CN103152823B - A kind of wireless indoor location method - Google Patents

Abstract

The present invention proposes a wireless indoor positioning method, comprising the steps of: using a smart phone to automatically collect wireless fingerprint data and user movement data, forming a fingerprint set F and a distance matrix D', and performing preprocessing; according to the preprocessed fingerprint set F and The distance matrix D′ constructs the fingerprint space; generates a stress-free floor plan, extracts key features, and performs spatial coordinate transformation. Among them, key feature extraction includes corridor identification, room identification, and reference point matching, and spatial coordinate transformation includes floor-level conversion and Room-level transformations. The wireless indoor positioning method of the present invention does not need manual on-site survey of the positioning area, and multiple users in the system can conveniently provide fingerprint information cooperatively. All redundant information can be used as upgrade information.

Description

A wireless indoor positioning method

technical field

The invention belongs to the field of wireless indoor positioning and relates to a wireless indoor positioning method.

Background technique

The popularity of mobile phones and the promotion of ubiquitous computing have spawned a large number of researches on wireless indoor positioning. Most current indoor positioning methods use received signal strength (Received Signal Strength, RSS) to determine the position. And RSS fingerprints can be easily obtained from off-the-shelf wireless network devices (such as WiFi or ZigBee compatible devices).

The traditional wireless fingerprint positioning technology adopts a two-stage model. The first stage is the training stage, or called the signal acquisition stage, that is, the wireless signal strength (such as the Wi-Fi signal strength or Zig-Bee signal strength of different wireless routes) at various positions in the room is artificially measured in advance. record, and store the record result in the entry corresponding to the physical location in the database. Due to the uncertainty of wireless signal propagation and indoor conditions, the acquisition of signal strength data requires a large number of repetitions. At the same time, the accuracy of the physical location also affects the accuracy of the final positioning result, which requires a lot of manpower and time for preparation and site survey (SiteSurvey), and the establishment of a wireless signal map (RadioMap) in advance. After the first stage of training, after the fingerprint database is established, the system enters the second stage, which is the actual service stage. The user can obtain his own wireless fingerprint information in the area where the wireless signal distribution map exists, and send the information to the positioning service module as the basis of the query. Through the positioning algorithm, the wireless fingerprint information sent by the user is compared with the wireless fingerprint information in the database, and the location information with the closest similarity is returned to the user.

Since the wireless fingerprint positioning technology utilizes the existing network facilities for positioning without increasing the additional cost of the system, scholars at home and abroad have conducted a lot of in-depth research on the wireless fingerprint positioning method. But as mentioned above, the existing wireless fingerprint positioning technology also has the following defects:

(1) High-cost and low-efficiency manual site surveys are required

On-site survey requires wireless signal fingerprint sampling and manual labeling for each position in the positioning area, which is time-consuming and laborious, and it is difficult to cover all positions in the positioning area.

(2) Poor tolerance to environmental changes and wireless signal fluctuations

One of the characteristics of the indoor environment is that the environment is changeable, and the wireless signal propagation characteristics are complex, and the signal volatility is large. The method based on site survey is difficult to adapt to the dynamic changes of the environment; the traditional method directly uses the Euclidean distance between wireless signal fingerprints as a feature, which is difficult to adapt to the volatility of wireless signals.

(3) Logical positioning at the room level cannot be achieved

Most of the traditional methods are aimed at the positioning of absolute coordinates, and cannot distinguish between different rooms. In fact, room information has greater application value in practice.

At present, smartphones have powerful computing and communication capabilities, built-in various sensors with rich functions, and are bound to users almost anytime and anywhere. Therefore, the mobile phone can be regarded as an increasingly important information interface between the user and the environment. Through the smartphone and its built-in sensors, it can not only perceive rich environmental data, but also capture the user's motion information, which makes it possible for off-site survey-type indoor positioning.

Contents of the invention

The present invention aims at solving one of the above technical problems at least to a certain extent or at least providing a useful commercial choice. Therefore, the object of the present invention is to propose a wireless indoor positioning method with flexible and convenient data collection and good logic of positioning results.

A wireless indoor positioning method according to an embodiment of the present invention includes the following steps: S1. Using a smart phone to automatically collect wireless fingerprint data and user movement data to form a fingerprint set F and a distance matrix D', and perform preprocessing; S2. The preprocessed fingerprint set F and distance matrix D′ construct a fingerprint space; S3. Generate a stress-free plan, perform key feature extraction, and perform spatial coordinate transformation, wherein the key feature extraction includes corridor identification and room identification Matching with the reference point, the spatial coordinate transformation includes floor-level transformation and room-level transformation.

In one embodiment of the present invention, the S1 further includes: S11. Collecting the signal of the wireless network and the readings of the acceleration sensor and the direction sensor through the smart phone, assuming that there are m wireless network access points in the area, the smart The RSS fingerprint obtained by the mobile phone at a certain position in the area is recorded as an m-dimensional vector f=(s ₁ ,s ₂ ,...,s _m ), where s _i represents the RSS of the i-th wireless network access point value, and define d′ _ij as the distance between f _i and f _j , that is, the number of steps the user walks between the two locations. After the fingerprint collection phase is over, the fingerprint set F={f _i ,i= 1...n}, where n is the number of fingerprints, and the distance matrix D′=[d′ _ij ]; S12. Preprocess the fingerprint set F, for two fingerprints f _i =(s ₁ , s ₂ ,...,s _m ) and f _j =(t ₁ ,t ₂ ,...,t _m ), define the difference between f _i and f _j as If δ _ij is less than the predetermined threshold ∈, f _i and f _j are regarded as the same point in the fingerprint space generation process, otherwise they are regarded as different points; S13. Preprocess the distance matrix D′, according to the shortest The path algorithm calculates the shortest distance between each pair of fingerprints, that is, if there is an intermediate node f _k between f _i and f _j and satisfies d′ _ij >d′ _ik +d′ _kj , then d′ _ij is updated to d ′ _ik +d′ _kj .

In one embodiment of the present invention, the S2 includes: according to the preprocessed fingerprint set F and the distance matrix D', map all the fingerprints to a d-dimensional Euclidean space through the MDS algorithm.

In one embodiment of the present invention, the corridor identification includes: using betweenness to obtain fingerprints on the corridor, wherein, according to the distance between fingerprints, a minimum spanning tree T is established, and the betweenness of each point on the minimum spanning tree is calculated , the part with high betweenness is regarded as the fingerprint on the corridor, which is recorded as the corridor fingerprint set F _c , where the betweenness B(v) is defined as: For a graph G=(V ,E), Among them, σ _st is the number of the shortest path from s to t, and σ _st (v) is the number of the shortest path from s to t passing through v.

In one embodiment of the present invention, the room identification includes: using clustering to obtain a room fingerprint set, wherein the corridor fingerprint set F _c is removed from all fingerprint spaces, and the K-means algorithm is used to analyze the remaining fingerprints FF _c clustering, get k k clusters (denoted as i=1,2,...,k), after clustering, consider the same All points in are from the same room.

In one embodiment of the present invention, the reference point matching includes: using the door of the room as a key reference point for establishing the connection between the stress-free floor plan and the fingerprint space, wherein, firstly, define and is the point closest to the door inside and outside the room door, then define the room door fingerprint set All fingerprints in the F _D can be organized into a chain in the minimum spanning tree T, so the F _D is represented as a vector form F _D =(f ₁ , f ₂ ,...,f _k ), and In the stress-free plan view, define P _D =(p ₁ ,p ₂ ,...,p _k ) as the sampling point closest to each door in the corridor, and the order in which each sampling point appears in P _D is along From one side of the corridor to the other, since the mapping from F _D to P _D is corresponding or just in reverse order, the reference point matching is realized to determine the fingerprint set of a room in the stress-free plan and a certain room in the plan corresponding to.

In one embodiment of the present invention, the conversion at the floor level includes: using a transformation matrix to realize the conversion at the floor level of the stress-free plan view and the fingerprint space visualization map, wherein it is assumed that the coordinates of a fingerprint in _FD are x _i =[x _i ¹ x _i ² ... x _i ^d ] ^T , where d is the dimension of the fingerprint space, and the coordinates of the corresponding position y _i =[y _i ¹ y _i ² ... y _i ^d ] ^T , define A as the transformation matrix of d×d, B=[b ₁ b ₂ ...b _d ] ^T , since there are k=|F _D | equations y _i = _Axi +B, all The above k equations are rewritten as H _i z=G _i , where H _i =[ _xi ^T 1], z=[AB] ^T , G _i =y _i ^T , combined with the k equations, Hz= G, where H _i and G _i are the i-th row of H and G, solved by the least square method Obtain A and B, so that for a fingerprint whose coordinates are x=[x _i ¹ x _i ² ... x _i ^d ] ^T , the sampling point closest to Ax+B can be regarded as its actual position.

In one embodiment of the present invention, the room-level conversion includes: further using the MDS algorithm to transform the fingerprints from the same room into a d-dimensional space, so that the sampling points of the corresponding room form a d-dimensional stress-free plan , using the nearest and farthest points from the room door as reference points, the transformation relationship between the fingerprint room and the physical room can be determined. By repeating the above operations for all the rooms one by one, the stress-free floor plan and the fingerprint space visualization map of each room are realized. room-level transformations.

The wireless indoor positioning method of the present invention does not need manual on-site survey of the positioning area, and multiple users in the system can conveniently provide fingerprint information cooperatively. All redundant information can be used as upgrade information.

Additional aspects and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.

Description of drawings

The above and/or additional aspects and advantages of the present invention will become apparent and comprehensible from the description of the embodiments in conjunction with the following drawings, wherein:

Fig. 1 is a flowchart of an indoor space positioning method according to an embodiment of the present invention.

Fig. 2 is a schematic diagram of the 2D fingerprint space provided by the example of the present invention.

Fig. 3 is a schematic diagram of the 3D fingerprint space provided by the example of the present invention.

Fig. 4 is a schematic diagram of the results of sampling on the indoor plan provided by the example of the present invention.

Fig. 5 is a 2D stress-free plan view generated by the sample points of the plan view provided by the example of the present invention.

Fig. 6 is a 3D stress-free plan view generated by the sample points of the plan view provided by the example of the present invention.

Fig. 7 is a schematic diagram of the minimum spanning tree of the 3D fingerprint space provided by the example of the present invention.

Fig. 8 is a schematic diagram of the clustering results of the corridor part removed from the fingerprint space provided by the example of the present invention.

detailed description

Embodiments of the present invention are described in detail below, examples of which are shown in the drawings, wherein the same or similar reference numerals designate the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the figures are exemplary and are intended to explain the present invention and should not be construed as limiting the present invention.

In describing the present invention, it is to be understood that the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", " Orientation or position indicated by "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", etc. The relationship is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, therefore It should not be construed as a limitation of the present invention.

In addition, the terms "first" and "second" are used for descriptive purposes only, and cannot be interpreted as indicating or implying relative importance or implicitly specifying the quantity of indicated technical features. Thus, a feature defined as "first" and "second" may explicitly or implicitly include one or more of these features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

In the present invention, unless otherwise clearly specified and limited, terms such as "installation", "connection", "connection" and "fixation" should be understood in a broad sense, for example, it can be a fixed connection or a detachable connection , or integrally connected; it may be mechanically connected or electrically connected; it may be directly connected or indirectly connected through an intermediary, and it may be the internal communication of two components. Those of ordinary skill in the art can understand the specific meanings of the above terms in the present invention according to specific situations.

In the present invention, unless otherwise clearly specified and limited, a first feature being "on" or "under" a second feature may include direct contact between the first and second features, and may also include the first and second features Not in direct contact but through another characteristic contact between them. Moreover, "above", "above" and "above" the first feature on the second feature include that the first feature is directly above and obliquely above the second feature, or simply means that the first feature is horizontally higher than the second feature. "Below", "beneath" and "under" the first feature to the second feature include that the first feature is directly below and obliquely below the second feature, or simply means that the first feature has a lower level than the second feature.

The invention belongs to the field of wireless indoor positioning, and relates to a method for realizing an off-site survey type wireless indoor positioning technology. The present invention uses a smart phone as a carrier, cooperates with a variety of embedded sensors, and uses an off-site survey type fingerprint positioning algorithm to deduce the physical distance between wireless fingerprints based on the user's moving path, and uses the MDS algorithm to generate fingerprints based on the physical distance Space system, and then convert it into physical coordinates to realize an invention of positioning service for indoor users.

As shown in FIG. 1 , the indoor space positioning method according to the embodiment of the present invention includes the following steps.

S1. Use the smart phone to automatically collect wireless fingerprint data and user movement data to form a fingerprint set F and a distance matrix D′, and perform preprocessing. Specifically, further include:

S11. Collect data

During the collection phase, the user does not need special training, and just needs to walk in the building and carry out daily activities as usual. The mobile phone carried by the user will collect the RSS characteristics of WiFi at various points on the walking path. At the same time, the acceleration sensor integrated in the mobile phone is used to calculate the number of walking steps between each sampling point of the user, so as to determine the physical distance between each point. Specifically, it is assumed that there are m wireless network access points in area A. For each location in area A, the RSS fingerprint can be defined as an m-dimensional vector f=(s ₁ ,s ₂ ,...,s _m ), where s _i represents the RSS of the i-th wireless network access point value. In addition, we agree that if the i-th wireless network access point cannot be monitored at this location, the value of _si is set to 0. Define d′ _ij as the distance between f _i and f _j , the calculation process is: d′ _ij is positive infinity by default; when the user records f _i at a certain location, he walks to another location to record f _j , d′ _ij That is, the number of steps the user walks between two locations. After the fingerprint collection phase is over, we get a fingerprint set F={f _i ,i=1...n} (n is the number of fingerprints) and a distance matrix D′=[d′ _ij ].

S12. Preprocessing the fingerprint set

In order to construct the fingerprint space, the data needs to be preprocessed. User actions are usually arbitrary and irregular, and walking paths may intersect. Therefore, fingerprints may overlap. The preprocessing process merges similar fingerprints, where "similar" means that they are likely to come from the same (or very close) location. We use δ to represent the difference between fingerprints. For two fingerprints f _i =(s ₁ ,s ₂ ,...,s _m ) and f _j =(t ₁ ,t ₂ ,...,t _m ), define Their difference is

${\mathrm{<span\; class="notranslate">\; \&delta;</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; |</span>$

For f _i and f _j , if their degree of difference δ _ij is less than a predefined threshold ∈, they are regarded as the same point in the fingerprint space generation process. Otherwise, f _i and f _j are treated as distinct points.

S13. Preprocessing the distance matrix

We need to preprocess the raw data from the accelerometer to calculate the walking distance. Theoretically, the distance can be obtained by integrating the time with the quadratic acceleration. However, due to the presence of noise, errors can accumulate rapidly. In order to avoid the accumulation of errors, we use the method of calculating the number of steps through the value of the acceleration sensor, and use the number of steps as an indicator of walking distance. For fingerprints f _i and f _j , if none of the users walked a path connecting them, the distance between them will not be available. However, f _i and f _j can be connected through multiple user paths. Therefore, d′ _ij can be estimated by the length of the shortest path connecting f _i and f _j . In addition, the measured d' _ij may also be updated. For example, we use the classic Floyd-Warshall algorithm (shortest path algorithm) to calculate the minimum distance between each pair of fingerprints. The algorithm complexity is O(n ³ ), where n is the number of fingerprints. If for an intermediate node k such that d' _ij >d' _ik +d' _kj , then d' _ij is updated to d' _ik +d' _kj . Doing so eliminates errors caused by the user walking around or pacing back and forth. For convenience, we still denote the processed distance matrix by D′.

S2. Construct the fingerprint space according to the preprocessed fingerprint set F and the distance matrix D′

Specifically, taking D′ as input, the MDS algorithm maps all fingerprints to a d-dimensional Euclidean space. The fingerprint spaces when d is 2 and 3 are shown in Figure 1 and Figure 2, respectively.

S3. Matching of fingerprint space and actual location

S31. Stress-free plan generation

In architecture and building engineering, a floor plan is a top view illustrating the relationships between rooms, spaces, and other physical features in a building. In floor plans, the distance between walls is often drawn to illustrate the size of the room and the length of the walls. Due to walls and other obstructions, the distance between two locations on the floor plan is not necessarily equal to the walking distance between them. To solve this problem, we introduce stress-free floor plans. As shown in Figure 3, on the plan view, we divide the area of interest with a grid and perform average sampling. According to the characteristics of the fingerprint-based positioning method, the grid spacing can usually be 1-3 meters. If the spacing is too large, the accuracy of positioning will be reduced, while if it is too small, the accuracy will be improved little. By calculating the distance between each sampling point, we get a distance matrix D=[d _ij ], where d _ij represents the walking distance between two sampling points p _i and p _j on the planar graph. Taking D as input, the MDS algorithm maps all sampling points to a d-dimensional Euclidean space. In a stress-free plan view, the Euclidean distance between two points reflects the walking distance between their corresponding locations in the actual plan view. For the convenience of observation, d is taken as 2 and 3, and the generated stress-free plan views are shown in Fig. 4 and Fig. 5, respectively.

S32. key feature extraction

Divided into three steps: corridor identification, room identification, reference point matching

(1) Corridor identification

In a building, corridors connect rooms. Users often have to walk through corridors to get from one room to another. This character of the corridors in real space is also reflected in the stress-free floor plan and fingerprint space. The fingerprints collected in the corridor are at a key position in the fingerprint space. Considering centrality in graph theory, these fingerprints have a large centrality value. In graph theory, the concentration of points can take degree, betweenness, compactness, etc. Examples of the present invention use betweenness to acquire fingerprints in hallways. Conceptually, a point has high betweenness if it has a high probability of appearing on any shortest path. For a graph G=(V,E) consisting of a vertex set V and an edge set E, the centrality is defined as

$\mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; the\; s</span>}<span\; class="notranslate">\; \&NotEqual;</span>\mathrm{<span\; class="notranslate">\; \&nu;</span>}<span\; class="notranslate">\; \&NotEqual;</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; V</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}\frac{{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; st</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&nu;</span>}<span\; class="notranslate">\; )</span>}{{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; st</span>}}}$

Among them, σ _st is the number of the shortest path from s to t, and σ _st (v) is the number of the shortest path from s to t passing through v.

According to the distance between fingerprints, we build a minimum spanning tree T, as shown in Figure 6. Use the betweenness to divide all fingerprints into two parts, the part with high betweenness is regarded as the fingerprint in the corridor, and this part of the fingerprint set is denoted as F _c .

(2) Room identification

Removing Fc from all _fingerprints , we observe that the remaining fingerprints can form several clusters, with the fingerprints between different clusters being spatially separated. Considering the computational efficiency, we use the K-means algorithm to cluster the remaining fingerprints. Taking the value of K as the number of rooms on the floor plan, FF _c is gathered into k clusters (denoted as i=1,2,...,k), as shown in Figure 7. After clustering, we consider the same All the points in are from the same room, and then we need to use reference point matching to determine which one of the floor plan this room corresponds to.

(3) Reference point matching

The inventive example utilizes the door of the room as a key reference point for establishing the connection between the stress-free floor plan and the fingerprint space. We define as follows

so, It can be regarded as the point closest to the door inside and outside the door. definition All fingerprints in F _D can be organized into a chain in the minimum spanning tree T, so that we can express F _D as a vector form F _D =(f ₁ ,f ₂ ,...,f _k ). In the stress-free plan view, define P _D =(p ₁ ,p ₂ ,...,p _k ) as the sampling point closest to each door in the corridor, and the order in which points appear in P _D is along the corridor from one side to the other side. In this way, the mapping from F _D to _PD has only two situations: corresponding or just opposite. Define l=(l ₁ ,l ₂ ,...,l _k-1 ), l _i =||p _i+1 -p _i ||, l'=(l' ₁ ,l' ₂ ,... ,l' _k-1 ), l _i =||p′ _i+1 -p′ _i ||. Define the similarity between l and l' as s ₁ , and the calculation method is

$\frac{\mathrm{<span\; class="notranslate">\; l</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; l</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; l</span>}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; l</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}$

Use the same method to calculate the similarity of the opposite vectors of l and l', denoted as s ₂ . If s ₁ ≥ s ₂ , it is considered that the points from F _D to _PD are corresponding; otherwise, the points from F _D to _PD are just opposite.

S33. Space coordinate conversion

Divided into two steps: conversion at the floor level and conversion at the room level

(1) Conversion at the floor level

As can be seen from the stress-free plan view and the visualization of the fingerprint space, their structures are very similar, but with some minor changes such as translation, rotation, and reflection. We utilize a transformation matrix to achieve floor-level matching.

Assume that the coordinates of a fingerprint in F _D are x _i =[ _xi ¹ x _i ² ... x _i ^d ] ^T , and d is the dimension of the fingerprint space. The coordinates of the corresponding position y _i =[y _i ¹ y _i ² ... y _i ^d ] ^T . Define A as a d×d transformation matrix, B=[b ₁ b ₂ ...b _d ] ^T . We have the equation k=|F _D |

y _i =Ax _i +B

We rewrite these k equations as

H _i z=G _i

Where H _i =[x _i ^T 1], z=[AB] ^T , G _i =y _i ^T . Combining these k equations, we have

Hz=G

H _i and G _i are the ith rows of H and G.

It can be obtained by the method of least squares

$\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; z</span>}}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; h</span>}}^{\mathrm{<span\; class="notranslate">\; T</span>}}\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; )</span>}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; h</span>}}^{\mathrm{<span\; class="notranslate">\; T</span>}}\mathrm{<span\; class="notranslate">\; G</span>}$

In this way, we can get A and B. For a fingerprint whose coordinates are x=[x _i ¹ x _i ² ... x _i ^d ] ^T , the sampling point closest to Ax+B can be regarded as its actual position.

(2) Room-level conversion

As mentioned above, the example of the present invention classifies fingerprints into different fingerprint rooms through clustering, and obtains the one-to-one correspondence between the fingerprint space and the physical space on the plan through matching of reference points. Using MDS again, the fingerprints from the same room are transformed into a d-dimensional space. Similarly, the sampling points corresponding to the room form a d-dimensional stress-free plan. Taking the nearest and farthest points from the room door as reference points, the transformation relationship between the fingerprint room and the physical room can be determined. By repeating this step for all rooms, we achieve room-level transformations.

To sum up, the wireless indoor positioning method of the present invention does not require manual on-site survey of the positioning area, and multiple users in the system can conveniently provide fingerprint information collaboratively. Redundant information generated during normal movement can be used as upgrade information.

It should be noted that any process or method descriptions described in flowcharts or otherwise described herein can be understood as representing codes including one or more steps of executable instructions for implementing specific logical functions or processes. modules, segments or parts, and the scope of the preferred embodiments of the present invention includes further implementations, which may be performed out of the order shown or discussed, including in substantially simultaneous fashion or in reverse order depending on the functions involved. function, which should be understood by those skilled in the art to which the embodiments of the present invention belong.

In the description of this specification, descriptions referring to the terms "one embodiment", "some embodiments", "example", "specific examples", or "some examples" mean that specific features described in connection with the embodiment or example , structure, material or characteristic is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be construed as limitations to the present invention. Variations, modifications, substitutions, and modifications to the above-described embodiments are possible within the scope of the present invention.

Claims (6)

1. a wireless indoor location method, is characterized in that, comprises the following steps:

S1. utilize smart mobile phone automatically to gather wireless fingerprint data and user's Mobile data, formed fingerprint collection F and Distance matrix D ', and carry out preliminary treatment;

S2. according to pretreated described fingerprint collection F and Distance matrix D ', build fingerprint space, wherein, described S2 comprises: according to pretreated described fingerprint collection F and Distance matrix D ', the Euclidean space all fingerprint map tieed up to a d by MDS algorithm;

S3. the Euclidean space tieed up according to described d generates unstressed plane graph, carries out key feature extraction, and carries out space coordinate conversion,

Wherein, described key feature extracts and comprises corridor recognition, room identification and reference point coupling, described space coordinate conversion comprises the conversion of floor-level and the conversion of room-level, described reference point coupling comprises: utilize the door in room as setting up the critical reference points contacted between unstressed plane graph and fingerprint space, wherein, first define with for inside and outside room door from the point that door is nearest, then define room door fingerprint collection described F _din all fingerprints can be organized into a chain in minimum spanning tree T, therefore by described F _dbe expressed as vector form F _d=(f ₁, f ₂..., f _k), and in unstressed plane graph, definition P _d=(p ₁, p ₂..., p _k) on corridor from each nearest sampled point, described in each, sampled point is at P _dthe order of middle appearance is along corridor from side to opposite side, due to from F _dto P _dbe mapped as corresponding or just reverse order is corresponding, achieve reference point coupling and determine the corresponding of certain room fingerprint collection and certain room in plane graph in unstressed plane graph.

2. wireless indoor location method as claimed in claim 1, it is characterized in that, described S1 comprises further:

S11. the signal of wireless network and the reading of acceleration transducer and direction sensor is gathered by smart mobile phone, suppose there be m wireless network access point in region, then described smart mobile phone RSS fingerprint obtained of certain position in region is designated as the vector f=(s of a m dimension ₁, s ₂..., s _m), wherein s _irepresent the RSS value of i-th wireless network access point, define d ' again _ijfor f _iand f _jbetween distance, be the step number that user walks between the two positions, after the fingerprint-collection stage terminates, obtain fingerprint collection F={f _i, i=1 ... n}, wherein n is the number of fingerprint, and Distance matrix D '=[d ' _ij];

S12. preliminary treatment is carried out, to two fingerprint f to described fingerprint collection F _i=(s ₁, s ₂..., s _m) and f _j=(t ₁, t ₂..., t _m), definition f _iand f _jdiversity factor is if δ _ijbe less than predetermined threshold, then f _iand f _jin the generative process of fingerprint space, be taken as identical point, otherwise be taken as different points;

S13. to described Distance matrix D ' carry out preliminary treatment, the beeline between often pair of fingerprint is calculated according to shortest path first, if namely at f _iand f _jcertain intermediate node of existence f _k, meet d ' _ij>d ' _ik+ d ' _kj, then d ' _ijbe updated to d ' _ik+ d ' _kj.

3. indoor orientation method as claimed in claim 1, it is characterized in that, described corridor recognition comprises: utilize the fingerprint on degree acquisition corridor, centre, wherein, according to the distance between fingerprint, set up minimum spanning tree T, calculate the middle degree of each point on described minimum spanning tree, the part that centre degree is high is considered as the fingerprint on corridor, is designated as corridor fingerprint collection F _c, wherein said middle degree B (v) is defined as: the figure G=(V, E) be made up of vertex set V and limit collection E for, wherein, σ _stfor the shortest path number from s to t, σ _stv () is from s to t and through the shortest path number of v.

4. indoor orientation method as claimed in claim 1, it is characterized in that, the identification of described room comprises: utilize cluster to obtain room fingerprint collection, wherein, remove described corridor fingerprint collection F in all fingerprint spaces _c, adopt K-means algorithm to remaining fingerprint F-F _ccarry out cluster, obtain individual the trooping of k k and (be designated as i=1,2 ..., k), think same after cluster in institute a little from same room with them.

5. indoor orientation method as claimed in claim 1, it is characterized in that, the conversion of described floor-level comprises: utilize a transformation matrix to realize the conversion of the floor-level of unstressed plane graph and fingerprint spatial visualization figure, wherein, suppose at F _din the coordinate of a fingerprint be x _i=[x _i ¹x _i ²x _i ^d] ^t, wherein d is the dimension in fingerprint space, the coordinate y of corresponding position _i=[y _i ¹y _i ²y _i ^d] ^t, definition A is the transformation matrix of d × d, B=[b ₁b ₂b _d] ^t, owing to there being k=|F _d| bar equation y _i=Ax _i+ B, is rewritten as H by described k bar equation _iz=G _i, wherein H _i=[x _i ^t1], z=[AB] ^t, G _i=y _i ^t, combine described k bar equation, have Hz=G, wherein H _iand G _ifor i-th row of H and G, solved by least square method try to achieve A and B, thus be x=[x for coordinate _i ¹x _i ²x _i ^d] ^ta fingerprint, can be considered its physical location from the sampled point that Ax+B is nearest.

6. indoor orientation method as claimed in claim 1, it is characterized in that, the conversion of described room-level comprises: utilize MDS algorithm the fingerprint from same room with them to be transformed to the space of a d dimension further, the sampled point in corresponding room is made to form the unstressed plane graph of a d dimension, by as a reference point with point farthest recently from room door, the transformation relation between fingerprint room and physical room can be determined, by repeating aforesaid operations one by one to all rooms, achieve the conversion of the unstressed plane graph in each room and the room-level of fingerprint spatial visualization figure.