CN104519571B - A kind of indoor orientation method based on RSS - Google Patents
A kind of indoor orientation method based on RSS Download PDFInfo
- Publication number
- CN104519571B CN104519571B CN201410831784.2A CN201410831784A CN104519571B CN 104519571 B CN104519571 B CN 104519571B CN 201410831784 A CN201410831784 A CN 201410831784A CN 104519571 B CN104519571 B CN 104519571B
- Authority
- CN
- China
- Prior art keywords
- signal
- positioning
- fingerprint
- rss
- signal strength
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- 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
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
- G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
- G01S5/0205—Details
- G01S5/0215—Interference
Abstract
The invention discloses a kind of indoor orientation method for being based on RSS (Received signal strength, received signal strength), its data acquisition is convenient, need not additionally increase receiving device, accurate positioning.This indoor orientation method based on RSS (Received signal strength, received signal strength), including step：(1) off-line phase, the radio signal reception strength information on some positions of collection space, fingerprint base is built；(2) the on-line testing stage, the signal intensity on walking path is collected；(3) by using rarefaction representation algorithm, time and space constraints is added, establishes location model；(4) position coordinates corresponding to signal value on path is calculated, result is optimized.
Description
Technical Field
The invention belongs to the technical field of Wireless Local Area Networks (WLAN) indoor positioning, and particularly relates to an RSS (Received signal strength) -based indoor positioning method.
Background
A Wireless Local Area Network (WLAN) is a brand new information acquisition platform, and can implement complex and large-scale positioning, monitoring and tracking tasks in a wide application field, and network node self-positioning is the basis and premise of most applications. The current popular Wi-Fi (Wireless Fidelity ) positioning is a positioning solution of IEEE802.11, which is a series of standards for Wireless local area networks. The system adopts a mode of combining an empirical test and a signal propagation model, is easy to install, needs few base stations, can adopt the same bottom wireless network structure and has high total precision. The main classification is as follows:
● Approximation method
The approximation method utilizes the characteristic that the coverage of the AP in the room is limited (different types of routers have different coverage), and determines the position of the mobile user according to the condition of the received signal strength of the terminal equipment and the position of the corresponding AP. When the user approaches a certain known location, the object is located by that location. That is, the position of an Access Point (AP) that uses the wireless terminal for data communication is approximated as an estimated position. The method can be used for detecting article contact, monitoring the access point of the cellular network and the like. The method does not need complex calculation, but the positioning accuracy is limited in the coverage area of the AP, only regional position judgment can be realized, and the prior information of the specific position of the AP is needed.
● Geometric measurement method
This method first requires mapping the signal strength value to the distance of signal propagation according to a propagation model (empirical or mathematical) of the radio signal. And on a two-dimensional plane, position estimation is carried out through the geometrical principle of trilateration according to the distances between the terminal equipment and at least three other APs. As shown in fig. 1, three black dots are reference points with known coordinates, and x is a point to be located, and the coordinates of x can be calculated by using the distances between x and the three reference points.
Microsoft RADAR (radio detection and ranging) is an indoor positioning system based on RSSI (Received Signal Strength Indication) technology, and is also the earliest indoor positioning system based on WLAN, and the position of a user node in a floor is determined by making full use of the existing WLAN facilities through the Received Signal Strength Indication in the 802.n standard. Two methods are commonly used to compute node positions, one of which is to use a theoretical model of signal propagation. The method has low accuracy, can save cost, does not need to establish a database in advance, and recalculates the parameters after the base station moves. However, in real-world environments, conditions such as temperature, obstacles, propagation modes and the like are often changed, so that the technology still has difficulty in practical application. The method is simple and high in calculation efficiency, but the positioning accuracy depends on whether the propagation model is correct or not and whether the propagation model is suitable for positioning a building structure with a complicated area or not. Due to the complexity of indoor radio wave propagation, the signal strength is affected by multipath propagation, reflection and the like, so that the actual indoor environment is difficult to be characterized by a fixed mathematical model.
● Scene analysis method
Scene analysis, also known as fingerprinting, does not map the measurements of signal strength values directly to signal propagation distances, but uses the scene features observed at a location to infer the observer's position, which can be seen as learning the internal laws between signal strength and position, and then matching the new measurements with the learned sample points.
The method generally comprises two stages of off-line measurement and on-line positioning to complete positioning. In the off-line measurement, a plurality of sample points are selected according to certain spacing distances in an area to be positioned to form a grid of the sample points, measurement is carried out on the positions of the sample points, signal strength measurement vectors from all APs are recorded, and the information forms a signal strength fingerprint library. The fingerprint library describes the relationship of signal strength to spatial location in this fixed positioning environment. In the on-line positioning, the signal intensity measurement vector measured in real time is compared with the information in the fingerprint database, and the position of the sample point with the closest signal intensity is taken as the estimated position.
Another method for computing node location in the RADAR system is to use empirical models of signal propagation. Before actual positioning, a plurality of test points are selected in a floor, the signal intensity received by each base station at the points is recorded, and an off-line database of the relation between the position and the signal intensity at each point is established. In actual positioning, the system adopts a deterministic matching algorithm, namely a K Nearest Neighbor (KNN) algorithm, compares the measured signal intensity with the signal intensity recorded in the database, and the coordinate of the point with the minimum mean square error of the signal intensity is the coordinate of the node. This method has a high accuracy, but requires a database of location and signal strength relationships to be pre-established and re-established as the base station moves.
The system can be embedded into any handheld terminal equipment with a Wifi adapter, and can be used for independently positioning and tracking without additional hardware support and line-of-sight transmission, so that the positioning range is wider than that of a Cricket system. However, due to the complexity of the indoor environment, such as multipath, shadow fading, interference, etc., indoor radio wave propagation has strong time-varying characteristics, so that the performance of positioning is affected to a certain extent.
Compared with a traditional radio wave propagation model, the fingerprint method can more accurately describe the relation between the RSS and the spatial position, and does not need prior information of specific AP positions, so that the fingerprint method is widely applied to an indoor positioning system based on the RSS.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the indoor positioning method based on RSS overcomes the defects of the prior art, is convenient to acquire data, does not need to additionally increase receiving equipment, and is accurate in positioning.
The technical solution of the invention is as follows: the indoor positioning method based on RSS comprises the following steps:
(1) In the off-line stage, wireless signal receiving intensity information on some positions in space is collected to construct a fingerprint database;
(2) In the on-line testing stage, collecting the signal intensity on the walking path;
(3) Establishing a positioning model by using a sparse representation algorithm and adding time and space constraint conditions;
(4) And calculating position coordinates corresponding to the signal values on the path, and optimizing the result.
The invention applies the sparse representation algorithm to establish the positioning model and adds the time and space constraint conditions, so that the data acquisition is convenient, the additional increase of receiving equipment is not needed, and the positioning is more accurate.
Drawings
Fig. 1 is a schematic illustration of position estimation based on geometric measurements.
Fig. 2 is a top view of a real experimental scenario.
3a, b, c, d are schematic diagrams of straight-line path positioning in a real scene according to the method herein, the K-nearest neighbor method, the sparse representation algorithm and the kernel method, respectively.
Detailed Description
The indoor positioning method based on RSS (Received signal strength) comprises the following steps:
(1) In the off-line stage, wireless signal receiving intensity information on some positions in space is collected to construct a fingerprint database;
(2) In the on-line testing stage, collecting the signal intensity on the walking path;
(3) Establishing a positioning model by using a sparse representation algorithm and adding time and space constraint conditions;
(4) And calculating position coordinates corresponding to the signal values on the path, and optimizing the result.
An off-line stage: also called training phase, the process of obtaining fingerprint data and constructing fingerprint database;
fingerprint data: the received signal strength data of the spatial known position refers to signal strength data of a plurality of wireless wifi nodes (APs) obtained by mobile equipment such as a smart phone at the known position;
fingerprint database: a set of spatial fingerprint data is given. The space is generally divided into grids, the fingerprint data of each grid node is recorded according to the position of each grid node, and the fingerprint data of all the nodes form a fingerprint database;
and (3) an online testing stage: in the moving process of the test object, recording the signal intensity on a moving path by using mobile equipment such as a smart phone and the like, and realizing the position estimation of the moving object by adopting a sparse representation model;
sparse representation algorithm: is a representation of a signal by representing a given signal as a linear combination of data in a dictionary, with the aid of a previously obtained dictionary, i.e., a library of fingerprints in this document.
The time constraint condition is as follows: during the moving process of the object, the signal receiving strength recorded on the path is continuously changed, so that the sparse representation of the continuous signals has time continuity;
the space constraint condition is as follows: the wireless signal strength has the characteristic of continuous distribution in space, that is, the signal receiving strength measured at a position in space has continuity and similarity with the signal strength measured at the position around the wireless signal receiving strength, so that the sparse representation of the signal strength at a certain position is only related to the sparse representation of the signal strength at the position in spatial proximity to the position.
The invention applies the sparse representation algorithm to establish the positioning model and adds the time and space constraint conditions, so that the data acquisition is convenient, the additional increase of receiving equipment is not needed, and the positioning is more accurate.
Preferably, the positioning model is obtained in the step (3) through formulas (5) to (7):
where λ 1,2 (which should be two threshold parameters) is the set threshold, Y = [ Y ] 1 ,y 2 ,...y n ]Continuous received signal strength, y, acquired for a moving object during movement i Representing the signal reception intensity vector acquired in the ith time slot, psi being the fingerprint library in step (1), each column in Y represents a sparse representation vector of each column signal in Y. Solved to obtainAccording to the signal position information in the fingerprint library psi, the position information of each column of signals in Y can be obtained, and therefore position positioning is achieved.
Preferably, the optimized result is obtained in step (4) through formula (8):
wherein R is a threshold value and R is theta i (ii) a set of positions greater than a threshold, (x) n ，y n ) Watch (A)The coordinate values shown at the n points are,is the weight at the nth position.
One specific example is given below:
the method mainly comprises sparse representation of signals and a reconstruction algorithm.
The following describes a mathematical model of sparse representation, a real finite one-dimensional discrete-time signal x, which can be regarded as an R N Spatial Nx 1-dimensional column vectors with elements of x [ N ]]N =1,2, …, N. In the case of an image or a high-dimensional data vector, it is converted into a long one-dimensional vector. R N Any signal in space can be represented by an N x 1-dimensional orthogonal basis vectorIs expressed in linear combinations. Handlebar vectorForming NxN orthogonal basis dictionary matrices as column vectorsAny signal x can be represented as
Where theta is a weighting coefficient theta i ＝<x,Formed Nx 1 column vectors, (.) T Representing a transpose operation. It is obvious that x and theta are equivalent representations of the same signal, x being the signal in the real domain and theta being the signal in the psi domain.
The basic idea of signal sparse representation is to represent the main information of a signal by using as few non-0 coefficients as possible, so that the solving process of a signal processing problem is simplified. The existing signal sparse representation method can be divided into two types of orthogonal basis sparse representation and redundant dictionary sparse representation.
Natural signals in the normal domain are all non-sparse, but may be sparse in some transform domains. The orthogonal basis sparse representation just utilizes the characteristic of the signal, and the signal is projected to an orthogonal transformation basis to obtain a sparse or approximately sparse transformation vector. General transformation bases can be flexibly selected according to the characteristics of signals, fourier coefficients and wavelet coefficients of smooth signals, total variation norms of bounded variation functions, gabor coefficients of oscillation signals, curvelet coefficients of image signals with discontinuous edges and the like have enough sparsity, and signals can be recovered through a CS theory. If the orthogonal basis sparse representation method is applied, how to find the orthogonal basis suitable for the wifi signal, and even how to construct the orthogonal basis suitable for the wifi signal to obtain the most sparse representation of the wifi signal, will be a key problem to be researched.
When the signal cannot be sparsely represented by orthogonal basis, a redundant dictionary sparse representation may be employed. The dictionary is selected to fit the structure of the signal to be approximated as good as possible without any restriction. If the redundant dictionary sparse representation method is adopted in the text, the following two problems will be the focus of research: how to construct a redundant dictionary suitable for wifi signals; (2) how to design a fast and effective sparse decomposition algorithm.
For the research method, the psi is constructed by adopting a redundant dictionary method, the specific process is that the signal intensity of N fingerprint points is collected for all N APs by assuming off-line training, an N-N matrix can be obtained,representing the ith fingerprint point, eachAre all N x 1 dimensions and represent measurements of N APs. With the sparse vector θ derived from compressed sensing, each row represents the column of the corresponding fingerprint library, and the value of the element represents the degree of influence of the column on the signal strength. The corresponding column of the fingerprint library represents a fingerprint point and a coordinate point, so that the coordinate position corresponding to the test signal can be obtained by solving theta, and the positioning result is finally obtained.
● Measurement signal y
Measurement signal y j A vector representing all signal reception values of all APs in the jth time period. p is a radical of formula i,j Representing the measured value of the ith AP during the jth time period. N is the number of APs. To obtain the following formula:
from equation (3), θ, which is a sparse vector representing the signal y with as few non-0 coefficients as possible, can be solved. In this document, a sparse representation is used to solve the position information, where y is expressed as formula (2), ψ is the signal strength of all fingerprint points, and θ is solved according to formula (3), where the i-th element in θ corresponds to the i-th column signal strength value in ψ, which corresponds to the i-th coordinate point. Therefore, by solving for θ, a position coordinate point can be obtained.
In the testing stage, a section of continuous signal Y is collected, Y is averagely divided into n sections, and then Y = [ Y ] 1 ,y 2 ,...y n ]，y i A vector representing the received values of all AP signals during the ith run time. Now the fingerprint library ψ can be obtained by sparse representation of the model, e.g. (4)In a series of theta i I.e. byEach column in it represents a sparse vector solution, θ, for each column of signals in Y i The position information of the signal in psi is corresponded, so that the position information of each column of signals in Y is obtained, thereby realizing the positioning of all positions.
The existing positioning algorithm only compares single signals to obtain the final position, and does not pay much attention to the relation of signal points in space and time. In the signal information collected in the training stage, the influence of points around the positioning points on the positioning result should be large in space, the signal information corresponding to the time should be connected in the space, the position information of each point is obtained through sparse representation, and on the basis, the position information of the point and the position information of the point are added with constraint to finally realize positioning.
For each fingerprint point, the signal intensity of the fingerprint points around the fingerprint point is close to the signal intensity, the positions of the fingerprint points are closer to each other, the influence of the points around the fingerprint point is larger, in the text, the value of the corresponding position in the sparse vector is larger, and for a fingerprint library with N fingerprint points, a fingerprint library is constructedSpace constraint, S is as follows
The spatial information constraint threshold should be met.
For the signal value corresponding to the same AP, the position of the last point and the next point in the walking processThe signal information phase difference corresponding to the position of the point should be relatively small, so the obtained theta i And theta j The distances between should not differ much. For a walking path with n location points, constructSuch a constraint, T is as defined in formula (6)
The temporal information constraint threshold should be met.
In summary, the model is obtained as follows:
lambda 1,2 is the set threshold and finally solved according to the above formulaAnd obtaining a final positioning result.
Because theta is not only a sparse vector of 1, a threshold value r is set, and a plurality of positions with values larger than r are taken for final position calculation, as follows:
the method mainly aims at the problem of positioning and identifying the walking position of a person in an actual application environment. And calculating the position point weight by using a sparse representation method. In the walking process of a person, the distance between each position point and the previous position point is close, so that a time constraint condition is added to limit the distance between each position point; each position point has a relation with only a few fingerprint points around the position point, so that the weight value of each fingerprint point is limited by adding space constraint.
An example of a practical environment is described in detail below.
1. Practical Environment embodiment
1. Establishment of real environment and sampling of signal strength data
In an indoor real environment, the experimental site is arranged in the three-layer Beijing university of Industrial science, namely the information north building, the length of the experimental site is 53 meters, and the width of the experimental site is 15 meters, as shown in the attached figure 2. In this experiment we co-sampled RSS values from 90 APs in this region. In the off-line training stage, one person walks in an experimental area with the mobile terminal, and simultaneously records the RSS value and the coordinate. A total of 210 samples were collected in the experiment. To avoid systematic errors, accurate measurements are obtained, and we sample 10 times per sample point. The mean of 10 samples was recorded as the final measurement at that sample point.
2. Designing a walking path
In the on-line testing stage, one person walks in an experimental area with the mobile terminal, a straight line path is designed, and the signal intensity of all position points on the path is obtained.
3. Adding a time and space constraint matrix, and positioning by a constructed sparse representation constraint model
In the invention, time and space constraints are mainly added to the sparse vector, the time constraint matrix mainly represents the constraint between the current position and the next position, the space constraint mainly represents the constraint on the position of the current point, and the influence of surrounding points on the current point is the maximum. And in the positioning stage, the sparse vector is calculated by using the tested signal intensity vector and the constructed fingerprint library through a model of a sparse representation algorithm added with constraint.
4. Obtaining position coordinates from sparse vectors
And obtaining online positioning coordinates by performing product operation on the sparse vectors and the coordinates of the fingerprint database.
The experimental results are as follows:
| error (rice) | Methods of the invention | K nearest neighbor method | Sparse representation algorithm | Nuclear method |
| Straight line path | 1.1284m | 1.5214m | 1.5339m | 1.2859m |
The error is the difference between the real path and the positioning result obtained using the different positioning methods. The K-nearest neighbor method, the sparse representation algorithm and the kernel method are relatively common positioning algorithms based on signal strength. The experimental scenario is a space of 53m × 15m, and the error is in units of meters. The units in the drawing are centimeters, the solid dots are real walking paths, and the star points on the curve are positions obtained by applying an algorithm.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the present invention in any way, and all simple modifications, equivalent variations and modifications made to the above embodiment according to the technical spirit of the present invention still belong to the protection scope of the technical solution of the present invention.
Claims (2)
1. An indoor positioning method based on RSS is characterized by comprising the following steps:
(1) In the off-line stage, wireless signal receiving intensity information on some positions in space is collected to construct a fingerprint database;
(2) In the on-line testing stage, collecting the signal intensity on the walking path;
(3) Establishing a positioning model by using a sparse representation algorithm and adding time and space constraint conditions;
(4) Calculating position coordinates corresponding to the signal values on the path, and optimizing the result;
in the step (3), the positioning model is obtained through formulas (5) to (7):
wherein N is the number of fingerprint points in the fingerprint database, N is the number of position points of the walking path, and lambda 1 ，λ 2 For a set threshold value, Y = [ Y = [) 1 ,y 2 ,...y n ]The continuous received signal strength, Y of Y, acquired during the movement of the moving object j Representing the signal receiving strength vector collected at the j-th time point, psi is the fingerprint library in step (1), each column in Y represents a sparse representation vector of each column of signals in Y; solved to obtainFrom the signal position confidence in the fingerprint library psiAnd obtaining the position information of each column of signals in the Y, thereby realizing position positioning.
2. The RSS based indoor positioning method of claim 1, wherein the step (4) obtains an optimized result through formula (8):
wherein R is a threshold value and R isGreater than threshold in (a) i ,b i ) A coordinate value representing the position point i,is the weight at location point i.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410831784.2A CN104519571B (en) | 2014-12-26 | 2014-12-26 | A kind of indoor orientation method based on RSS |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410831784.2A CN104519571B (en) | 2014-12-26 | 2014-12-26 | A kind of indoor orientation method based on RSS |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104519571A CN104519571A (en) | 2015-04-15 |
| CN104519571B true CN104519571B (en) | 2018-03-09 |
Family
ID=52794144
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410831784.2A Active CN104519571B (en) | 2014-12-26 | 2014-12-26 | A kind of indoor orientation method based on RSS |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104519571B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105813022B (en) * | 2016-03-15 | 2019-04-30 | 广州杰赛科技股份有限公司 | A kind of RF finger print data base construction method and device |
| CN108519577B (en) * | 2018-03-12 | 2023-09-15 | 中国矿业大学(北京) | Distributed positioning method based on compressed sensing TOA characteristic signal fingerprint library |
| CN108627798B (en) * | 2018-04-04 | 2022-03-11 | 北京工业大学 | WLAN indoor positioning algorithm based on linear discriminant analysis and gradient lifting tree |
| CN108668249B (en) * | 2018-07-10 | 2021-01-22 | 北京物资学院 | Indoor positioning method and device for mobile terminal |
| CN109327797B (en) * | 2018-10-15 | 2020-09-29 | 山东科技大学 | Indoor positioning system of mobile robot based on WiFi network signal |
| CN110519704B (en) * | 2019-08-28 | 2021-06-15 | 中国银行股份有限公司 | Positioning method and system of signal sparse representation model based on time constraint |
| CN110516878A (en) * | 2019-08-28 | 2019-11-29 | 中国银行股份有限公司 | A kind of localization method and system of the sparse signal representation model based on space-time restriction |
| CN110505573B (en) * | 2019-08-28 | 2021-08-03 | 中国银行股份有限公司 | Positioning method and system of signal sparse representation model based on space constraint |
| CN110519705A (en) * | 2019-08-28 | 2019-11-29 | 中国银行股份有限公司 | A kind of localization method and system based on sparse signal representation model |
| CN112533137B (en) * | 2020-11-26 | 2023-10-17 | 北京爱笔科技有限公司 | Positioning method and device of equipment, electronic equipment and computer storage medium |
| CN112729301A (en) * | 2020-12-10 | 2021-04-30 | 深圳大学 | Indoor positioning method based on multi-source data fusion |
| CN115086973B (en) * | 2022-08-19 | 2022-11-11 | 深圳市桑尼奇科技有限公司 | Intelligent household human body induction method and device |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103139907A (en) * | 2013-02-04 | 2013-06-05 | 北京工业大学 | Indoor wireless positioning method by utilizing fingerprint technique |
| CN103167606A (en) * | 2013-03-12 | 2013-06-19 | 钱钢 | Wireless local area network (WLAN) indoor positioning method based on sparse representation |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2004095790A1 (en) * | 2003-03-28 | 2004-11-04 | University Of Maryland | Method and system for determining user location in a wireless communication network |
-
2014
- 2014-12-26 CN CN201410831784.2A patent/CN104519571B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103139907A (en) * | 2013-02-04 | 2013-06-05 | 北京工业大学 | Indoor wireless positioning method by utilizing fingerprint technique |
| CN103167606A (en) * | 2013-03-12 | 2013-06-19 | 钱钢 | Wireless local area network (WLAN) indoor positioning method based on sparse representation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104519571A (en) | 2015-04-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104519571B (en) | A kind of indoor orientation method based on RSS | |
| Hsieh et al. | Deep learning-based indoor localization using received signal strength and channel state information | |
| Bisio et al. | Smart probabilistic fingerprinting for WiFi-based indoor positioning with mobile devices | |
| CN103139907B (en) | A kind of indoor wireless positioning method utilizing fingerprint technique | |
| CN103874118B (en) | Radio Map bearing calibrations in WiFi indoor positionings based on Bayesian regression | |
| CN104038901B (en) | Indoor positioning method for reducing fingerprint data acquisition workload | |
| Zhang et al. | An efficient machine learning approach for indoor localization | |
| CN104469937A (en) | Efficient sensor deployment method used in compressed sensing positioning technology | |
| CN102231912A (en) | RSSI ranging-based positioning method for indoor wireless sensor network | |
| CN103220777A (en) | Mobile device positioning system | |
| He et al. | A novel radio map construction method to reduce collection effort for indoor localization | |
| CN110933628B (en) | Fingerprint indoor positioning method based on twin network | |
| Zhang et al. | Using compressive sensing to reduce fingerprint collection for indoor localization | |
| CN103945531A (en) | Method for WLAN indoor positioning Radio Map updating based on information entropy | |
| CN108225332B (en) | Indoor positioning fingerprint map dimension reduction method based on supervision | |
| Ni et al. | Fingerprint-MDS based algorithm for indoor wireless localization | |
| Zhou et al. | Application of backpropagation neural networks to both stages of fingerprinting based WIPS | |
| Zheng et al. | Convex optimization algorithms for cooperative RSS-based sensor localization | |
| Yang et al. | An indoor RFID location algorithm based on support vector regression and particle swarm optimization | |
| Tan et al. | An efficient fingerprint database construction approach based on matrix completion for indoor localization | |
| Milioris et al. | Building complete training maps for indoor location estimation | |
| Arai et al. | Color radiomap interpolation for efficient fingerprint wifi-based indoor location estimation | |
| CN107682822B (en) | Compressed sensing outdoor positioning method based on electromagnetic field intensity | |
| Gunathillake et al. | Maximum likelihood topology maps for wireless sensor networks using an automated robot | |
| Zhang et al. | Kernel-based particle filtering for indoor tracking in WLANs |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |