 8 is that reverse car seeking model generates schematic diagram.
Specific implementation mode
With reference to embodiment, the invention will be further described, but should not be construed the above-mentioned subject area of the present invention only It is limited to following embodiments.Without departing from the idea case in the present invention described above, according to ordinary skill knowledge and used With means, various replacements and change are made, should all include within the scope of the present invention.
Embodiment 1：
A kind of parking lot reverse car seeking method based on Wi-Fi, mainly includes the following steps that：
1) car park areas is determined.The parking lot is divided into m domain block.M is natural number.
The parking lot is covered with wireless Wi-Fi network.The wireless Wi-Fi network is mainly by n Wi-Fi wireless access Point APxComposition.X=1,2 ..., n.
2) it is based on the wireless Wi-Fi network, establishes parking lot Indoor Locating Model.Key step is as follows：
2.1) the signal sequence D and received signal strength RSS data sample set R for parking lot indoor positioning are established.
The key step for establishing signal sequence D and received signal strength RSS data sample set R is as follows：
2.1.1 the regional center of m domain block) is denoted as sampled point P respectivelyy.Y=1,2 ... m.The sampled point PyHave Unique position coordinates, to forming position fingerprint.
Location fingerprint " connects the position in actual environment with certain " fingerprint ", and a position corresponds to a uniqueness Fingerprint.This fingerprint can be one-dimensional or multidimensional, for example equipment to be positioned is receiving or sending information, then fingerprint can To be a feature or multiple features (most commonly signal strength) of this information or signal.
2.1.2) in the 0-T periods, the signal strength of m sampled point Py is obtained respectively, to build signal sequence D.
In formula, dfgRepresent sampled point PfIn the signal strength at g moment.F=1,2 ..., m.G=1,2 ..., T.If connect Signal is can not receive, then use -90 indicates signal strength.
2.1.3) in the 0-T periods, the sampled point P is obtained respectivelyyReceive n Wi-Fi wireless access point APxSignal Intensity RSS data, to constitute the received signal strength RSS data sample set R for indoor positioning.
What RSS was inherently calculated or was measured whithin a period of time, therefore it is unreasonable only to acquire a RSS sample. In WiFi network, AP be often required to send a beacon frame, contain some network informations, service group ID (wireless network Name), support transmission rate and some other system informations.This beacon frame is used in a lot in WiFi One of control frame, its about 100ms are sent once, and RSS typically uses this beacon frame to measure.Beacon frames are not It is encrypted, even so a closed network (mobile device fails in connection) can be provided for positioning.Beacon frames are close It is sent in periodically, but is not completely periodically, to need to postpone to send when detecting transmission medium obstruction, It is sent not when obstruction once detecting, sending next time still can be at the 100ms moment estimated before, even if from upper one It is secondary to send also less than 100ms.
The set of data samples R indicates as follows：
R={ R1, R2..., Rt..., RT}。 (2)
In formula, R1, R2..., Rt..., RTRespectively the 1st moment, the 2nd moment ..., t moment ..., described in the T moment Sampled point PyReceive n Wi-Fi wireless access point APxSignal strength RSS data.
Wherein, sampled point P described in t momentyReceive n Wi-Fi wireless access point APxSignal strength RSS data RtTable Show as follows：
In formula,For sampled point PhWi-Fi wireless access point APs are received in t momentuSignal strength RSS data.H= 1,2 ..., m.U=1,2 ..., n.If not receiving signal, use -90 indicates signal strength.T is the time.
2.2) data in the received signal strength data sample set R are pre-processed, after obtaining pretreatment Received signal strength data sample set R*.Data in the signal sequence D are pre-processed, after obtaining pretreatment Signal sequence D*.The pretreatment includes mainly denoising and normalization.
Further, pretreated signal sequence D*As follows：
In formula,Represent pretreatment post-sampling point PfIn the signal strength at g moment.F=1,2 ..., m.G=1,2 ..., T。
Pretreated received signal strength data sample set R*As follows：
R={ R1 *, R2 *..., Rt *..., RT *}。 (5)
In formula, R1 *, R2 *..., Rt *..., RT *1st moment after respectively pre-processing, the 2nd moment ..., t moment ..., the Sampled point P described in the T momentyReceive n Wi-Fi wireless access point APxSignal strength RSS data.
Wherein, sampled point P described in t moment after pretreatmentyReceive n Wi-Fi wireless access point APxSignal strength RSS Data Rt *It indicates as follows：
In formula,To pre-process post-sampling point PhWi-Fi wireless access point APs are received in t momentuSignal strength RSS Data.H=1,2 ..., m.U=1,2 ..., n.T is the time.
2.3) all places that vehicle is parked in the parking lot are detected, stop C is denoted ase.E=1,2 ..., s.S is The sum to park cars in the parking lot.
Further, the key step for detecting all places that vehicle is parked in the parking lot is as follows：
2.3.1 morther wavelet ψ) is seta,b(t) it is：
In formula, a is scale factor, and b is shift factor.T is the time.
2.3.2 signal x (t)) is set in wavelet basis function ψa,b(t) projection on, i.e. parameter wavelet are WTx(a,b).Parameter Small echo WTx(a, b) is as follows：
In formula, a is scale factor, and b is shift factor.a0=2.b0=1.ψa,b(t) it is morther wavelet.X (t) is signal sequence D*In signal.
2.3.3) by morther wavelet ψa,b(t) and parameter wavelet WTx(a, b) discretization.
After discrete, scale factor a and shift factor b are as follows：
In formula, a0=2.b0=1.Z is set of integers, k, j ∈ Z.K and j is discrete conversion coefficient.
Discrete parameter small echoAs follows：
In formula, a0=2.b0=1.Z is set of integers.k,j∈Z.ψa,b(t) it is morther wavelet.X (t) is signal sequence D*In Signal.K and j is discrete conversion coefficient.
2.3.4) by the data set D*First row, i.e. signal sequence D1 *As process object.
2.3.5) to signal sequence D1 *The fitting for carrying out Gaussian function, obtains discrete signal sequence
By the discrete signal sequenceIt substitutes into formula 9, obtains signal sequence D1 *Signal sequence after decomposition
By signal sequencePreceding L data be fitted to Gaussian Profile, wherein probability density function f (x) is as follows：
In formula, x is signal sequenceArbitrary data.μ is mean value.σ2For variance.
2.3.6) judge signal sequenceRear m-L data whether meet the Gaussian Profile.If not meeting, count According to corresponding sampled point PyFor stop.Extract the signal strength RSS to park cars in the stop.
2.3.7) respectively by the data set D*Secondary series, third row ... T row as process object.Repeat step 5 With step 6, all stops are detected, and extract the signal strength RSS to park cars in each stop.
2.4) extraction stop CeThe signal strength RSS to park cars.
2.5) according to stop CeWith the signal strength RSS to park cars, using piling up, regression tree structure parking lot is real-time Location model.
Further, the key step for building real-time location model is as follows：
2.5.1) by n column data sample sets R*Respectively as input sample.Determine the initial weight distribution D1 of all samples, Initial weight is distributed D2 ..., and initial weight is distributed Dn。
In formula, m is sampled point PySum.ω1vFor set of data samples R1 *The weight of v-th of data.
2.5.2) to weight DnThe data set R at place* nStudy, obtains basic classification device.Basic classification device is as follows：
In formula, v is sampled point PySequence number.X is set of data samples R*In data.
2.4.3) calculate Rn(x) in data set R* nOn error in classification rate en.Error in classification rate enAs follows：
In formula, yvFor v-th of sampled point.Pro is probability.I (*) is characterized function.ωnvFor set of data samples Rn *V-th The weight of data.M is sampled point PySum.
Wherein, Rn(x) factor alphanAs follows：
In formula, enFor error in classification rate.
2.5.4 data set R) is calculated*Weight distribution Dn+1.Weight distribution Dn+1As follows：
Dn+1=(ωn+1,1,···,ωn+1,v,···,ωn+1,m)。 (16)
In formula, m is sampled point PySum.ωn+1,mFor set of data samples Rn *The weight of than the m-th data.
Wherein, weights omegan+1,vAs follows：
In formula, AnFor standardizing factor.αnFor Rn(x) coefficient.yvFor v-th of sampled point.ωn,vFor weight.
Wherein, standardizing factor AnAs follows：
In formula, αnFor Rn(x) coefficient.yvFor v-th of sampled point.M is total number of sample points.xvFor Rn(x) v-th of number in According to.
2.5.5 the linear combination of basic classification device) is built：
In formula, n is Wi-Fi wireless access point APsxSum.αβFor Rβ(x) coefficient.Rβ(x) it is basic classification device.
2.5.6 final classification device G (x)) is obtained.Final classification device G (x) is as follows：
In formula, sign (*) is sign function.αnFor Rn(x) coefficient.
Further, the data flow for establishing real-time location model is as follows：
Input variables of the selected data collection x1 as one tree, form are：
The interstitial content of this tree is set as N(1), and set same weight for each sample It is trained using decision Tree algorithms, if some sample is fallen on node, which is 1, remaining node is 0,
Training result h1 is obtained, output is：
Wherein N1The number of nodes set for the 1st.
The label of training result h1 and feature are compared, the sample weights of mistake classification increase, the sample correctly classified Weight is reduced, and obtains new sample weightsUsing x1 as new tree, repeated according to new sample weights One step obtains training result h2.
Wherein N2The number of nodes set for the 2nd.
It repeats f times, obtains the weight of the output result hf and each sample of the f treeTraining Terminate, hf is：
Wherein NfThe number of nodes set for the f.
By above-mentioned hi, i=1 ... f and original feature all connect, and are characterized as after being learnt
Wherein N is the sum of the number of nodes of all trees, wherein
6) feature set that said extracted arrivesAs positioning input data set, it is trained modeling using GT algorithms, is obtained GT algorithm models.
It is some reference mode that the model, which exports result,：
Wherein 1,2,3 ... n numbers for reference mode, m1, m2, m3…mnFor the access node signal measured by certain measurement point Intensity, f1, f2, f3 ... fn are specific signal strength values.
3) target vehicle to be positioned is determined.
4) according to the parking lot Indoor Locating Model, the target vehicle position is obtained, realizes reverse car seeking.
Further, the key step that corresponding indoor location is obtained by Indoor Locating Model is as follows：
4.1) real-time parking data and reverse car seeking data are acquired, and respectively constitute matrix T and matrix F.Parking in real time Data are mainly the signal strength RSS of vehicle.Reverse car seeking data are mainly the signal strength RSS of car owner.And respectively constitute square Battle array T and matrix F.
4.2) using the data in matrix T as the signal in the input of real-time location model and the real-time location model Intensity data is matched, to obtain the position of the target vehicle.
4.3) the wireless Wi-Fi network, intelligent terminal is utilized to detect and record the current location of car owner.
4.4) according to the position of the current location of car owner and the target vehicle, intelligent terminal generates navigation routine, realizes Reverse car seeking.
Fig. 1 to Fig. 4 is, to the abnormality detection of actual signal, wherein horizontal axis is the moment, and the longitudinal axis is wireless with WDF algorithms The intensity of Wi-Fi access points AP, the point in circle are abnormal point, the point that algorithm started for the 65th moment in actual test At the time of being predicted as abnormal occur, the state change occurred with practical 67 moment is almost consistent, and algorithm performance meets the requirements.Fig. 5 For the prediction with GT algorithms to actual anchor point, dot is the label of practical parking stall, and dot adds the point of block form to be pre- The parking stall label of survey, being laminated in for the two is completely correct, and using 2.5 meters as standard, test effect also meets actual utilization It is required that.

Claims (6)

1. a kind of parking lot reverse car seeking method based on Wi-Fi, which is characterized in that mainly include the following steps that：
1) car park areas is determined；The parking lot is divided into m domain block；M is natural number；
The parking lot is covered with the wireless Wi-Fi network；The wireless Wi-Fi network is mainly by n Wi-Fi wireless access Point APxComposition；X=1,2 ..., n；
2) it is based on the wireless Wi-Fi network, establishes parking lot Indoor Locating Model；Key step is as follows：
2.1) the signal sequence D and received signal strength RSS data sample set R for parking lot indoor positioning are established；
2.2) data in the received signal strength data sample set R are pre-processed, to obtain pretreated connect Receive signal strength data sample set R*；Data in the signal sequence D are pre-processed, to obtain pretreated letter Number sequence D*；The pretreatment includes mainly denoising and normalization；
2.3) all places that vehicle is parked in the parking lot are detected, stop C is denoted ase；E=1,2 ..., s；S is described The sum to park cars in parking lot；
2.4) extraction stop CeThe signal characteristic to park cars；
2.5) according to stop CeWith the signal characteristic to park cars, the real-time location model in parking lot is built using regression tree is piled up；
3) target vehicle to be positioned is determined；
4) according to the parking lot Indoor Locating Model, the target vehicle position is obtained, realizes reverse car seeking.
2. a kind of parking lot reverse car seeking method based on Wi-Fi according to claim 1, which is characterized in that establish letter Number sequence D and the key step of received signal strength RSS data sample set R are as follows：
1) regional center of m domain block is denoted as sampled point P respectivelyy；Y=1,2 ... m；The sampled point PyWith unique position Set coordinate；
2) in the 0-T periods, m sampled point P is obtained respectivelyySignal strength, to build signal sequence D；
In formula, dfgRepresent sampled point PfIn the signal strength at g moment；F=1,2 ..., m；G=1,2 ..., T；
3) in the 0-T periods, the sampled point P is obtained respectivelyyReceive n Wi-Fi wireless access point APxSignal strength RSS numbers According to constitute the received signal strength RSS data sample set R for indoor positioning；The set of data samples R indicates as follows：
R={ R1, R2..., Rt..., RT}； (1)
In formula, R1, R2..., Rt..., RTRespectively the 1st moment, the 2nd moment ..., t moment ..., sampled point described in the T moment PyReceive n Wi-Fi wireless access point APxSignal strength RSS data；
Wherein, sampled point P described in t momentyReceive n Wi-Fi wireless access point APxSignal strength RSS data RtIt indicates such as Under：
In formula,For sampled point PhWi-Fi wireless access point APs are received in t momentuSignal strength RSS data；T is the time； H=1,2 ..., m；U=1,2 ..., n.
3. a kind of parking lot reverse car seeking method based on Wi-Fi according to claim 1, which is characterized in that pretreatment Signal sequence D afterwards*As follows：
In formula,Represent pretreatment post-sampling point PfIn the signal strength at g moment；F=1,2 ..., m；G=1,2 ..., T；
Pretreated received signal strength data sample set R*As follows：
R={ R1 *, R2 *..., Rt *..., RT *}； (5)
In formula, R1 *, R2 *..., Rt *..., RT *Respectively pre-process after the 1st moment, the 2nd moment ..., t moment ..., T when Carve the sampled point PyReceive n Wi-Fi wireless access point APxSignal strength RSS data；
Wherein, sampled point P described in t moment after pretreatmentyReceive n Wi-Fi wireless access point APxSignal strength RSS data Rt *It indicates as follows：
In formula,To pre-process post-sampling point PhWi-Fi wireless access point APs are received in t momentuSignal strength RSS data； H=1,2 ..., m；U=1,2 ..., n；T is the time.
4. a kind of parking lot reverse car seeking method based on Wi-Fi according to claim 1, which is characterized in that detection institute The key step for stating all places that vehicle is parked in parking lot is as follows：
1) morther wavelet ψ is seta,b(t) it is：
In formula, a is scale factor, and b is shift factor；T is the time；
2) signal x (t) is set in wavelet basis function ψa,b(t) projection on, i.e. parameter wavelet are WTx(a,b)；Parameter wavelet WTx(a, B) as follows：
In formula, a is scale factor, and b is shift factor；, a0=2；b0=1；ψa,b(t) it is morther wavelet；X (t) is signal sequence D* In signal；
3) by morther wavelet ψa,b(t) and parameter wavelet WTx(a, b) discretization；
After discrete, scale factor a and shift factor b are as follows：
In formula, a0=2；b0=1；Z is set of integers；k,j∈Z；K and j is discrete conversion coefficient；
Discrete parameter small echoAs follows：
In formula, a0=2；b0=1；Z is set of integers；k,j∈Z；ψa,b(t) it is morther wavelet；X (t) is signal sequence D*In signal； K and j is discrete conversion coefficient；
4) by the data set D*First row, i.e. signal sequence D1 *As process object；
5) to signal sequence D1 *The fitting for carrying out Gaussian function, obtains discrete signal sequence
By the discrete signal sequenceIt substitutes into formula 9, obtains signal sequence D1 *Signal sequence after decomposition
By signal sequencePreceding L data be fitted to Gaussian Profile, wherein probability density function f (x) is as follows：
In formula, x is signal sequenceArbitrary data；μ is mean value；σ2For variance；
6) judge signal sequenceRear m-L data whether meet the Gaussian Profile；If not meeting, data are corresponding to adopt Sampling point PyFor stop；Extract the signal strength RSS to park cars in the stop；
7) respectively by the data set D*Secondary series, third row ... T row as process object；Step 5 and step 6 are repeated, It detects all stops, and extracts the signal strength RSS to park cars in each stop.
5. a kind of parking lot reverse car seeking method based on Wi-Fi according to claim 1, which is characterized in that structure is real When location model key step it is as follows：
1) by n column data sample sets R*Respectively as input sample；Determine the initial weight distribution D1 of all samples, initial weight It is distributed D2 ..., initial weight is distributed Dn；
In formula, m is sampled point PySum；ω1vFor set of data samples R1 *The weight of v-th of data；
2) to weight DnThe data set R at place* nStudy, obtains basic classification device；Basic classification device is as follows：
In formula, v is sampled point PySequence number；X is set of data samples R*In data；
3) R is calculatedn(x) in data set R* nOn error in classification rate en；Error in classification rate enAs follows：
In formula, yvFor v-th of sampled point；Pro is probability；I (*) is characterized function；ωnvFor set of data samples Rn *V-th of data Weight；M is sampled point PySum；
Wherein, Rn(x) factor alphanAs follows：
In formula, enFor error in classification rate；
4) data set R is calculated*Weight distribution Dn+1；Weight distribution Dn+1As follows：
Dn+1=(ωn+1,1,···,ωn+1,v,···,ωn+1,m)； (16)
In formula, m is sampled point PySum；ωn+1,mFor set of data samples Rn *The weight of than the m-th data；
Wherein, weights omegan+1,vAs follows：
In formula, AnFor standardizing factor；αnFor Rn(x) coefficient；yvFor v-th of sampled point；ωn,vFor weight；
Wherein, standardizing factor AnAs follows：
In formula, αnFor Rn(x) coefficient；yvFor v-th of sampled point；M is total number of sample points；xvFor Rn(x) v-th of data in；
5) linear combination of basic classification device is built：
In formula, n is Wi-Fi wireless access point APsxSum；αβFor Rβ(x) coefficient；Rβ(x) it is basic classification device；
6) final classification device G (x) is obtained；Final classification device G (x) is as follows：
In formula, sign (*) is sign function；αnFor Rn(x) coefficient.
6. a kind of parking lot reverse car seeking method based on Wi-Fi according to claim 1, which is characterized in that pass through room The key step that interior location model obtains corresponding indoor location is as follows：
1) the signal strength RSS of real-time parking data, that is, signal strength RSS and reverse car seeking data, that is, car owner at this time is acquired, and Respectively constitute matrix T and matrix F；
2) using the data in matrix T and F as the signal strength in the input of real-time location model and the real-time location model Data are matched, to obtain the position of the target vehicle；
3) the wireless Wi-Fi network, intelligent terminal is utilized to detect and record the current location of car owner.
4) according to the position of the current location of car owner and the target vehicle, intelligent terminal generates navigation routine, and realization is reversely sought Vehicle.
CN201810198678.3A 2018-03-12 2018-03-12 WIFI-based reverse car searching method for parking lot Active CN108668234B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810198678.3A CN108668234B (en) 2018-03-12 2018-03-12 WIFI-based reverse car searching method for parking lot

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810198678.3A CN108668234B (en) 2018-03-12 2018-03-12 WIFI-based reverse car searching method for parking lot

Publications (2)

Publication Number Publication Date
CN108668234A true CN108668234A (en) 2018-10-16
CN108668234B CN108668234B (en) 2020-07-28

Family

ID=63784114

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810198678.3A Active CN108668234B (en) 2018-03-12 2018-03-12 WIFI-based reverse car searching method for parking lot

Country Status (1)

Country Link
CN (1) CN108668234B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109166347A (en) * 2018-10-22 2019-01-08 天津传世科技有限公司 A kind of underground parking parking method

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070129974A1 (en) * 2005-12-06 2007-06-07 Sin Etke Technology Co., Ltd. Parking lot reservation system with electronic identification
CN103313389A (en) * 2013-07-02 2013-09-18 公安部第三研究所 Reverse vehicle locating and parking control achieving system and method on basis of wireless positioning technology
CN104270818A (en) * 2014-10-16 2015-01-07 北京北斗天成科技有限公司 Wireless positioning method for vehicles in park
CN104616526A (en) * 2015-01-12 2015-05-13 江苏省交通规划设计院股份有限公司 Vehicle positioning and reverse vehicle searching system and method
CN204463460U (en) * 2015-01-12 2015-07-08 江苏省交通规划设计院股份有限公司 A kind of vehicle location and oppositely seek the system of car
CN105654764A (en) * 2016-02-05 2016-06-08 浙江浙大中控信息技术有限公司 Indoor parking lot positioning method
CN105931481A (en) * 2016-05-23 2016-09-07 王�华 Wireless sensor and radio signal positioning-based intelligent reverse vehicle seeking system

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070129974A1 (en) * 2005-12-06 2007-06-07 Sin Etke Technology Co., Ltd. Parking lot reservation system with electronic identification
CN103313389A (en) * 2013-07-02 2013-09-18 公安部第三研究所 Reverse vehicle locating and parking control achieving system and method on basis of wireless positioning technology
CN104270818A (en) * 2014-10-16 2015-01-07 北京北斗天成科技有限公司 Wireless positioning method for vehicles in park
CN104616526A (en) * 2015-01-12 2015-05-13 江苏省交通规划设计院股份有限公司 Vehicle positioning and reverse vehicle searching system and method
CN204463460U (en) * 2015-01-12 2015-07-08 江苏省交通规划设计院股份有限公司 A kind of vehicle location and oppositely seek the system of car
CN105654764A (en) * 2016-02-05 2016-06-08 浙江浙大中控信息技术有限公司 Indoor parking lot positioning method
CN105931481A (en) * 2016-05-23 2016-09-07 王�华 Wireless sensor and radio signal positioning-based intelligent reverse vehicle seeking system

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109166347A (en) * 2018-10-22 2019-01-08 天津传世科技有限公司 A kind of underground parking parking method

Also Published As

Publication number Publication date
CN108668234B (en) 2020-07-28

Similar Documents

Publication Publication Date Title
CN107247938B (en) high-resolution remote sensing image urban building function classification method
CN108989976B (en) Fingerprint positioning method and system in intelligent classroom
CN109165623B (en) Rice disease spot detection method and system based on deep learning
CN107290741B (en) Indoor human body posture identification method based on weighted joint distance time-frequency transformation
CN109640269B (en) Fingerprint positioning method based on CSI and time domain fusion algorithm
CN109325418A (en) Based on pedestrian recognition method under the road traffic environment for improving YOLOv3
CN108109423B (en) Underground parking lot intelligent navigation method and system based on WiFi indoor positioning
CN106446895A (en) License plate recognition method based on deep convolutional neural network
CN109151750B (en) LTE indoor positioning floor distinguishing method based on recurrent neural network model
CN106792465A (en) A kind of indoor fingerprint map constructing method based on mass-rent fingerprint
CN107037399A (en) A kind of Wi Fi indoor orientation methods based on deep learning
CN107949054A (en) Based on high-precision fingerprint positioning method in deep learning visible ray room
CN107564290A (en) A kind of urban road intersection saturation volume rate computational methods
CN107817466B (en) Indoor positioning method based on stacking limited Boltzmann machine and random forest algorithm
CN108171119B (en) SAR image change detection method based on residual error network
CN109151995A (en) A kind of deep learning recurrence fusion and positioning method based on signal strength
CN112653991A (en) WLAN indoor positioning method of TebNet neural network model based on deep learning
CN110933633B (en) Onboard environment indoor positioning method based on CSI fingerprint feature migration
CN110536256A (en) A kind of indoor orientation method based on double layer grid
CN110717496A (en) Complex scene tree detection method based on neural network
CN113836999A (en) Tunnel construction risk intelligent identification method and system based on ground penetrating radar
CN113596989A (en) Indoor positioning method and system for intelligent workshop
CN108632763A (en) A kind of indoor positioning weighting k nearest neighbor method based on WiFi fingerprints
CN103345552A (en) Method and device for assessing reliability of power ICT communication network
CN112950926A (en) Urban trunk road speed prediction method based on big data and deep learning

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