CN105157699A - Indoor parking lot navigation method based on fusion of WiFi and sensor network technology - Google Patents

Indoor parking lot navigation method based on fusion of WiFi and sensor network technology Download PDF

Info

Publication number
CN105157699A
CN105157699A CN201510340593.0A CN201510340593A CN105157699A CN 105157699 A CN105157699 A CN 105157699A CN 201510340593 A CN201510340593 A CN 201510340593A CN 105157699 A CN105157699 A CN 105157699A
Authority
CN
China
Prior art keywords
geomagnetic
parking
wifi
parking stall
intensity
Prior art date
Application number
CN201510340593.0A
Other languages
Chinese (zh)
Other versions
CN105157699B (en
Inventor
李飞
陶兴亮
陈建新
季薇
程媛
Original Assignee
南京邮电大学
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 南京邮电大学 filed Critical 南京邮电大学
Priority to CN201510340593.0A priority Critical patent/CN105157699B/en
Publication of CN105157699A publication Critical patent/CN105157699A/en
Application granted granted Critical
Publication of CN105157699B publication Critical patent/CN105157699B/en

Links

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/04Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by terrestrial means
    • G01C21/08Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by terrestrial means involving use of the magnetic field of the earth

Abstract

The invention discloses an indoor parking lot navigation method based on fusion of WiFi and sensor network technology. The indoor parking lot navigation method comprises a parking cell detection module and a fusion positioning module, wherein the fusion positioning module determines whether a parking cell is occupied by a vehicle according to intensity changes of an earth magnetic field before and after parking. The parking cell detection method has the advantages of low cost, easiness in installation and the like compared with a traditional detection method. The fusion positioning module adopts a fusion positioning method in which geomagnetic sensing network and WiFi positioning technology are fused together. The geomagnetic sensing network can acquire the area and lane where a vehicle is located in real time; since the geomagnetic sensing network and the WiFi positioning technology are fused together for positioning, positioning precision can be greatly improved, and the precise position of the vehicle in a parking lot is provided; and through combined usage of information about unoccupied parking cell provided by parking cell detection technology, fast parking by the user is facilitated.

Description

A kind of parking garage air navigation aid based on WiFi and Sensor Network technological incorporation

Technical field

The present invention relates to field of navigation technology, particularly a kind of parking garage air navigation aid based on WiFi and Sensor Network technological incorporation.

Background technology

Since reform and opening-up, the economy of China is in the state of high speed development always, and people's living standard improves day by day, and car has driven the gate of every household into, bring great convenience to the trip of people, but also make simultaneously automobile parking and parking lot management problem increasingly severe.

As the main place of city parking, parking lot is the key breakthrough points solving Urban Parking Difficulty.The construction in a large amount of rise parking lot, enough parking stall quantity is provided to demander of parking, although can the effective difficult problem of stopping at release, but owing to being restricted in the administrative skill in parking lot, can not fully effectively utilize these parking resources, significantly increase the enlarging difficulty in parking lot.

On the other hand, from current urban parking area administrative skill aspect deep study and analysis, there is following two problems in discovery: 1. in parking stall measure, is all at present to arrange a detection module on each parking stall, adds difficulty to parking lot cost and construction complexity; Can not be provided him accurate location in parking lot to user in the parking lot in 2. current city, conveniently find best parking stall with user, add the resource consumption that some are unnecessary.

Solve the problem of these two aspects above-mentioned, improving operational efficiency and the parking stall utilization factor of urban parking area, alleviate in development of automobile socialization process that the road occurred is crowded, parking difficulty and park construction handling cost is high etc. that problem serves crucial effect, advance the paces of research and development intelligent parking lot parking guidance system.

In parking stall measure, current parking lot mainly uses following several technology:

1) inductive coil detection technique

Current inductive coil detection technique is domestic and international most widely used one in the detection technique of parking lot, this technology is by the coil be buried under parking stall and can detect that the electronic equipment that this coil inductance changes forms, principle adds that a voltage produces corresponding electric current to coil, magnetic field is produced again by electric current, vehicle enters out-of-date, the metal parts of vehicle body can disturbing magnetic field, the inductance of coil is caused to change, finally detected by electronic equipment, system just can realize the renewal of parking space state according to this testing result.This technology is also applied to vehicle detection at present, can identify vehicle [3].Although this technology is reliable and cheap, can breaking road surface thus affect traffic when construction.

2) ultrasonic detection technology

This technology installs a ultrasonic detector at the top of each parking stall, measures the duty information of parking stall by measuring transmitted wave.Detecting device is by directional transmissions ultrasound wave, and hyperacoustic a part of emitted energy runs into the receiving end that road surface and car body surface can reflex to detecting device, just can obtain the status information of parking stall after processing.When vehicle has driven hyperacoustic transmitting boundary into, ultrasound wave can from the top reflective of vehicle to receiver, inconsistent to the distance of receiver with ground launch, will produce one like this and the signal that vehicle is put in storage be detected.

3) infrared detection technology

Infrared detection technology is divided into active infrared detection technique and passive infrared detection technique two kinds.Active infrared detection technique is similar to ultrasonic detection technology, and adopt infrared light supply to launch infrared ray to measurement parking stall, when infrared ray runs into vehicle, part energy can reflex to receiver, with this, can provide the status information of parking stall.Passive infrared detection technique is the duty information being judged parking stall by the difference of infrared radiation of detector detection external environment infrared radiation and motor car engine, and the Detection results of this detection technique is not very desirable.

4) earth magnetism sensing detection technology

Geomagnetic sensing technology be utilize terrestrial magnetic field ferromagnet (iron and steel parts of automobile) by time change detect parking stall duty information, geomagnetic sensor can detect the intensity in magnetic field of the earth 1/10000 and the change in direction accurately.It is different from the geomagnetic field intensity after parking that parking stall is stopped front, and therefore geomagnetic sensor can detect the status information of parking stall.The impact of this detection technique not climate, installation, maintenance is convenient, and does not need closed track, destroys road surface, only needs the top providing a support installing in parking stall.

Because geomagnetic sensor is very responsive to the change of geomagnetic field intensity, and be not easy by environment impact (like rain, snow, wind mist etc.), be widely used in recently stop detect in.E.Sifuentes, O.Casas, the people such as R.Pallas-Areny propose and achieve a kind of method using the wireless sensing net node of arouse machine processed to carry out parking stall measure, the wireless sensing net node that they use is made up of geomagnetic sensor and optical sensor, optical sensor can detect the illumination reduction caused by the arrival of vehicle or other objects, once there be object close, optical sensor can wake geomagnetic sensor up to carry out parking stall measure, adopts the consumption that more can reduce node electric energy in this way.XiangkeGuan, the people such as ZushengZhang propose a kind of relative extreme Parking Cell Detection Algorithms based on wireless magnetometric sensor network, this proposes on the basis of minimax algorithm, and their detection system is through the test of half a year, and detecting successful probability can reach 98.8%.

At present, using geomagnetic sensor to carry out parking stall measure is all need to arrange a node in each parking stall, does not take into full account cost undoubtedly like this.

In current intelligent parking lot inducible system, user's function of locating all is not provided, the technology of indoor positioning is not applied in intelligent parking lot inducible system.WiFi location fingerprint indoor orientation method at present based on received signal strength is widely used in indoor positioning, and many scholars have also done many research to the WiFi location fingerprint indoor positioning algorithms based on received signal strength.

In 2000, Microsoft proposed the indoor locating system that is named as RADAR, and this system is the positioning system of use WiFi signal the earliest as basis on location.This system is in WiFi signal coverage, the received signal strength value of each WiFi node is gathered in certain some reference position point, reference mode coordinate and each received signal strength value are formed a fingerprint, finally a lot of fingerprints is deposited and put in database, form fingerprint database.After fingerprint database has built, in the tuning on-line stage, collected the received signal strength value of visible WiFi node around by terminal to be positioned, form one group of signal observed value be associated.Finally use NN algorithm to mate with the data in fingerprint database, select the estimated position of mating most, be position location.

The algorithm be applied at present based on the WiFi location fingerprint indoor positioning of received signal strength mainly contains NN and KNN algorithm.Their position fixing process is all first carry out the WiFi signal intensity of sample reference node in off-line phase and set up fingerprint database, then carries out location matches when on-line stage.NN algorithm (nearest neighbor algorithm) select in matching result Euclidean distance minimum as positioning result, and KNN algorithm (k nearest neighbor algorithm) is K the matching result that before selecting, K Euclidean distance is minimum, centroid algorithm is then utilized to ask the barycenter of this K result as final positioning result.Zhang Xiaoliang, Zhao's equality people proposes a kind of KNN algorithm of optimization, and this algorithm can under the prerequisite of precision ensureing location fingerprint indoor positioning, the effective calculated amount reduced in position fixing process.Lu Henghui, Liu Xingchuan, Zhang Chao, the people such as Lin Xiaokang locate the WiFi based on triangle and location fingerprint recognizer and compare, result of study shows that the WiFi compared based on triangle algorithm locates, and not only there is larger advantage the WiFi location of position-based algorithm for recognizing fingerprint in availability, and is greatly improved in positioning precision, under their experimental situation, positioning precision is maximum can improve 92.08%.Indoor positioning technologies based on WiFi is applied by Zhu Zhongyi in museum; spectators are made can oneself to control to visit rhythm; the deficiency that museum's Traditional Man explanation mode exists is made up; so not only solve museum and have the contradiction between the protection of collection and displaying for a long time; also drive the information system management of other business in museum, improve work efficiency.Current WiFi indoor positioning technologies is merely able to the possible error of guarantee 50% ~ 60% in 2m, precision need improve, and there is the defect that error is larger.

Summary of the invention

Technical matters to be solved by this invention overcomes the deficiencies in the prior art and provides a kind of parking garage air navigation aid based on WiFi and Sensor Network technological incorporation, adopts the targeting scheme that geomagnetic sensor node and WiFi location fingerprint localization method merge significantly can improve positioning precision.

The present invention is for solving the problems of the technologies described above by the following technical solutions:

According to a kind of parking garage air navigation aid based on WiFi and Sensor Network technological incorporation that the present invention proposes, carry out indoor parking navigation in conjunction with parking space state infomation detection and user location, wherein:

Parking space state infomation detection step is as follows:

Step one, in parking lot, arrange n geomagnetic sensor node, arrange a geomagnetic sensor node between every two adjacent parking stalls, the adjacent parking stall of q geomagnetic sensor node both sides is respectively q parking stall and q+1 parking stall; Wherein, q is integer and 1≤q≤n, and the direction in setting sensing the earth's core, parking stall is l 1direction, the direction that q+1 points to parking stall q parking stall is l 2direction;

Step 2, in advance respectively l is arranged to the position at n geomagnetic sensor node place 1direction geomagnetic field intensity threshold value a q, l 2direction geomagnetic field intensity threshold value b q;

Step 3, initialization, each geomagnetic sensor node detects l respectively 2direction geomagnetic field intensity l 1direction geomagnetic field intensity

Step 4, each geomagnetic sensor node detect current l respectively 1direction geomagnetic field intensity value ZMag q, calculate current with l during initialization 1the difference of direction geomagnetic field intensity is

Step 401: if then enter step 5;

Step 402: if then enter step 6;

Step 5, each geomagnetic sensor node detect current l respectively 2direction geomagnetic field intensity value YMag q, calculate current with l during initialization 2the difference of direction geomagnetic field intensity is

Step 501: if then having car to enter q parking stall, has stopped in q parking stall; Return step 3;

Step 502: if then having car to enter q+1 parking stall, has stopped in q+1 parking stall; Return step 3;

Step 6, each geomagnetic sensor node is adopted to detect current l respectively 2direction geomagnetic field intensity value YMag q, calculate current with l during initialization 2the difference of direction geomagnetic field intensity is

Step 601: if then have car to leave q parking stall, q parking stall is without car; Return step 3;

Step 602: if then have car to leave q+1 parking stall, q+1 parking stall is without car; Return step 3;

User's positioning step is as follows:

Steps A, n geomagnetic sensor node is set in parking lot, between every two adjacent parking stalls, a geomagnetic sensor node is set;

Step B, in off-line training step, set up location fingerprint database; Specific as follows: in parking lot, be set to reference position every the position of predeterminable range, travel through these reference positions and on each reference position, gather WiFi location fingerprint feature simultaneously, and by this stored in location fingerprint database; Subregion is carried out to position fingerprint database in the region detected according to geomagnetic sensor node;

Step C, in the tuning on-line stage, first the WiFi information of user present position is gathered, then region residing for user is detected to the data analysis of geomagnetic sensor node collection, mating doing with gathered WiFi information with the location fingerprint database of region respective regions residing for user, adopting nearest neighbor algorithm or k nearest neighbor algorithm to calculate the rough location of user;

Step D, employing projection theory, be corrected to the concrete orientation in track, car place, this projection fed back to user as final position location by rough location.

As a kind of further prioritization scheme of parking garage air navigation aid based on WiFi and Sensor Network technological incorporation of the present invention, described in zMag i, YMag ithe mode de-jitter being all through smothing filtering obtains.

As a kind of further prioritization scheme of parking garage air navigation aid based on WiFi and Sensor Network technological incorporation of the present invention, described WiFi location fingerprint feature comprises the location coordinate information of the MAC Address of WiFi node, received signal strength value and reference point.

As a kind of further prioritization scheme of parking garage air navigation aid based on WiFi and Sensor Network technological incorporation of the present invention, the predeterminable range in described step B is 1 meter.

As a kind of further prioritization scheme of parking garage air navigation aid based on WiFi and Sensor Network technological incorporation of the present invention, described geomagnetic sensor node is HMC584 geomagnetic sensor.

The present invention adopts above technical scheme compared with prior art, there is following technique effect: the parking stall measure part of native system adopts geomagnetic sensor node to detect the state of parking stall, this method for detecting parking stalls compared to traditional detection method, have cost low, the advantages such as simple are installed.Fusion localization part adopts the amalgamation location technology based on geomagnetic sensor node and WiFi, the existing geomagnetic sensor resource in parking lot can be utilized to carry out fusion location, increase substantially the precision of location, there is provided the exact position of car in parking lot, the empty parking space information provided in conjunction with parking stall measure technology can help user fast to stop.

Accompanying drawing explanation

Fig. 1 is present system structural representation.

Fig. 2 is three direction of principal axis geomagnetic field intensity schematic diagram.

Fig. 3 is geomagnetic sensor node deployment and placement schematic diagram.

Fig. 4 is the geomagnetic sensor node image data analogous diagram without smothing filtering; Wherein, (a) is X axis geomagnetic field intensity Value Data, and (b) is Y-axis geomagnetic field intensity Value Data, and (c) is Z-axis direction geomagnetic field intensity Value Data, and (d) is three axle geomagnetic field intensity Value Datas.

Fig. 5 is through the geomagnetic sensor node image data analogous diagram of smothing filtering; Wherein, (a) is X axis geomagnetic field intensity Value Data, and (b) is Y-axis geomagnetic field intensity Value Data, and (c) is Z-axis direction geomagnetic field intensity Value Data, and (d) is three axle geomagnetic field intensity Value Datas.

Fig. 6 is geomagnetic field intensity change analogous diagram before and after stopping in parking stall; Wherein, (a) is X axis geomagnetic field intensity Value Data, and (b) is Y-axis geomagnetic field intensity Value Data, and (c) is Z-axis direction geomagnetic field intensity Value Data, and (d) is three axle geomagnetic field intensity Value Datas.

Fig. 7 is geomagnetic field intensity mutation analysis figure before and after stopping in adjacent two parking stalls; Wherein, (a) is X axis geomagnetic field intensity Value Data, and (b) is Y-axis geomagnetic field intensity Value Data, and (c) is Z-axis direction geomagnetic field intensity Value Data, and (d) is three axle geomagnetic field intensity Value Datas.

Fig. 8 is the amalgamation indoor orientation method off-line phase schematic diagram based on geomagnetic sensor and WiFi.

Fig. 9 is the amalgamation indoor orientation method tuning on-line stage schematic diagram based on geomagnetic sensor and WiFi.

Figure 10 is fusion and positioning method position correction schematic diagram.

Figure 11 be vehicle by during geomagnetic sensor to geomagnetic field intensity impact analysis figure; Wherein, (a) is X axis geomagnetic field intensity Value Data, and (b) is Y-axis geomagnetic field intensity Value Data, and (c) is Z-axis direction geomagnetic field intensity Value Data, and (d) is three axle geomagnetic field intensity Value Datas.

Figure 12 is NN location algorithm and fusion location algorithm comparative analysis figure.

Figure 13 is KNN location algorithm and fusion location algorithm comparative analysis figure.

Embodiment

Below in conjunction with accompanying drawing, technical scheme of the present invention is described in further detail:

System architecture of the present invention as shown in Figure 1, comprises parking stall measure module and merges locating module.The collection of parking stall measure module in charge parking space information, for user provides empty parking space information; Merge locating module to be responsible for providing car positioning function, for user provides him at the particular location in parking lot, facilitate it to look for best parking stall.

Parking stall measure module of the present invention adopts geomagnetic sensor to detect the duty information of parking stall, and wherein detect adjacent two parking stall measure for individual node, geomagnetic sensor node deployment as shown in Figure 3.In testing process, first pre-service is carried out to the data of each geomagnetic sensor node collection, all can have an impact to the geomagnetic field intensity in this region when vehicle stops leaving parking stall into parking stall or vehicle, utilize geomagnetic sensor node to detect geomagnetic field intensity, adopt the Parking Cell Detection Algorithms of the present invention's design to detect the status information of parking stall.The deployment orientation of base area Magnetic Sensor node in Parking Cell Detection Algorithms, adopts corresponding geomagnetic field intensity uniaxially to detect parking stall.

Fusion locating module of the present invention uses and locates based on the amalgamation indoor positioning algorithms of geomagnetic sensor and WiFi, this algorithm uses WiFi location fingerprint indoor orientation method to carry out Primary Location, utilizes the Primary Location result of existing geomagnetic sensor network to WiFi location fingerprint indoor orientation method to be optimized and improves positioning precision further.

Described carries out de-jitter to the data of geomagnetic sensor node collection, the data that each geomagnetic sensor node collects are the three axial geomagnetic field intensities at geomagnetic sensor node present position place, as Fig. 2, the unit of node locality magnetic field intensity is mGause.Geomagnetic field intensity is wherein XMag, YMag, ZMag are respectively the magnetic field intensity on X-axis, Y-axis and Z axis.Due to the impact of neighbourhood noise, this value has certain fluctuation, shakes the metrical error brought here for preventing, and adopts the mode of smothing filtering to carry out Key dithering.Absolute force after each level and smooth is:

A ( i ) = G ( i ) + G ( i - 1 ) + &CenterDot; &CenterDot; &CenterDot; + G ( 1 ) i , i < L G ( i ) + G ( i - 1 ) + &CenterDot; &CenterDot; &CenterDot; + G ( i - L + 1 ) L , i &GreaterEqual; L - - - ( 1 )

Wherein, i=1,2,3 ..., be the number of times of geomagnetic sensor locality magnetic field intensity, G (i) is the geomagnetic field intensity value that geomagnetic sensor gathers for i-th time, L is the width of the window of smoothing filter, and A (i) is the field strength values after de-jitter.

The setting of described parking stall measure threshold value, in parking lot, geomagnetic field intensity value on each position is different, therefore an empirical value can not be obtained as threshold value, after the difference of the geomagnetic field intensity of front and back of stopping to node is analyzed, the difference of the geomagnetic field intensity value before and after stopping with the surveyed area of each node sets threshold value as a reference.In native system, base area Magnetic Sensor node put orientation, adopt and corresponding axially parking stall to be detected.Here for the geomagnetic sensor node putting position shown in Fig. 3, the change of the geomagnetic field intensity of Z-axis direction is adopted to judge berthing of vehicle and leave parking stall.Here need to arrange a threshold value for judging the change of Z-axis direction geomagnetic field intensity, the impact of stress release treatment.Here adopt the change of the geomagnetic field intensity of Y-axis for judging parking stall, vehicle parking (or leaving) node left (parking stall 1) or parking stall, right (parking stall 2).Equally here also need to arrange a threshold value for judging the change of Y-axis geomagnetic field intensity, the impact of stress release treatment.

Described Parking Cell Detection Algorithms, here taking to detect Y-axis geomagnetic field intensity and Z-axis direction geomagnetic field intensity mode (specifically uses the geomagnetic field intensity of certain axis to do parking stall measure and node is put relevant, as Fig. 3), detect Y-axis geomagnetic field intensity be for distinguish car leave with stop into, detecting Z-axis direction geomagnetic field intensity is leave (stop into) left or right parking stall for car, need all respectively when detecting the geomagnetic field intensity of two axis to arrange a threshold value, exceed threshold value and just indicate that corresponding action occurs.Here the threshold value detecting Z axis geomagnetic field intensity gets a, and the threshold value detecting Y-axis geomagnetic field intensity is b, and detection algorithm step is as follows:

Step 1: initialization, detects Y, Z-axis direction geomagnetic field intensity, record Y-axis earth magnetism field intensity average YMag_Mean, and record Z axis earth magnetism field intensity average ZMag_Mean, enters step 2;

Step 2: detect current Z axis geomagnetic field intensity value ZMag, calculates ZMag-ZMag_Mean;

Step 201: if ZMag-YMag_Mean>a, has car to enter parking stall, enters step 3;

Step 202: if ZMag-YMag_Mean<a, has car to leave parking stall, enters step 4:;

Step 3: detect current Y-axis geomagnetic field intensity value YMag, calculates YMag-YMag_Mean;

Step 301: if YMag-YMag_Mean<b, car enters parking stall 1, enters step 1 and parking space information is back to user;

Step 302: if YMag-YMag_Mean>b, car enters parking stall 2, enters step 1 and parking space information is back to user;

Step 4: detect current Y-axis geomagnetic field intensity value YMag, calculates YMag-YMag_Mean;

Step 401: if YMag-YMag_Mean>b, car leaves parking stall 1, enters step 1 and parking space information is back to user;

Step 402: if YMag-YMag_Mean<b, car leaves parking stall 2, enters step 1 and parking space information is back to user;

Described peak detection algorithm, vehicle is out-of-date from node bypass, to the shake that geomagnetic field intensity herein can be once very strong, can see an obvious crest in oscillograph.In algorithm, first de-jitter is carried out to raw data, after waveform is level and smooth, carries out the judgement of peak value, here using noise in environment on 2 of the size that geomagnetic field intensity affects times of threshold values as peakvalue's checking, namely all peak values are all higher than mean value, its difference threshold value for this reason.Due to the length of vehicle and the restriction of speed, only have at most a car to pass through in 1s, therefore can only produce a crest, so in the data gathered for 1s in the algorithm, if there is multiple crest, only get wherein maximum one.When vehicle is by certain node, adopts peak detection algorithm can be tested with car and pass through, according to the deployed position of node, the region residing for vehicle can be learnt.In this programme, the both sides of two-way traffic is all laid with geomagnetic sensor node, as shown in Figure 10, node serial number near left-lane is all odd number, near right lane is all even number, when vehicle is by geomagnetic sensor, can judge the track residing for vehicle according to the numbering of node.Corresponding fingerprint base can be chosen carry out location matches in region residing for vehicle at fusion positioning stage, according to the track that vehicle travels, matching result is corrected, thus improve positioning precision.

The described amalgamation indoor positioning algorithms based on geomagnetic sensor and WiFi is on the basis of WiFi location fingerprint indoor orientation method, utilizes the existing geomagnetic sensor network of parking position detection module to carry out merging.WiFi location fingerprint indoor orientation method comprises two stages, off-line training step and tuning on-line stage.

Off-line training step

The object in this stage is to set up a location fingerprint database in the server, in parking lot, reference position is set to every the position of a meter, on each reference position, gather the received signal strength value from existing WiFi node while of traveling through these reference positions, the location coordinate information of the MAC Address of each WiFi node, received signal strength value and reference point is formed the tlv triple data be associated and be kept in location fingerprint database.Subregion is carried out to position fingerprint database in the region detected according to geomagnetic sensor, zoning unit size track area size as shown in Figure 3.

The tuning on-line stage

The corresponding location fingerprint storehouse that geomagnetic sensor provides be according to when car from geomagnetic sensor side by time, the geomagnetic field intensity at geomagnetic sensor place has one and significantly shakes.Therefore, adopt peak detection algorithm, what geomagnetic sensor can be real-time detects passing through of car, thus obtains the region residing for vehicle, adopts the location fingerprint reservoir area coupling of respective regions during On-line matching.The numbering of the node passed through according to there being car can judge the track residing for vehicle, and position correction can be carried out to matching result in the track travelled for known vehicle, finally obtains the position residing for vehicle.

Need the user of location in locating area, use the received signal strength value being provided with all WiFi nodes of android mobile phone Real-time Collection of native system positioning software, and MAC Address and received signal strength value are formed two tuples, as the input data of location matches algorithm, carry out coupling estimating user rough location with the corresponding location fingerprint storehouse that specific matching algorithm and geomagnetic sensor provide, be employed herein NN (nearest neighbor algorithm) algorithm or KNN (k nearest neighbor algorithm) algorithm.

Here track is two-way traffic, and car may travel at left-lane, right lane, is the precision improving location, needs the lane detection that travelled by vehicle out here, for further position correction.In this programme, determine by the numbering of node the track that vehicle travels according to vehicle.The step of algorithm is as follows:

Step 1: system initialization;

Step 2: wait for Location Request;

Step 201: the request for location services receiving user, enters step 3;

Step 202: the request for location services not receiving user, enters step 2;

Step 3: obtain the MAC Address of each WiFi node that scanning input arrives and two tuples of received signal strength value composition, enter step 4;

Step 4: geomagnetic sensor network inspection vehicle region, obtains corresponding location fingerprint storehouse, these two tuples is mated with the location fingerprint in location fingerprint storehouse, adopts NN algorithm or KNN algorithm to estimate the rough location of user, enters step 5;

Step 5: in search geomagnetic sensor network, car is by the information of node, changed by the absolute force of inspection vehicle region geomagnetic sensor X axis (putting position of base area Magnetic Sensor is determined), passing through of vehicle is detected by peak detection algorithm, judge the concrete orientation (left-lane or right lane) being obtained track, vehicle place by the numbering of node, enter step 6;

Step 6: adopt projection theory, rough location is corrected to the concrete orientation in track, car place, this projection is fed back to user as final position location, enters step 2;

The parking garage navigational system of the WiFi that the present invention proposes and Sensor Network technological incorporation is primarily of parking stall measure module and merge locating module composition, the data of geomagnetic sensor collection are transmitted by ZigBee, by Serial Port Transmission to server, server carries out process to upgrade parking stall duty information to the data gathered.User is that the android smart phone installation client software carried by oneself obtains parking space information and positional information.At positioning stage, user passes through MAC Address and the received signal strength value of all WiFi nodes around mobile telephone scanning, scanning information is formed two tuples and send to server by WiFi network, server calculates truck position according to this information in conjunction with fusion location algorithm of the present invention and feeds back to user, user checks the position of oneself car by mobile phone, and look for a best parking stall with this, the structural drawing of system is as shown in Figure 1.

The model of the geomagnetic sensor that geomagnetic sensor node of the present invention adopts is HMC584, and it is three axle geomagnetic sensors, and can detect the intensity in magnetic field of the earth 1/10000 and the change in direction, numerical value unit is mGause.In native system, the sampling rate of geomagnetic sensor is 10hz.Fig. 4 is geomagnetic field intensity that geomagnetic sensor collects and geomagnetic field intensity three number of axle size variation according to noise effect in actual environment.Have the shake of ± 10 in figure as seen, this can bring certain error effect in parking stall measure, therefore adopts the pretreatment mode of smothing filtering to carry out Key dithering here, and as described in formula (1), L gets 50 here.Pretreated effect as shown in Figure 5.

The actual width recording each parking stall is 2.5m, and length is 5m, and track is two-way traffic, and lane width is about 6m.Native system parking stall measure module judges the status information (whether parking stall has car) of parking stall to the geomagnetic field intensity change before and after the parking of parking stall place by geomagnetic sensor.Here by the centre of two parking stalls adjacent in geomagnetic sensor node deployment and parking lot, dispose as shown in Figure 3, adopt the state of individually Magnetic Sensor monitoring nodes two parking stalls, reach and reduce node consuming cost.In actual test, the geomagnetic field intensity before and after stopping to two adjacent parking stalls is analyzed, and analysis result figure as shown in Figure 6.Many experiments data find, vehicle parking is put in storage or left parking stall and all can have an impact to the geomagnetic field intensity of Z-axis direction.Here adopt node disposing way shown in Fig. 3, parking warehouse-in can make Z-axis direction geomagnetic field intensity increase, and leaves parking stall and Z-axis direction geomagnetic field intensity can be made to reduce.Therefore need here to arrange a threshold value a, after namely stopping and before stopping, (before leaving parking stall and after leaving parking stall) Z axis geomagnetic field intensity difference exceedes this threshold value and can judge have car to stop into (having car to leave).During parking warehouse-in, shown in Fig. 3, parking stall 1 parking can make the geomagnetic field intensity value of Y-axis increase, and parking stall 2 parking can make the geomagnetic field intensity value of Y-axis reduce, and leaves parking stall situation contrary, shown in actual data analysis Fig. 7.Therefore also need here to arrange a threshold value b, for judging vehicle parking (leaving) parking stall 1 or parking stall 2.

The present invention carries out indoor parking navigation in conjunction with parking space state infomation detection and user location, can generate guidance path, carry out indoor parking navigation according to parking space state information, user's locating information.

By parking lot WiFi network, parking space information is shown, for finding empty parking space at cell-phone customer terminal in real time in native system.Parking space information is stored in MySQL database.

The fore-end of node deployment in the middle of two parking stalls of two parking stall measure, as shown in Figure 3.Have multiple adjacent parking stall unit as shown in Figure 3 in parking lot, each unit in a database storage mode is as shown in table 1, is numbered, parking stall 1 state, parking stall 2 is numbered and parking stall 2 state forms by unit number, node serial number, parking stall 1.Be similar to six parking stall measure, its threshold value a, b analyze and try to achieve in real data.

Table 1

In the test of reality, carried out the measurement of twice metrical error according to the difference of temperature, measurement result is as shown in table 2.

Table 2

Temperature Actual stop frequency Correct detection number of times Accuracy 18 degrees Celsius 300 294 98% -4 degrees Celsius 215 212 98.6%

The amalgamation indoor orientation method adopting geomagnetic sensor and WiFi is merged in locating module in the present invention, setting up the WiFi network that WiFi fingerprint database uses is existing WiFi node, and tuning on-line stage client and server communication use WiFi network.

Off-line training step server adopts MySql database to carry out store and management WiFi fingerprint database, use java language service end WiFi fingerprint base data manipulation routine, gather WiFi location fingerprint and use millet 1S (android4.0) mobile phone, parking lot size is 40m*80m.When setting up fingerprint base using in parking lot every the place of 1m as a reference position, gather location fingerprint data, stored in fingerprint database in each reference position.Data layout is made up of the location coordinate information of the MAC Address of each WiFi node, received signal strength value and reference point.The received signal strength value of all WiFi nodes is gathered once every 2s in each reference position, due to the instability of WiFi signal, not at every turn can be collected for the node more weak from the reference position long-range guided missile number of writing, therefore gather the WiFi node that collected number of times in 10 times is less than 7 times to be excluded, finally the received signal strength value of the WiFi node do not excluded is averaging, using the received signal strength value of this mean value as this WiFi node, stored in fingerprint database.

Tuning on-line stage service end uses java language positioning service program, user uses the received signal strength value being provided with all WiFi nodes of android mobile phone Real-time Collection in parking lot of client, and two tuples of MAC Address and received signal strength value composition are sent to server.Server obtains car region by the Detection Information of geomagnetic sensor network, adopts corresponding location fingerprint database to go coupling two tuple, obtains position according to a preliminary estimate.The Detection Information of base area Magnetic Sensor obtains the track that vehicle travels, and adopts projecting method to correct position according to a preliminary estimate, as shown in Figure 10, obtains the accurate location of user's car and feed back to user.Fusion and positioning method comprises off-line phase and on-line stage, and schematic diagram as shown in Figure 8 and Figure 9.

Here arest neighbors and k nearest neighbor location matches algorithm is adopted to carry out Primary Location to user truck position, arest neighbors location matches algorithm (NN) is qualitative positional fingerprint matching algorithm substantially the most really, and it proposes in the RADAR positioning system of Microsoft first time.The method is the matching process based on analogical learning, uses the sampling sample of online positioning stage and the sampling sample of off-line training step to carry out similarity mode.The received signal strength average of training stage is called location fingerprint, and use Euclidean distance describes the similarity before location fingerprint and location fingerprint, finally, gets the highest location fingerprint of similarity as estimated position.

It is R that the online received signal strength value of definition t measures vector twith the fingerprint vector at reference point j place in fingerprint database euclidean distance as shown in formula (2):

DIST ( R i , R j &OverBar; ) = | | R i - R j &OverBar; | | 2 - - - ( 2 )

Wherein, R j &OverBar; = ( rssi 1 , j &OverBar; , rssi 2 , j &OverBar; , &CenterDot; &CenterDot; &CenterDot; , rssi n , j &OverBar; ) , represent the received signal strength mean value from n WiFi node at reference point j place in WiFi fingerprint database.Finally, estimated position is the minimum point of Euclidean distance, as shown in formula (3):

Location = min DIST ( R t , R j &OverBar; ) - - - ( 3 )

Simultaneously, because the impact of numerous disturbing factors on the received signal strength value of WiFi node of environment is larger, also been proposed k nearest neighbor algorithm (KNN), by choosing K less reference point of distance in formula (4), by trying to achieve the average of this K reference position coordinate, be used as last estimated position with this average;

Location &OverBar; = 1 K &Sigma; j = 1 K Location j - - - ( 4 )

Location jfor a jth position in K the Euclidean distance minimum position that k nearest neighbor algorithmic match goes out.

In geomagnetic sensor network, the position coordinates of each geomagnetic sensor in parking lot is known, and when user drives by geomagnetic sensor, the geomagnetic field intensity change at geomagnetic sensor place as shown in figure 11.By peak detection algorithm, the geomagnetic field intensity in X-axis is detected, can pass through this node by real-time judge user car, so just can detect the region at car place in real time, do positional information coupling by the location fingerprint storehouse choosing respective regions, draw the Primary Location result merging location.During this programme is implemented, base area Magnetic Sensor node location is numbered geomagnetic sensor, car is when left-lane travels, the nodal test that only can be numbered as odd number goes out, similar, car is when right lane travels, and the geomagnetic sensor nodal test that only can be numbered as even number goes out, and therefore can detect the track that vehicle travels in this way.When detecting the concrete orientation in track, car place, so just can correct fusion and positioning method Primary Location result, project in the concrete orientation in track by the mode of projection, as shown in figure 12 (hypothesis detects that car passes through in the middle of track) here.

The coordinate supposing the two ends in concrete orientation, track, car place is (x 1, y 1), (x 2, y 2), the Primary Location position of fusion and positioning method is (x 3, y 3), the final positioning result of fusion and positioning method is subpoint place Primary Location result projected in the concrete orientation in track, car place, and being set to (x, y) this result can pass through following solving equations:

( x - x 3 , y - y 3 ) &CenterDot; ( x 1 - x 2 , y 1 - y 2 ) = 0 x - x 1 y - y 1 = x 1 - x 2 y 1 - y 2 - - - ( 5 )

NN algorithm and KNN algorithm compare with the amalgamation indoor positioning algorithms based on geomagnetic sensor and WiFi respectively by the present invention under actual scene, find in the analysis to 189 positioning results, adopt the targeting scheme that geomagnetic sensor and WiFi location fingerprint localization method merge significantly can improve positioning precision, locating effect as shown in Figure 12 and Figure 13.

More than just the preferred embodiment of the present invention is described.Concerning those skilled in the art, other advantage and distortion can be associated easily according to above embodiment.Therefore, the present invention is not limited to above-mentioned embodiment, and it carries out detailed, exemplary explanation as just example to a kind of form of the present invention.Not deviating from the scope of present inventive concept, the usual change that those of ordinary skill in the art carry out in the aspects of the technology of the present invention and replacement, all should be included within protection scope of the present invention.

Claims (5)

1. based on a parking garage air navigation aid for WiFi and Sensor Network technological incorporation, it is characterized in that, carry out indoor parking navigation in conjunction with parking space state infomation detection and user location, wherein:
Parking space state infomation detection step is as follows:
Step one, in parking lot, arrange n geomagnetic sensor node, arrange a geomagnetic sensor node between every two adjacent parking stalls, the adjacent parking stall of q geomagnetic sensor node both sides is respectively q parking stall and q+1 parking stall; Wherein, q is integer and 1≤q≤n, and the direction in setting sensing the earth's core, parking stall is l 1direction, the direction that q+1 points to parking stall q parking stall is l 2direction;
Step 2, in advance respectively l is arranged to the position at n geomagnetic sensor node place 1direction geomagnetic field intensity threshold value a q, l 2direction geomagnetic field intensity threshold value b q;
Step 3, initialization, each geomagnetic sensor node detects l respectively 2direction geomagnetic field intensity l 1direction geomagnetic field intensity
Step 4, each geomagnetic sensor node detect current l respectively 1direction geomagnetic field intensity value ZMag q, calculate current with l during initialization 1the difference of direction geomagnetic field intensity is
Step 401: if then enter step 5;
Step 402: if then enter step 6;
Step 5, each geomagnetic sensor node detect current l respectively 2direction geomagnetic field intensity value YMag q, calculate current with l during initialization 2the difference of direction geomagnetic field intensity is
Step 501: if then having car to enter q parking stall, has stopped in q parking stall; Return step 3;
Step 502: if then having car to enter q+1 parking stall, has stopped in q+1 parking stall; Return step 3;
Step 6, each geomagnetic sensor node is adopted to detect current l respectively 2direction geomagnetic field intensity value YMag q, calculate current with l during initialization 2the difference of direction geomagnetic field intensity is
Step 601: if then have car to leave q parking stall, q parking stall is without car; Return step 3;
Step 602: if then have car to leave q+1 parking stall, q+1 parking stall is without car; Return step 3;
User's positioning step is as follows:
Steps A, n geomagnetic sensor node is set in parking lot, between every two adjacent parking stalls, a geomagnetic sensor node is set;
Step B, in off-line training step, set up location fingerprint database; Specific as follows: in parking lot, be set to reference position every the position of predeterminable range, travel through these reference positions and on each reference position, gather WiFi location fingerprint feature simultaneously, and by this stored in location fingerprint database; Subregion is carried out to position fingerprint database in the region detected according to geomagnetic sensor node;
Step C, in the tuning on-line stage, first the WiFi information of user present position is gathered, then region residing for user is detected to the data analysis of geomagnetic sensor node collection, mating doing with gathered WiFi information with the location fingerprint database of region respective regions residing for user, adopting nearest neighbor algorithm or k nearest neighbor algorithm to calculate the rough location of user;
Step D, employing projection theory, be corrected to the concrete orientation in track, car place, this projection fed back to user as final position location by rough location.
2. a kind of parking garage air navigation aid based on WiFi and Sensor Network technological incorporation according to claim 1, is characterized in that, described in zMag i, YMag ithe mode de-jitter being all through smothing filtering obtains.
3. a kind of parking garage air navigation aid based on WiFi and Sensor Network technological incorporation according to claim 1, it is characterized in that, described WiFi location fingerprint feature comprises the location coordinate information of the MAC Address of WiFi node, received signal strength value and reference point.
4. a kind of parking garage air navigation aid based on WiFi and Sensor Network technological incorporation according to claim 1, it is characterized in that, the predeterminable range in described step B is 1 meter.
5. a kind of parking garage air navigation aid based on WiFi and Sensor Network technological incorporation according to claim 1, it is characterized in that, described geomagnetic sensor node is HMC584 geomagnetic sensor.
CN201510340593.0A 2015-06-18 2015-06-18 A kind of parking garage air navigation aid based on WiFi and Sensor Network technological incorporation CN105157699B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510340593.0A CN105157699B (en) 2015-06-18 2015-06-18 A kind of parking garage air navigation aid based on WiFi and Sensor Network technological incorporation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510340593.0A CN105157699B (en) 2015-06-18 2015-06-18 A kind of parking garage air navigation aid based on WiFi and Sensor Network technological incorporation

Publications (2)

Publication Number Publication Date
CN105157699A true CN105157699A (en) 2015-12-16
CN105157699B CN105157699B (en) 2018-03-16

Family

ID=54798640

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510340593.0A CN105157699B (en) 2015-06-18 2015-06-18 A kind of parking garage air navigation aid based on WiFi and Sensor Network technological incorporation

Country Status (1)

Country Link
CN (1) CN105157699B (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105405315A (en) * 2015-12-17 2016-03-16 杭州优橙科技有限公司 Magnetic field sensor intelligent parking detection method based on K neighbor machine learning
CN106412839A (en) * 2016-09-12 2017-02-15 南京邮电大学 Indoor positioning and tracking method based on secondary partition and gradient fingerprint match
CN107289935A (en) * 2016-04-01 2017-10-24 中国航空工业第六八研究所 A kind of indoor navigation algorithm suitable for wearable device
CN107607118A (en) * 2017-08-10 2018-01-19 浙江科技学院 A kind of vehicle positioning method of parking garage
WO2019084935A1 (en) * 2017-11-03 2019-05-09 深圳市沃特沃德股份有限公司 Parking lot positioning method, apparatus and system
CN110035392A (en) * 2018-11-09 2019-07-19 阿里巴巴集团控股有限公司 Whether equipment is located at recognition methods and device and the electronic equipment of target area

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060253226A1 (en) * 2005-04-12 2006-11-09 Ehud Mendelson System and method of detecting and navigating to empty parking spaces
CN102760363A (en) * 2012-07-18 2012-10-31 无锡赛思汇智科技有限公司 Parking place monitoring system based on combination of weak magnetic sensor and self-organized network technology
US20140088872A1 (en) * 2011-06-13 2014-03-27 Vcc Engineering Llc System and method for aligning a particular passenger with a unique vehicle
CN103761789A (en) * 2014-01-22 2014-04-30 哈尔滨工程大学 Total management system and management method based on Internet of Things and cloud computing and for parking lot
CN103985271A (en) * 2014-05-18 2014-08-13 吉林大学 Intelligent parking guiding system based on cloud internet of things

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060253226A1 (en) * 2005-04-12 2006-11-09 Ehud Mendelson System and method of detecting and navigating to empty parking spaces
US20140088872A1 (en) * 2011-06-13 2014-03-27 Vcc Engineering Llc System and method for aligning a particular passenger with a unique vehicle
CN102760363A (en) * 2012-07-18 2012-10-31 无锡赛思汇智科技有限公司 Parking place monitoring system based on combination of weak magnetic sensor and self-organized network technology
CN103761789A (en) * 2014-01-22 2014-04-30 哈尔滨工程大学 Total management system and management method based on Internet of Things and cloud computing and for parking lot
CN103985271A (en) * 2014-05-18 2014-08-13 吉林大学 Intelligent parking guiding system based on cloud internet of things

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
赵聘 等: "利用现有无线局域网进行室内定位算法研究", 《信号处理》 *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105405315A (en) * 2015-12-17 2016-03-16 杭州优橙科技有限公司 Magnetic field sensor intelligent parking detection method based on K neighbor machine learning
CN105405315B (en) * 2015-12-17 2017-06-30 杭州优橙科技有限公司 A kind of magnetic field sensor intelligent parking detection method based on k nearest neighbor machine learning
CN107289935A (en) * 2016-04-01 2017-10-24 中国航空工业第六八研究所 A kind of indoor navigation algorithm suitable for wearable device
CN106412839A (en) * 2016-09-12 2017-02-15 南京邮电大学 Indoor positioning and tracking method based on secondary partition and gradient fingerprint match
CN106412839B (en) * 2016-09-12 2019-09-10 南京邮电大学 Based on secondary partition and the matched indoor positioning of fingerprint gradient and tracking
CN107607118A (en) * 2017-08-10 2018-01-19 浙江科技学院 A kind of vehicle positioning method of parking garage
WO2019084935A1 (en) * 2017-11-03 2019-05-09 深圳市沃特沃德股份有限公司 Parking lot positioning method, apparatus and system
CN110035392A (en) * 2018-11-09 2019-07-19 阿里巴巴集团控股有限公司 Whether equipment is located at recognition methods and device and the electronic equipment of target area
CN110035392B (en) * 2018-11-09 2020-05-12 阿里巴巴集团控股有限公司 Method and device for identifying whether equipment is located in target area or not and electronic equipment

Also Published As

Publication number Publication date
CN105157699B (en) 2018-03-16

Similar Documents

Publication Publication Date Title
CN105118303B (en) Vehicle enters position detecting method under intelligent parking monitoring management system and car-parking model
CN104064031B (en) The vehicle peccancy monitoring of a kind of applying electronic car plate and tracking location system
US20170267233A1 (en) Method for autonomous vehicle parking
JP6431075B2 (en) System, method and apparatus for guidance and alignment for electric vehicles and charging stations
CN105270410B (en) Exact curvature algorithm for estimating for the path planning of autonomous land vehicle
CN106205114B (en) A kind of Freeway Conditions information real time acquiring method based on data fusion
CN104751631B (en) The method that Trip chain mode of transportation is judged based on GPS location and fuzzy theory
CN103033180B (en) A kind of accurate Position Fixing Navigation System of indoor vehicle and method thereof
CN103363988B (en) A kind of method utilizing intelligent mobile phone sensor to realize the navigation of earth magnetism indoor positioning
Nawaz et al. Parksense: A smartphone based sensing system for on-street parking
CN103985250B (en) The holographic road traffic state vision inspection apparatus of lightweight
CN102930744B (en) Parking plot guiding and reverse vehicle tracking method for parking lot
US10466056B2 (en) Trajectory matching using ambient signals
US10336203B2 (en) Charging/discharging station, in particular for a battery-operated vehicle
JP5704236B2 (en) Recommended information provision system
CN101270997B (en) Floating car dynamic real-time traffic information processing method based on GPS data
JP2020115348A (en) Sparse map for autonomous vehicle navigation
CN104914865B (en) Intelligent Mobile Robot Position Fixing Navigation System and method
CN102708680B (en) Commute travel mode identification method based on AGPS technology
CN102779428B (en) Guiding method and system of parking lot
CN103925923B (en) A kind of earth magnetism indoor locating system based on adaptive particle filter device algorithm
JP4568782B2 (en) Location registration device, navigation device, route search device, location registration method, location registration program, and recording medium
CN104180805B (en) Smart phone-based indoor pedestrian positioning and tracking method
CN106908075B (en) Big data acquisition and processing system and electric vehicle endurance estimation method based on big data acquisition and processing system
CN103810851B (en) A kind of traffic trip mode identification method based on mobile phone location

Legal Events

Date Code Title Description
PB01 Publication
C06 Publication
SE01 Entry into force of request for substantive examination
C10 Entry into substantive examination
GR01 Patent grant
GR01 Patent grant
EE01 Entry into force of recordation of patent licensing contract

Application publication date: 20151216

Assignee: Nanjing Nanyou Information Industry Technology Research Institute Co. Ltd.

Assignor: Nanjing Post & Telecommunication Univ.

Contract record no.: X2019980001257

Denomination of invention: Indoor parking lot navigation method based on fusion of WiFi and sensor network technology

Granted publication date: 20180316

License type: Common License

Record date: 20191224

EE01 Entry into force of recordation of patent licensing contract