CN106525052A - Graph model constrained indoor positioning method - Google Patents
Graph model constrained indoor positioning method Download PDFInfo
- Publication number
- CN106525052A CN106525052A CN201611153178.5A CN201611153178A CN106525052A CN 106525052 A CN106525052 A CN 106525052A CN 201611153178 A CN201611153178 A CN 201611153178A CN 106525052 A CN106525052 A CN 106525052A
- Authority
- CN
- China
- Prior art keywords
- samplenode
- particle
- model
- graph model
- sampling
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/20—Instruments for performing navigational calculations
- G01C21/206—Instruments for performing navigational calculations specially adapted for indoor navigation
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/10—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration
- G01C21/12—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning
- G01C21/16—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning by integrating acceleration or speed, i.e. inertial navigation
- G01C21/165—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning by integrating acceleration or speed, i.e. inertial navigation combined with non-inertial navigation instruments
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Navigation (AREA)
Abstract
The invention discloses a graph model constrained indoor positioning method. The method includes the steps that indoor positioning environments are classified according to terrains and functional characteristics; based on classification provided by a map, sampleNode sampling points are obtained according to a given generation rule; sampleNode is sampled; a track push measuring model for measuring the step number and the step length is established; in the moving process of a to-be-positioned target, track positions are subjected to particle filtering; when the to-be-positioned target is moved to the sampleNode, a primary calibration event is triggered, and integrator drift of the track speculation model is calibrated. Compared with the prior art, information hidden in the map is fully utilized, a graph model with a topological structure is established, and in the earlier sampling process, sampling is carried out according to mark points obtained from the model. In the navigation process, by means of the particle filtering algorithm and based on a topological graph, high-precision positioning can be achieved only by scattering a small number of particles.
Description
Technical field
The present invention relates to indoor positioning technologies, more particularly to a kind of indoor orientation method of graph model constraint.
Background technology
Because gps signal cannot cover interior, indoor positioning environment is again complicated and changeable, cannot use conventional positioning side
Method.Recently as the development of intelligent mobile phone sensor technology, using all kinds of perception devices carried on mobile phone, obtain in environment
Scene, wireless signal, the feature such as earth magnetism, merge Various types of data Jing after particular procedure and realize positioning.Common localization method has
Bluetooth positioning, WIFI fingerprint locations, reckoning positioning, earth magnetism fingerprint location.
Some above-mentioned technical methods, realize the positioning of degree of precision, but, some of which needs arrangement substantial amounts of outer
Install it is standby, and some then previous work need the exploration of substantial amounts of manpower, or the complexity that algorithm is realized is higher, it is impossible to individually exist
Process and complete on the limited smart mobile phone of endurance.
The content of the invention
To overcome the deficiencies in the prior art, the present invention to propose a kind of indoor orientation method of graph model constraint, fully profit
It is with the information implied in navigation map, multiple to reduce algorithm during the intricate operation degree and real-time positioning of early stage sampling
Miscellaneous degree.
The technical scheme is that what is be achieved in that：
A kind of indoor orientation method of graph model constraint, including step
S1：Indoor positioning environment is divided into into gallery, half open zone of action according to landform and functional character and is opened
Wealthy zone of action；
S2：Based on given classification is schemed, sampleNode sampled points, the generation rule are drawn according to given create-rule
It is then：
The two ends of gallery, bifurcation and turning point are taken as sampleNode,
The gateway and range of activity corner of half open zone of action is taken as sampleNode；
Multiple gateways of open zone of action are taken as sampleNode, and gateway is carried out shortest path line, line
On sampleNode is taken with one step, while appropriate plane extension is carried out to the sampleNode on line；
S3：The sampleNode is sampled, sampled data includes the ground quantity of magnetism of sampled point and rotation from side to side
Corner；
S4：Set up for measuring the flying track conjecture amount model of step number and step-length；
S5：In the motor process of target to be positioned, particle filter is carried out to track position；
S6：When target to be positioned moves to sampleNode, primary calibration event, the product to flying track conjecture model are triggered
Divide drift calibration.
Further, gallery described in step S1 refers to corridor and balcony, and half open area refers to room, institute
State open zone of action and refer to hall.
Further, the model of flying track conjecture amount described in step S4 includes step number measurement module and adaptive step estimation mould
Block.
Further, step S5 includes step：
S51：Particle is initialized, and each particle is set to identical initial weight；
S52：Particle propagation, including sampling side propagate and sampling node propagate, wherein sampling side propagate when, choose angle with
Propagate on the minimum side of flight-path angle difference；Sampling node is propagated and then carries out probability propagation according to estimating step length；
S53：Right value update, is divided into the renewal based on ground constraint diagram and the renewal based on sampleNode；
S54：State estimation, does weighted sum, then does average, draw estimated position, this estimated position to all of particle
The position of user is presented to for system then；
S55：Particle resampling, removes the too low particle of weights, while replicating the particle of high weight so that total number of particles is protected
Hold constant.
Further, primary calibration event is triggered in step S6 needs to meet three conditions：
The position that last system is estimated is near sampleNode to be triggered；
The anglec of rotation for detecting is consistent with the anglec of rotation being pre-stored in sampleNode storehouses；With
The earth magnetism for measuring is matched with the ground quantity of magnetism being pre-stored in sampleNode storehouses.
The beneficial effects of the present invention is, compared with prior art, the information that the present invention is implied in making full use of map is built
The vertical graph model with topological structure, in early stage sampling process, is sampled according to the mark points drawn from model.Navigate through
Cheng Zhong, using particle filter algorithm, on the basis of topological diagram, it is only necessary to spread a small amount of particle and just can obtain determining for degree of precision
Position.
Description of the drawings
Fig. 1 is the indoor orientation method flow chart of graph model constraint of the present invention.
Fig. 2 is the sampleNode point schematic diagrams that the classification in one embodiment of the invention in step S2 according to figure is generated.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation is described, it is clear that described embodiment is only a part of embodiment of the invention, rather than the embodiment of whole.It is based on
Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under the premise of creative work is not made
Embodiment, belongs to the scope of protection of the invention.
The sensor that the present invention need to be used includes：Gyroscope, accelerometer, magnetometer, take on common smart mobile phone
These sensors are carried, it is possible to select on smart mobile phone to implement the program.
Refer to Fig. 1, a kind of indoor orientation method of graph model constraint of the invention, including
Step S1：The plane structure chart of building is taken first, according to landform and functional character by the zone of action in map
It is divided into three classes：Gallery, half open zone of action, open zone of action.Wherein, gallery refers to corridor, balcony, and half
Open area is primarily referred to as room, and open zone of action refers to the interior such as hall open ground.
Step S2：Based on given classification is schemed, sampleNode sampled points are drawn according to given rule, set up
SampleNode Sample Storehouses.Generate sampleNode rule be：
Gallery：Selection corridor two ends, bifurcation, turning point are sampleNode；
Half open behaviour area：SampleNode is taken at gateway and range of activity inside lock；
Open behaviour area：SampleNode is taken in multiple gateways, gateway is carried out into shortest path line, on line with
One step takes sampleNode, while appropriate plane extension is carried out to the sampleNode on line.
The sampleNode points of generation are as shown in Figure 2.
Step S3：Sampled for the sampleNode of labelling in step 2, sampled data includes：The earth magnetism of sampled point
Amount, the anglec of rotation from side to side.
The measuring principle of magnetometer is that measurement earth's magnetic field is incident upon the projection on three coordinate axess of magnetometer, then basis
The ground quantity of magnetism combination of each axle converses ground magnetic vector.People in handheld mobile phone walking process, cannot predict, handss by the attitude of mobile phone
The determination of machine attitude depends on course angle, and in complicated indoor earth's magnetic field environment, it is impossible to by the ground magnetic vector being disturbed
Obtain accurate course angle.So the ground magnetic vector of record collection is nonsensical, in order that Geomagnetism Information does not receive attitude
Affect to match in front and back, in implementation process, only record the absolute force of collection.
The collection of the anglec of rotation, the angle being primarily referred to as when original route is when other are gone to, as this angle is held
The continuous time is shorter, be relative angle, hardly deposits integrator drift.With obvious distinguishing characteristic, therefore can be used in the range of restriction
Triggering calibration event.
Step S4：Flying track conjecture amount model is set up, for measuring cadence and step-length.Flight path measurement model is divided into two moulds
Block, respectively cadence measurement module and adaptive step estimation block.
Step number is detected：The accelerometer data of collection is mainly fast fourier transform (FFT), FFT by step number statistics
Formula is as follows：
Wherein x [n] represents the acceleration that accelerometer is detected, and X [k] is output valve, and j is the number of each frequency wave component
Group, N represent the acceleration information points for participating in calculating.Using sliding window algorithm, newest continuous 100 data points are chosen
As input operator.Again from the element that selected value in X [k] is maximum, the array position Jing conversions residing for element are cadence, only
Frequency just assert it is to take a step in the range of 1Hz-2Hz.
For the estimation of step-length, there are some researches show, step-length and the cadence of normal person's walking and initial acceleration is in when taking a step
Linear relationship：
L=α * f+ β * a+ γ
Wherein L is step-length, and f represents cadence, and a represents starting acceleration of taking a step, and α, β, γ are variable element.Close according to more than
Be formula, linear regression done with method of least square, can draw the value of tri- parameters of α, β, γ, then the parameter for drawing is substituted into into former relation
Formula just can draw step-length estimation dynamical equation.
Step S5：In implementation process, particle filter is carried out to target, including：Particle initialization, particle propagation, weights are more
Newly, state estimation, five stages of particle resampling.
Particle is initialized：In algorithm initialization, the original state of all particles need to be set.Current invention assumes that pedestrian's is first
Beginning position is, it is known that with initial position as geometric center, the N number of particle of uniformly dispersing, the initial weight of each particle are arranged to 1/
N。
Particle propagation：In positioning, when the generation of walking event is detected, particle is propagated under state transition model.
Particle state is made up of node n, direction θ two parts, according to the state transition model that Posterior probability distribution formula can obtain particle is：
{nt, θt}=p (nt, θt|nt-1, θt-1, zθ, t, zD, t, G)
Wherein, nt, θtNode n residing for t particle is represented respectively and towards θ, zθ, t, zD, tStep-length estimation is represented respectively
The direction for drawing and step-length in model, G represent the topological graph model under current environment.During particle propagation, Xia Yizhuan
State is mainly determined by direction and propagation distance, and θ in above formulatOnly by zθ, t, θt-1, G decisions, ntOnly by nt-1, zD, t, G decisions.With
Above formula can be decomposed into by this：
{nt, θt}=p (nt, θt|nt-1, θt-1, zθ, t, zD, t, G)
=p (nt|nt-1, θt-1, ZD, t, G) and p (θt|zθ, t, G)
When selecting side to propagate, it is the multiformity for keeping particle, after step-length model assessment result is obtained, one is carried out to particle
Secondary Gauss sampling, then particle state is drawn according to above formula, select currently towards θtWith select to carry out probability when angle is minimum
Propagate, its expression formula is：
Wherein, etRepresent the side of selection.
Right value update：Right value update is that the weight to particle is reassigned, and can be divided into two classes according to the event of triggering：Base
Renewal in the right value update of walking event and based on sampleNode events.
For the renewal of walking event, when the generation of walking event is detected, z is drawn by step-length appraising modelθ, t, zD, t,
Current particle state is drawn by particle state metastasis model, so as to z can be drawnθ, t, zD, tPosterior probability, right value update functional expression
It is represented by：
wt=wt-1·p(zθ, t|θt, G) and p (zD, t|nt, G)
Wherein, wtRepresent the particle weights of t.
It is for the renewal of sampleNode events, as step-length appraising model has integrator drift, of long duration to produce
Error can constantly accumulate change greatly, and the renewal of sampleNode events then can eliminate this error to a certain extent.According to upper
Formula right value update functional expression, can obtain：
wt=wt-1·p(sθ|θt, G) and p (sn|nt, G)
Wherein sθ、snRefer to the steering angle and coordinate position in sampleNode storehouses respectively.
State estimation：Estimate the current coordinate of target, weighted sum is done to all of particle, just can show that particle is current
Coordinate：
Particle resampling：Particle resampling is to remove the low particle of weights, retains and replicate the high particle of weights, in order that
Obtain particle and do not lose multiformity, in given range, generate superseded factor u at random, particle weights are added up one by one, weights is chosen prominent
The particle (accumulated value is more than upper particle weights during superseded factor u) of change, after choosing n times, you can eliminate low weights grain
Son.
Topological constraint diagram is introduced, computational complexity is substantially reduce the number so that the method for the present invention may be implemented in smart mobile phone
In Deng portable terminal device, it is easy to promote.
The above is the preferred embodiment of the present invention, it is noted that for those skilled in the art
For, under the premise without departing from the principles of the invention, some improvements and modifications can also be made, these improvements and modifications are also considered as
Protection scope of the present invention.
Claims (5)
1. the indoor orientation method that a kind of graph model is constrained, it is characterised in that including step
S1：Indoor positioning environment is divided into into gallery, half open zone of action and open work according to landform and functional character
Dynamic region；
S2：Based on given classification is schemed, sampleNode sampled points, the create-rule are drawn according to given create-rule
For：
The two ends of gallery, bifurcation and turning point are taken as sampleNode,
The gateway and range of activity corner of half open zone of action is taken as sampleNode；
Multiple gateways of open zone of action are taken as sampleNode, and gateway is carried out shortest path line, on line with
One step takes sampleNode, while appropriate plane extension is carried out to the sampleNode on line；
S3：The sampleNode is sampled, sampled data includes the ground quantity of magnetism and the anglec of rotation from side to side of sampled point；
S4：Set up for measuring the flying track conjecture amount model of step number and step-length；
S5：In the motor process of target to be positioned, particle filter is carried out to track position；
S6：When target to be positioned moves to sampleNode, primary calibration event is triggered, the integration of flying track conjecture model is floated
Move calibration.
2. the indoor orientation method that graph model as claimed in claim 1 is constrained, it is characterised in that long and narrow logical described in step S1
Road refers to corridor and balcony, and half open area refers to room, and the open zone of action refers to hall.
3. the indoor orientation method that graph model as claimed in claim 1 is constrained, it is characterised in that flight path is pushed away described in step S4
Measurement model includes step number measurement module and adaptive step estimation block.
4. the indoor orientation method that graph model as claimed in claim 1 is constrained, it is characterised in that step S5 includes step：
S51：Particle is initialized, and each particle is set to identical initial weight；
S52：Particle propagation, including sampling side is propagated and sampling node is propagated, wherein when sampling side is propagated, choosing angle and flight path
Propagate on difference minimum side in angle；Sampling node is propagated and then carries out probability propagation according to estimating step length；
S53：Right value update, is divided into the renewal based on ground constraint diagram and the renewal based on sampleNode；
S54：State estimation, does weighted sum, then does average, draw estimated position to all of particle, and this estimated position is then
System presents to the position of user；
S55：Particle resampling, removes the too low particle of weights, while replicating the particle of high weight so that total number of particles keeps not
Become.
5. the indoor orientation method that graph model as claimed in claim 1 is constrained, it is characterised in that once school is triggered in step S6
Quasi- event needs to meet three conditions：
The position that last system is estimated is near sampleNode to be triggered；
The anglec of rotation for detecting is consistent with the anglec of rotation being pre-stored in sampleNode storehouses；With
During trigger event, the earth magnetism for measuring is matched with the ground quantity of magnetism being pre-stored in sampleNode storehouses.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611153178.5A CN106525052A (en) | 2016-12-14 | 2016-12-14 | Graph model constrained indoor positioning method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611153178.5A CN106525052A (en) | 2016-12-14 | 2016-12-14 | Graph model constrained indoor positioning method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106525052A true CN106525052A (en) | 2017-03-22 |
Family
ID=58339454
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611153178.5A Pending CN106525052A (en) | 2016-12-14 | 2016-12-14 | Graph model constrained indoor positioning method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106525052A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107462247A (en) * | 2017-07-18 | 2017-12-12 | 深圳天珑无线科技有限公司 | A kind of indoor orientation method, device and computer-readable recording medium |
| CN110060102A (en) * | 2019-04-18 | 2019-07-26 | 重庆邮电大学 | Retail shop where user based on inclined label study positions big data prediction technique |
| CN107289941B (en) * | 2017-06-14 | 2021-01-08 | 湖南格纳微信息科技有限公司 | Inertial navigation-based indoor positioning method and device |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103983266A (en) * | 2014-05-28 | 2014-08-13 | 北京天地方元科技有限公司 | Indoor locating method based on geomagnetic information and indoor locating system based on geomagnetic information |
| CN105737826A (en) * | 2016-02-24 | 2016-07-06 | 中国地质大学（武汉） | Indoor positioning method for pedestrian |
-
2016
- 2016-12-14 CN CN201611153178.5A patent/CN106525052A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103983266A (en) * | 2014-05-28 | 2014-08-13 | 北京天地方元科技有限公司 | Indoor locating method based on geomagnetic information and indoor locating system based on geomagnetic information |
| CN105737826A (en) * | 2016-02-24 | 2016-07-06 | 中国地质大学（武汉） | Indoor positioning method for pedestrian |
Non-Patent Citations (4)
| Title |
|---|
| SEBASTIAN HILSENBECK等: "Graph-based Data Fusion of Pedometer and WIFI Measurements for Miobile Indoor Positioning", 《ACM PRESS THE 2014 ACM INTERNATIONAL JOINT CONFERENCE-SEATTLE,WASHINGTON》 * |
| 张聪聪等: "基于地磁场的室内定位和地图构建", 《仪器仪表学报》 * |
| 熊明亮等: "基于地球磁场的室内定位***的研究", 《无线互联科技.无线天地》 * |
| 黎大鹏等: "基于VWMC的传感器网络移动节点定位算法", 《计算机工程与设计》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107289941B (en) * | 2017-06-14 | 2021-01-08 | 湖南格纳微信息科技有限公司 | Inertial navigation-based indoor positioning method and device |
| CN107462247A (en) * | 2017-07-18 | 2017-12-12 | 深圳天珑无线科技有限公司 | A kind of indoor orientation method, device and computer-readable recording medium |
| CN110060102A (en) * | 2019-04-18 | 2019-07-26 | 重庆邮电大学 | Retail shop where user based on inclined label study positions big data prediction technique |
| CN110060102B (en) * | 2019-04-18 | 2022-05-03 | 重庆邮电大学 | Bias label learning-based method for predicting positioning big data of shops where users are located |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110118560B (en) | Indoor positioning method based on LSTM and multi-sensor fusion | |
| Li et al. | Improving positioning accuracy of vehicular navigation system during GPS outages utilizing ensemble learning algorithm | |
| Zhou et al. | Activity sequence-based indoor pedestrian localization using smartphones | |
| Zhang et al. | A handheld inertial pedestrian navigation system with accurate step modes and device poses recognition | |
| Pei et al. | Optimal heading estimation based multidimensional particle filter for pedestrian indoor positioning | |
| CN106840163A (en) | A kind of indoor orientation method and system | |
| CN104781686B (en) | Pathway matching | |
| CN110455291B (en) | Method and apparatus for indoor location-finding using magnetic field anomalies | |
| CN105723242B (en) | Measure the traffic speed in road network | |
| CN107830862A (en) | A kind of method of the indoor positioning pedestrian tracking based on smart mobile phone | |
| CN106714110A (en) | Auto building method and system of Wi-Fi position fingerprint map | |
| CN107255474B (en) | PDR course angle determination method integrating electronic compass and gyroscope | |
| CN105043380A (en) | Indoor navigation method based on a micro electro mechanical system, WiFi (Wireless Fidelity) positioning and magnetic field matching | |
| CN109164411B (en) | Personnel positioning method based on multi-data fusion | |
| CN108604407A (en) | Method and system for generating expected gait of march | |
| CN106767828A (en) | A kind of mobile phone indoor positioning solution | |
| CN104897154A (en) | Mobile phone-based indoor geomagnetic navigation method and apparatus thereof | |
| CN106461753A (en) | Location transition determination | |
| Liu et al. | Indoor localization using smartphone inertial sensors | |
| CN110260855B (en) | Indoor pedestrian navigation positioning method integrating pedestrian dead reckoning, geomagnetic information and indoor map information | |
| CN108020813A (en) | Localization method, positioner and electronic equipment | |
| CN106525052A (en) | Graph model constrained indoor positioning method | |
| Hao et al. | A multi-floor location method based on multi-sensor and WiFi fingerprint fusion | |
| CN108801267B (en) | Indoor anchor-free positioning method integrating multiple sensors | |
| Hasan et al. | Smart phone based sensor fusion by using Madgwick filter for 3D indoor navigation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170322 |