CN109212464B - Method and equipment for estimating terminal distance and position planning - Google Patents
Method and equipment for estimating terminal distance and position planning Download PDFInfo
- Publication number
- CN109212464B CN109212464B CN201811138609.XA CN201811138609A CN109212464B CN 109212464 B CN109212464 B CN 109212464B CN 201811138609 A CN201811138609 A CN 201811138609A CN 109212464 B CN109212464 B CN 109212464B
- Authority
- CN
- China
- Prior art keywords
- distance
- wireless access
- access point
- terminal
- terminals
- 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
Images
Classifications
-
- 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
- G01S1/00—Beacons or beacon systems transmitting signals having a characteristic or characteristics capable of being detected by non-directional receivers and defining directions, positions, or position lines fixed relatively to the beacon transmitters; Receivers co-operating therewith
- G01S1/02—Beacons or beacon systems transmitting signals having a characteristic or characteristics capable of being detected by non-directional receivers and defining directions, positions, or position lines fixed relatively to the beacon transmitters; Receivers co-operating therewith using radio waves
- G01S1/08—Systems for determining direction or position line
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Position Fixing By Use Of Radio Waves (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The present application is a divisional application entitled method, terminal and apparatus for estimating terminal spacing and location planning, having application number 201610395410.X, the filing date of the parent application (201610395410.X) is 2016, 6/year. The invention provides a method, equipment and a terminal for estimating terminal distance and position planning, wherein the method mainly comprises the steps of acquiring wireless access point information acquired by a terminal with the distance to be estimated; extracting a characteristic vector of the wireless access point information, and acquiring the distance between any two terminals in the terminals of the distance to be estimated according to a distance calculation function; and on the basis of the distance, obtaining the position planning among the terminals by adopting a dimension reduction processing mode, and obtaining the planning position coordinates of each terminal. The method accurately obtains the relative distance between the terminals and the terminal position with less resource consumption and higher speed, has low requirement on the hardware accuracy of the terminals, and can be widely applied to the existing user terminals.
Description
The present application is a divisional application entitled method, terminal and apparatus for estimating terminal spacing and location planning, having application number 201610395410.X, the filing date of the parent application (201610395410.X) is 2016, 6/year.
Technical Field
The present invention relates to the field of wireless communication, and in particular, to a method, a device, and a terminal for estimating a distance between terminals without using a third-party server.
Background
With the rapid development of intelligent terminal devices and wireless internet technologies, Location Based Services (LBS) and problems related to location acquisition are widely used. Currently, satellite positioning technology such as Global Positioning System (GPS) and Beidou positioning system is mostly adopted for outdoor positioning, and when the satellite connection condition is good, the positioning accuracy can reach within 1 m. But due to building shielding, satellite positioning technology is not applicable in the field of indoor positioning, and the initial indoor positioning technology comprises wireless positioning technology based on infrared, ultrasonic wave and RFID signals; the current application is that the indoor position is estimated based on the fingerprint matching algorithm of wireless network signals such as Bluetooth and Wi-Fi, such as the indoor positioning method which is provided by Liudingjun and combines Wi-Fi and sensing data. The average indoor positioning accuracy can be improved to 3-5 m by an indoor positioning algorithm based on technologies such as fingerprint matching and the like. However, as LBS develops, more ways are provided for establishing social relationships between people, and the emerging near-field social concepts gradually move into the field of view of people. APP such as presence, Mixin and the like all provide a new social contact mode based on a near-field social concept, and strangers in the same area can communicate and know with each other more quickly and naturally. Therefore, it is important to accurately obtain the position relationship between persons in a specific area.
In the prior art, for example, differences of AP information in a surrounding environment obtained by a terminal at different positions are used, and a current position of a user is calculated based on signal arrival Time (TOA) or signal arrival Time Difference (TDOA) and Received Signal Strength (RSSI), for example, absolute time synchronization between the AP and the user terminal is required, so that a requirement on accuracy of the device is high, and devices such as a smart phone and a smart watch, which are commonly used by users in a market, often cannot meet the accuracy requirement, or the consumption of system resources or traffic resources is severe when the requirement is barely met.
In addition, in the patent document CN104459612A, the name of the invention is a mobile terminal with energy supply for measuring the distance and direction to the WI-FI device, the distance and direction of the mobile terminal are measured by a distance measuring antenna and a phase discrimination distance measuring module. The method is characterized in that two sides needing distance measurement are limited in the range which can be sensed by the sensor, and point-to-point direct sensing is realized through the sensor module. In CN104812061A, the MIMO-OFDM channel state is utilized, and anchor AP information of known location in the positioning space is combined, and the path from different anchor APs to the terminal is calculated, and then the final user positioning is obtained by calculation.
The above prior art has at least the following common drawbacks: (1) positioning based on map information often requires receiving additional information such as GPS information to acquire the terminal position; (2) the positioning technology based on wifi signals usually needs a large amount of calculation and measurement to determine the relative position between terminals, and the calculation amount needs to be set at a server end for processing and pre-storing comparison information to achieve real-time positioning, and because calibration work is complicated and calibration data cannot be shared between buildings, manpower and material resources are consumed greatly; (3) because the complex and changeable indoor environment causes large loss to the signal propagation and the multipath effect exists in the signal propagation process, the positioning result precision of the method is not high; (4) the method can not adapt to the environment of the changeable indoor signal source well, and the larger the time span is, the larger the signal source difference is.
Disclosure of Invention
In view of the above, to solve the problems in the prior art, the present invention first provides a method for estimating a terminal distance, including:
acquiring wireless access point information acquired by a terminal with a distance to be estimated;
extracting a characteristic vector of the wireless access point information, and acquiring the distance between any two terminals in the terminals of the distance to be estimated according to a distance calculation function; the distance calculation function may be a function obtained in various ways, such as a distance function obtained by fitting on the basis of an empirical value obtained by detecting a fixed point in the positioning region;
the wireless access point information includes at least: MAC information and received signal strength of the wireless access point;
the distance calculation function is obtained by machine learning. The machine learning method can be realized by adopting a conventional artificial neural network method, such as a BP neural network, and the like, and can also be realized by adopting algorithms such as a support vector machine and the like.
Preferably, the machine learning further comprises:
acquiring wireless access point information at different positions in a positioning area;
establishing a relation between the position and the wireless access point information collected at the position to form different position data;
grouping the position data pairwise, removing repeated groups, and performing feature extraction on each group of the position data;
and obtaining a distance calculation function at least partially based on the extracted feature vectors in a machine learning mode.
Preferably, the feature vector includes at least one of: the number of wireless access points, the signal difference of the same wireless access point, the ratio of the number of the same wireless access points to the number of the total wireless access points, and the like. These feature vectors may be combined arbitrarily, or combined with other feature vectors, according to specific machine learning needs, accuracy needs, and the like.
Preferably, after feature extraction is performed on each group of the position data, a distance between two points in the group is calculated as a label of the extracted feature vector;
forming feature data based on the label and the feature vector;
based on the feature data, a distance calculation function is obtained.
Preferably, after the feature data is formed, the feature data is normalized. The normalization processing is not necessary, and when the magnitude difference of the feature vectors is not large, the step of normalization processing may not be added, or adjustment may be performed according to the specific calculation requirement.
Preferably, the machine learning uses a support vector machine model, and the support vector machine kernel function uses a radial basis function;
the support vector machine model adopts a support vector regression classifier;
and the machine learning process adopts a gradient descent method to search the optimal regression parameters.
In another aspect, the present invention further provides a near field terminal position planning method, including:
acquiring wireless access point information acquired by a terminal with a distance to be estimated;
extracting a characteristic vector of the wireless access point information, and acquiring the distance between any two terminals in the terminals of the distance to be estimated according to a distance calculation function;
converting the terminal with the distance to be estimated into two-dimensional position distribution according to the distance to obtain a planning position of the terminal;
the wireless access point information includes at least: MAC information of the wireless access point, received signal strength.
Preferably, wireless access point information at different positions in a positioning area is collected;
establishing a relation between the position and the wireless access point information collected at the position to form different position data;
grouping the position data pairwise, removing repeated grouping, performing feature extraction on each group of position data, and calculating the distance between two points in the group as a label of the extracted feature vector;
forming feature data based on the label and the feature vector;
and obtaining a distance calculation function at least partially based on the characteristic data in a machine learning mode. The machine learning method can be realized by adopting a conventional artificial neural network method, such as a BP neural network, and the like, and can also be realized by adopting algorithms such as a support vector machine and the like.
Preferably, the obtaining the planned position of the terminal with the distance to be estimated further includes:
according to the distance between any two terminals, a data set I is formed, and a distance matrix between the terminals is established based on the data set, wherein the distance matrix can be expressed as:
wherein d isi,jRepresents the spacing of the ith and jth variables in the dataset, I, j e 1.
Further preferably, the distance may be an euclidean distance between arbitrary terminals.
Preferably, the eigenvalue decomposition is performed on the distance matrix, and the specific method is as follows:
construct a matrix X, T, let
Then, from the above equation:
wherein, XiIs RNN is a spatial dimension and is not less than 1 and not more than N at the ith coordinate point in the space;
and (3) carrying out matrix decomposition on the T of the matrix:
wherein U is a characteristic vector, and Λ is a characteristic value matrix;
order:
and finishing the dimension reduction processing of the distance matrix.
Preferably, the average distance between the terminals of the distance to be estimated is obtained according to the coordinates of the planned position and the actual position coordinates of the corresponding terminals;
and obtaining the evaluation parameters of the planning position based on the average distance and the coordinate distance between the two farthest positions in the coordinates of the planning position.
Preferably, the evaluation parameter is calculated as follows:
wherein D ismaxThe coordinate distance between the two farthest positions in the coordinates of the planning positions; dmeanIs the average distance. When the beta isMDSThe closer to 0, the better the fit of the planned position to the real position.
In yet another aspect of the present invention, the present invention further provides a terminal capable of estimating a terminal distance, wherein the terminal includes:
a request receiving unit, configured to send and/or receive a distance acquisition request;
the wireless access point information acquisition unit is used for acquiring the wireless access point information acquired by the terminal and receiving the acquired wireless access point information sent by other terminals;
the distance acquisition unit is used for extracting the characteristic vectors of the terminal information of the terminal and other terminals after the terminal receives the terminal information sent by other terminals, and acquiring the relative distance between the terminal and other terminals through a distance calculation function;
preferably, the wireless access point information includes at least: MAC information of the wireless access point, received signal strength.
Preferably, the feature vector comprises at least one of: the wireless access point number, the signal difference of the same wireless access point, and the ratio of the same wireless access point to the total wireless access point.
Preferably, the terminal further comprises a storage unit for storing the distance calculation function.
Preferably, the distance calculation function is obtained by means of machine learning;
further preferably, the machine learning comprises:
acquiring wireless access point information at different positions in a positioning area;
establishing a relation between the position and the wireless access point information collected at the position to form different position data;
grouping the position data pairwise, removing repeated groups, and performing feature extraction on each group of the position data;
and obtaining a distance calculation function at least partially based on the extracted feature vectors in a machine learning mode.
Preferably, after feature extraction is performed on each set of the position data, the method further includes:
calculating the distance between two points in the group as a label of the extracted feature vector;
forming feature data based on the label and the feature vector;
based on the feature data, a distance calculation function is obtained.
Preferably, the terminal further includes:
the position planning unit is used for obtaining the planned position of the terminal according to the distance between any two terminals in the terminal and the other terminals;
the planned position is a position in a two-dimensional space.
Preferably, the position planning unit further includes a dimension reduction processing unit, configured to establish a distance matrix according to a distance between any two terminals, and perform dimension reduction processing on the distance matrix to obtain the coordinates of the planned position.
Preferably, the dimension reduction process employs a eigenvalue decomposition.
In another aspect of the present invention, the present invention further provides an apparatus for acquiring a planned position of a terminal, where the apparatus includes:
the information acquisition unit is used for acquiring wireless access point information acquired by a terminal with a distance to be estimated;
the characteristic extraction unit is used for extracting a characteristic vector of the wireless access point information;
the distance calculation unit is used for acquiring the distance between any two terminals in the terminals of the distance to be estimated according to a distance calculation function;
the position planning unit is used for obtaining the planning position of the terminal according to the distance calculated by the distance calculation unit; the planned position is a position in a two-dimensional space;
the wireless access point information includes at least: MAC information of the wireless access point, received signal strength.
Preferably, the device further includes a distance function obtaining unit, configured to obtain the distance calculation function, specifically in the following manner:
acquiring wireless access point information at different positions in a positioning area;
establishing a relation between the position and the wireless access point information collected at the position to form different position data;
grouping the position data pairwise, removing repeated grouping, performing feature extraction on each group of position data, and calculating the distance between two points in the group as a label of the extracted feature vector;
forming feature data based on the label and the feature vector;
and obtaining a distance calculation function at least partially based on the characteristic data in a machine learning mode.
Preferably, the position planning unit forms a data set according to the distance between any two terminals obtained by the distance calculation unit, and establishes a distance matrix between the terminals based on the data set; and
and performing dimension reduction processing on the distance matrix to obtain a planning position coordinate of the terminal.
Preferably, the device further comprises a planning position evaluation unit, configured to obtain an average distance between the terminals of the distance to be estimated according to the coordinates of the planning position and the actual position coordinates of the corresponding terminals;
and obtaining the evaluation parameters of the planning position based on the average distance and the coordinate distance between the two farthest positions in the coordinates of the planning position.
Preferably, the evaluation parameter is calculated as follows:
wherein D ismaxThe coordinate distance between the two farthest positions in the coordinates of the planning positions; dmeanIs the average distance;
βMDSthe closer to 0, the better the fit of the planned position to the real position.
Compared with the prior art, the technical scheme of the invention does not need additional third-party positioning information, can obtain the estimation of the distance between the terminals and the determination of the position of the terminal only by the wireless access point information around the terminal, does not need accurate consistent time between the terminals, has low requirement on the performance of the terminal, and can be widely applied to the existing user terminals.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flowchart illustrating a method for estimating a terminal distance according to an embodiment of the present invention;
fig. 2 is a flowchart of a near field terminal position planning method according to an embodiment of the present invention;
FIG. 3 is a diagram of a terminal structure according to an embodiment of the present invention;
FIG. 4 is a data path acquisition diagram according to an embodiment of the present invention;
FIG. 5 is a distribution of collection points for experiment 1 in accordance with one embodiment of the present invention;
FIG. 6 is a comparison of the distance estimation results of experiment 1 according to an embodiment of the present invention;
fig. 7 is a comparison graph of the position planning result of experiment 1 according to an embodiment of the present invention.
Detailed Description
An application program recommendation method and apparatus according to an embodiment of the present invention are described in detail below with reference to the accompanying drawings. It should be understood that the described embodiments are only some embodiments of the invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It will be appreciated by those of skill in the art that the following specific examples or embodiments are a series of presently preferred arrangements of the invention to further explain the principles of the invention, and that such arrangements may be used in conjunction or association with one another, unless it is expressly stated that some or all of the specific examples or embodiments are not in association or association with other examples or embodiments. Meanwhile, the following specific examples or embodiments are only provided as an optimized arrangement mode and are not to be understood as limiting the protection scope of the present invention.
Example 1:
in a specific embodiment, as shown in fig. 1, the present invention provides a method of estimating terminal separation, the method comprising:
acquiring wireless access point information acquired by a terminal with a distance to be estimated;
extracting a characteristic vector of the wireless access point information, and acquiring the distance between any two terminals in the terminals of the distance to be estimated according to a distance calculation function; the distance calculation function may be a function obtained in various ways, such as a distance function obtained by fitting on the basis of an empirical value obtained by detecting a fixed point in the positioning region;
the wireless access point information includes at least: MAC information and received signal strength of the wireless access point;
the distance calculation function is obtained by machine learning. The machine learning method can be realized by adopting a conventional artificial neural network method, such as a BP neural network, and the like, and can also be realized by adopting algorithms such as a support vector machine and the like.
In a specific embodiment, the machine learning further comprises:
acquiring wireless access point information at different positions in a positioning area;
establishing a relation between the position and the wireless access point information collected at the position to form different position data;
grouping the position data pairwise, removing repeated groups, and performing feature extraction on each group of the position data;
and obtaining a distance calculation function at least partially based on the extracted feature vectors in a machine learning mode.
Specifically, in the data acquisition process, wireless Access Point (AP) information is scanned at different positions in a positioning area, AP data at different positions in the area is acquired, a relationship between the position and the AP is established, and a data format may be set as follows:
＜X,Y,MAC1,RSSI1,...,MACn,RSSIn＞
x, Y are position coordinates, MACiIs APiMAC address, RSSIiN is the number of APs searched for the location for the corresponding received signal strength.
In a specific embodiment, the feature vector includes at least one of the following: the number of wireless access points, the signal difference of the same wireless access point, the ratio of the number of the same wireless access points to the number of the total wireless access points, and the like. These feature vectors may be combined arbitrarily, or combined with other feature vectors, according to specific machine learning needs, accuracy needs, and the like.
In a specific embodiment, when performing machine learning and performing example calculation on a terminal, the feature extraction may be performed as follows:
and grouping the acquired data of different positions pairwise in a handshake mode, namely if the Wi-Fi signal data of n points are acquired, grouping each point and other n-1 points for 1 time respectively, and removing repeated groups to obtain n x (n-1)/2 groups of data. Performing feature extraction on each group of data, and calculating the same AP Number (NUM) between each group of datasameAP) Maximum value of same AP Signal Difference (RSSID)max) Minimum value of same AP Signal Difference (RSSID)min) Average of same AP signal differences (RSSID)mean) Ratio of the number of the same AP to the total number of APs in the group (NUM)sameAP/NUMallAP) As a feature, the coordinate distance D of two points in the group is calculated as a data tag,
establishing the characteristic data, the format can be set as follows:
＜D 1：NUMsameAP2：RSSIDmax3：RSSIDmin4：RSSIDmean5:NUMsameAP/NUMallAP＞
in order to avoid the problem of overlarge dimension difference of the feature data, normalization processing can be carried out on the data, the feature value is normalized to the range of [ -1,1], and a data file for machine learning training is generated.
In a specific embodiment, the machine learning employs a support vector machine model, and the support vector machine kernel function employs a radial basis function;
the support vector machine model adopts a support vector regression classifier;
and the machine learning process adopts a gradient descent method to search the optimal regression parameters.
In a specific embodiment, in the specific distance calculation, the regression model generated as described above, i.e., the distance calculation function, is used to perform regression on the test data, the distance between 2 points in each group is predicted from each group of feature data, and the experimental result is represented by the pearson correlation coefficient (ρ) between the regression value and the true valuepearson) And average Error (Error)Mean) Evaluation was carried out.
Example 2:
in a specific embodiment, as shown in fig. 2, the present invention further provides a near field terminal position planning method, including:
acquiring wireless access point information acquired by a terminal with a distance to be estimated;
extracting a characteristic vector of the wireless access point information, and acquiring the distance between any two terminals in the terminals of the distance to be estimated according to a distance calculation function;
converting the terminal with the distance to be estimated into two-dimensional position distribution according to the distance to obtain a planning position of the terminal;
the wireless access point information includes at least: MAC information of the wireless access point, received signal strength.
In a specific embodiment, wireless access point information at different positions in a positioning area is collected;
establishing a relation between the position and the wireless access point information collected at the position to form different position data;
grouping the position data pairwise, removing repeated grouping, performing feature extraction on each group of position data, and calculating the distance between two points in the group as a label of the extracted feature vector;
forming feature data based on the label and the feature vector;
and obtaining a distance calculation function at least partially based on the characteristic data in a machine learning mode.
In a specific embodiment, the above-described machine learning and feature extraction method can be performed as in embodiment 1.
The machine learning method can be realized by adopting a conventional artificial neural network method, such as a BP neural network, and the like, and can also be realized by adopting algorithms such as a support vector machine and the like.
In a specific embodiment, the obtaining a planned position of the terminal with the distance to be estimated further includes:
according to the distance between any two terminals, a data set I is formed, and a distance matrix between the terminals is established based on the data set, wherein the distance matrix can be expressed as:
wherein d isi,jRepresents the spacing of the ith and jth variables in the dataset, I, j e 1.
The objective of the multidimensional analysis is to obtain a set of vectors x of size I1,...,xI∈RNFor all I, j ∈ 1i-xj||≈di,jAnd | | · | | represents the vector modulo. The vector norm may be the euclidean distance between variables, but in a broad sense it may also refer to an arbitrary distance function. In multidimensional analysis, the essence is to search a data set I to R on the basis of keeping the relative distance between variables unchangedNThe mapping relationship between them. If dimension N is chosen to be 2 or 3, vector xiNamely, the structural relationship of each variable in the data set I can be reflected in a two-dimensional plane or a three-dimensional space. Finally, the multidimensional analysis described above can be converted into calculations
And the optimization problem can be solved by adopting a matrix eigenvalue solution.
In a specific embodiment, the eigenvalue decomposition is performed on the distance matrix by the following specific method:
construct a matrix X, T, let
Then, from the above equation:
namely, it is
Wherein, XiIs RNN is a spatial dimension and is not less than 1 and not more than N at the ith coordinate point in the space;
and (3) carrying out matrix decomposition on the T of the matrix:
wherein U is a characteristic vector, and Λ is a characteristic value matrix;
order:
and finishing the dimension reduction processing of the distance matrix.
In a specific implementation manner, according to the coordinates of the planned position and the actual position coordinates of the corresponding terminal, the average distance between the terminals of the distance to be estimated is obtained;
and obtaining the evaluation parameters of the planning position based on the average distance and the coordinate distance between the two farthest positions in the coordinates of the planning position.
Calculating the average distance D between the coordinates of the planned point and the corresponding coordinates of the actual pointmeanIs solved by the following formula, wherein (x)i_pre,yi_pre) To predict point coordinates, (x)i_real,yi_real) The coordinate of the actual point is shown, and n is the number of the experimental points;
calculating the distance D between two points with the farthest distance in the predicted point setmaxBy mean distance DmeanAt a maximum distance DmaxRatio of (beta)MDSAs a parameter for evaluating the results of the MDS location planning,
in a specific embodiment, the evaluation parameter is calculated as follows:
wherein D ismaxThe coordinate distance between the two farthest positions in the coordinates of the planning positions; dmeanIs the average distance. When the beta isMDSThe closer to 0, the better the fit of the planned position to the real position.
Example 3:
in yet another aspect of the present invention, as shown in fig. 3, the present invention provides a terminal capable of estimating a terminal pitch, the terminal comprising:
a request receiving unit, configured to send and/or receive a distance acquisition request;
the wireless access point information acquisition unit is used for acquiring the wireless access point information acquired by the terminal and receiving the acquired wireless access point information sent by other terminals;
the distance acquisition unit is used for extracting the characteristic vectors of the terminal information of the terminal and other terminals after the terminal receives the terminal information sent by other terminals, and acquiring the relative distance between the terminal and other terminals through a distance calculation function;
preferably, the wireless access point information includes at least: MAC information of the wireless access point, received signal strength.
In a specific embodiment, the feature vector includes at least one of: the wireless access point number, the signal difference of the same wireless access point, and the ratio of the same wireless access point to the total wireless access point. The specific arrangement of the feature vector may be as in embodiment 1.
In a specific embodiment, the terminal further comprises a storage unit for storing the distance calculation function.
In a specific embodiment, the distance calculation function is obtained by machine learning;
further preferably, the machine learning comprises:
acquiring wireless access point information at different positions in a positioning area;
establishing a relation between the position and the wireless access point information collected at the position to form different position data;
grouping the position data pairwise, removing repeated groups, and performing feature extraction on each group of the position data;
and obtaining a distance calculation function at least partially based on the extracted feature vectors in a machine learning mode.
In a specific embodiment, after performing feature extraction on each set of the position data, the method further includes:
calculating the distance between two points in the group as a label of the extracted feature vector;
forming feature data based on the label and the feature vector;
based on the feature data, a distance calculation function is obtained.
The above-mentioned machine learning and distance function obtaining method can adopt the specific mode in embodiment 1.
In a specific embodiment, the terminal may also have a function of location planning, that is, it further includes:
the position planning unit is used for obtaining the planned position of the terminal according to the distance between any two terminals in the terminal and the other terminals;
the planned position is a position in a two-dimensional space.
In a specific embodiment, the position planning unit further includes a dimension reduction processing unit, configured to establish a distance matrix according to a distance between any two terminals, and perform dimension reduction processing on the distance matrix to obtain the coordinates of the planned position.
Preferably, the dimension reduction process employs a eigenvalue decomposition.
The specific location planning method and principle described above may be implemented in the specific manner in embodiment 2.
Example 4:
in another aspect of the present invention, the present invention further provides an apparatus for acquiring a planned position of a terminal, where the apparatus includes:
the information acquisition unit is used for acquiring wireless access point information acquired by a terminal with a distance to be estimated;
the characteristic extraction unit is used for extracting a characteristic vector of the wireless access point information;
the distance calculation unit is used for acquiring the distance between any two terminals in the terminals of the distance to be estimated according to a distance calculation function;
the position planning unit is used for obtaining the planning position of the terminal according to the distance calculated by the distance calculation unit; the planned position is a position in a two-dimensional space;
the wireless access point information includes at least: MAC information of the wireless access point, received signal strength.
In a specific embodiment, the apparatus further includes a distance function obtaining unit, configured to obtain the distance calculation function, specifically by:
acquiring wireless access point information at different positions in a positioning area;
establishing a relation between the position and the wireless access point information collected at the position to form different position data;
grouping the position data pairwise, removing repeated grouping, performing feature extraction on each group of position data, and calculating the distance between two points in the group as a label of the extracted feature vector;
forming feature data based on the label and the feature vector;
and obtaining a distance calculation function at least partially based on the characteristic data in a machine learning mode.
In a specific embodiment, the position planning unit forms a data set according to the distance between any two terminals obtained by the distance calculation unit, and establishes a distance matrix between the terminals based on the data set; and
and performing dimension reduction processing on the distance matrix to obtain a planning position coordinate of the terminal.
In a specific embodiment, the apparatus further includes a planning position evaluation unit, configured to calculate an average distance between the terminals with the distance to be estimated according to coordinates of the planning position and actual position coordinates of the terminals corresponding to the planning position;
and obtaining the evaluation parameters of the planning position based on the average distance and the coordinate distance between the two farthest positions in the coordinates of the planning position.
In a specific embodiment, the evaluation parameter is calculated as follows:
wherein D ismaxThe coordinate distance between the two farthest positions in the coordinates of the planning positions; dmeanIs the average distance;
βMDSthe closer to 0, the better the fit of the planned position to the real position.
Example 5:
in order to further explain the optimized technical solution of the present invention, in this embodiment, a specific example is combined to describe the application process and effect of the technical solution of the present invention.
The data used in this example is Wi-Fi data from one floor of a shopping mall in a commercial area in Shanghai, with a total acquisition area of 114233 pixels squared (about 12600 square meters), for 1367 data points. The data acquisition path is shown in fig. 4, where the acquisition path is marked with a black line. The experimental data of the terminal distance calculating part in the experiment is obtained by performing feature extraction and data preprocessing process calculation on original fingerprint data, the distance unit in the experiment is pixel (px), and the SVR experiment is verified by adopting a leave-one-out method.
In this example, 7 sub-regions from the collection region were selected for specific experiments, and the experimental profiles are shown in the following table:
in public places such as shopping malls, modules such as pedestrian walkways (linear type), storefronts (rectangular type), and halls (area type) are basically configured, and in this example, experiment 1 of the above 7 experiments is used as a representative example to perform a concrete experimental analysis.
Experiment 1: in the experiment, a linear type acquisition area is selected, 50 pieces of Wi-Fi data fingerprint data, 45 pieces of training data and 5 pieces of test data are acquired in total, and the distribution of acquisition points is shown in figure 5.
After SVR training and regression, the prediction result pairs of 5 test points are shown in FIG. 6, and the SVR regression result rho in experiment 1pearson＝0.932，ErrorMeanWhen 4.578(px), the linear region regression results better, with an average error rate of less than 5%.
The distance prediction results of 5 points are subjected to dimension reduction by adopting multi-dimensional scale analysis, the position planning result is shown in figure 7, and betaMDSThe experimental planning results are good in the linear region when the value is 0.070.
The comparison and analysis of the experimental results can obtain that the dimension reduction effects of support vector regression and multidimensional scale analysis are reduced along with the increase of the area of the distribution area of the sampling points, but compared with the planning result of MDS, the prediction result of SVR keeps better stability and accuracy, and the accumulated error exists between the SVR step and the MDS step.
Meanwhile, the embodiment shows that in the planning range of 100 square meters, the correlation between the calculation result of the terminal distance and the actual distance is more than 90%, and the error is less than 10%; the accuracy of the phase position planning result is more than 80%. In the range of 300 square meters, the correlation between the calculation result of the terminal distance and the actual distance is more than 85%, and the error is less than 20%; the terminal position planning method can achieve more than 70% of calculation accuracy. Through verification, the mobile terminal distance calculation and position planning method based on the Wi-Fi signal characteristics has a good overall effect. Compared with other indoor positioning methods, the method has the advantages of high calculation speed, no limitation of the position of the acquisition point on the prediction result and strong usability in a near field area.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.
The above description is only for the specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (6)
1. A method for estimating terminal separation, comprising:
acquiring wireless access point information acquired by a terminal with a distance to be estimated;
extracting a characteristic vector of the wireless access point information, and acquiring the distance between any two terminals in the terminals with the distance to be estimated according to a distance calculation function;
the wireless access point information includes at least: MAC information and received signal strength of the wireless access point;
the distance calculation function is obtained by machine learning in the following way:
acquiring wireless access point information at different positions in a positioning area;
establishing a relation between the position and the wireless access point information collected at the position to form different position data;
grouping the position data pairwise, removing repeated groups, and extracting feature vectors of each group of position data;
obtaining a distance calculation function at least partially based on the extracted feature vectors in a machine learning manner;
the relationship between the position and the wireless access point information collected at the position is established, and the data format is as follows:
＜X,Y,MAC1,RSSI1,...,MACn,RSSIn＞
x, Y are position coordinates, MACiIs APiMAC address, RSSIiN is the number of wireless access points searched for the position for the corresponding received signal strength;
the feature extraction adopts the following mode:
grouping the collected data of different positions pairwise in a handshaking mode, performing feature extraction on each group of data, and calculating the coordinate distance D of two points in the group as a data tag:
establishing a feature vector, wherein the format is as follows:
<D：NUMsameAP:RSSIDmax:RSSIDmin：RSSIDmean:NUMsameAP/NUMallAP>
wherein (x)1,y1)、(x2、y2) As coordinates of two points in the group, NUMsameAPFor the same AP number, RSSID, between each set of datamaxMaximum value of the same AP signal difference, RSSIDminIs the minimum value of the same AP signal difference, RSSIDmeanIs the average of the differences of the signals of the same AP, NUMsameAP/NUMallAPIs the ratio of the number of the same AP to the total number of the APs in the group.
2. The method of claim 1, wherein:
the feature vector includes at least one of: the wireless access point number, the same wireless access point signal difference, and the ratio of the number of the same wireless access points to the total number of the wireless access points.
3. The method of claim 1, wherein: and after the feature vector is formed, normalizing the feature vector.
4. The method of claim 1, wherein:
the machine learning adopts a support vector machine model, and a kernel function of the support vector machine adopts a radial basis function;
the support vector machine model adopts a support vector regression classifier;
the machine learning process adopts a gradient descent method to search for the optimal regression parameters.
5. The method according to claim 1, wherein the distance calculation function obtained by machine learning is evaluated by a Pearson correlation coefficient between a regression value and a true value and an average error.
6. An apparatus for estimating the terminal separation, the apparatus comprising at least one processor coupled to at least one memory device, the memory device storing at least instructions for being invoked and executed by the processor, wherein the instructions are configured to perform the method according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811138609.XA CN109212464B (en) | 2016-06-06 | 2016-06-06 | Method and equipment for estimating terminal distance and position planning |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610395410.XA CN106093844B (en) | 2016-06-06 | 2016-06-06 | Estimate terminal room away from and position planning method, terminal and equipment |
| CN201811138609.XA CN109212464B (en) | 2016-06-06 | 2016-06-06 | Method and equipment for estimating terminal distance and position planning |
Related Parent Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610395410.XA Division CN106093844B (en) | 2016-06-06 | 2016-06-06 | Estimate terminal room away from and position planning method, terminal and equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109212464A CN109212464A (en) | 2019-01-15 |
| CN109212464B true CN109212464B (en) | 2020-01-14 |
Family
ID=57448620
Family Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610395410.XA Active CN106093844B (en) | 2016-06-06 | 2016-06-06 | Estimate terminal room away from and position planning method, terminal and equipment |
| CN201811138609.XA Active CN109212464B (en) | 2016-06-06 | 2016-06-06 | Method and equipment for estimating terminal distance and position planning |
Family Applications Before (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610395410.XA Active CN106093844B (en) | 2016-06-06 | 2016-06-06 | Estimate terminal room away from and position planning method, terminal and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (2) | CN106093844B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107907856B (en) * | 2017-10-24 | 2021-07-27 | 东南大学 | RFID positioning method and system based on virtual reference label |
| CN113365216B (en) * | 2021-04-19 | 2023-04-07 | 杭州羿腾科技有限公司 | Tracking early warning method, system and equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1715951A (en) * | 2004-06-29 | 2006-01-04 | 微软公司 | Proximity detection using wireless signal strenghts |
| CN102713955A (en) * | 2009-07-29 | 2012-10-03 | 肖普基克公司 | Method and system for presence detection |
| CN102960010A (en) * | 2010-07-07 | 2013-03-06 | 索尼公司 | Communication control device, communication control method, communication system, and communication device |
| CN104703281A (en) * | 2013-12-06 | 2015-06-10 | 华为技术有限公司 | Channel distributing method, channel distributing device and Wireless Local Area Network system |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102186242B (en) * | 2011-05-09 | 2014-04-16 | 江南大学 | Method for positioning mobile node of wireless sensor network in fixed area |
| CN102413564B (en) * | 2011-11-25 | 2014-07-23 | 北京工业大学 | Indoor positioning method based on BP neural network and improved centroid algorithm |
| WO2015027118A1 (en) * | 2013-08-22 | 2015-02-26 | Qualcomm Incorporated | Utilizing a reference signal for indoor positioning |
| CN104185275B (en) * | 2014-09-10 | 2017-11-17 | 北京航空航天大学 | A kind of indoor orientation method based on WLAN |
| CN104320759B (en) * | 2014-10-15 | 2018-08-03 | 上海交通大学 | Based on fixedly target indoor locating system fingerprint base construction method |
-
2016
- 2016-06-06 CN CN201610395410.XA patent/CN106093844B/en active Active
- 2016-06-06 CN CN201811138609.XA patent/CN109212464B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1715951A (en) * | 2004-06-29 | 2006-01-04 | 微软公司 | Proximity detection using wireless signal strenghts |
| CN102713955A (en) * | 2009-07-29 | 2012-10-03 | 肖普基克公司 | Method and system for presence detection |
| CN102960010A (en) * | 2010-07-07 | 2013-03-06 | 索尼公司 | Communication control device, communication control method, communication system, and communication device |
| CN104703281A (en) * | 2013-12-06 | 2015-06-10 | 华为技术有限公司 | Channel distributing method, channel distributing device and Wireless Local Area Network system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106093844A (en) | 2016-11-09 |
| CN106093844B (en) | 2019-03-12 |
| CN109212464A (en) | 2019-01-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhou et al. | Integrated statistical test of signal distributions and access point contributions for Wi-Fi indoor localization | |
| CN106793086B (en) | Indoor positioning method | |
| Campos et al. | Wi-Fi multi-floor indoor positioning considering architectural aspects and controlled computational complexity | |
| Zhang et al. | DeepPositioning: Intelligent fusion of pervasive magnetic field and WiFi fingerprinting for smartphone indoor localization via deep learning | |
| Elbakly et al. | A robust zero-calibration RF-based localization system for realistic environments | |
| WO2017185828A1 (en) | Fingerprint positioning method and apparatus | |
| KR101483998B1 (en) | Method for estimating indoor position based on wireless lan, sever and terminal | |
| CN103916954B (en) | Probabilistic Localization Methods and positioner based on WLAN | |
| WO2019062734A1 (en) | Indoor positioning method and device based on wi-fi hot spots | |
| CN105636201B (en) | Indoor orientation method based on sparse signal fingerprint database | |
| JP2015531053A (en) | System, method, and computer program for dynamically creating a radio map | |
| CN104661300B (en) | Localization method, device, system and mobile terminal | |
| CN112906750B (en) | Hyperspectral image-based material analysis method and system | |
| Yang et al. | Research on Wi-Fi indoor positioning in a smart exhibition hall based on received signal strength indication | |
| Gao et al. | Employing spatial analysis in indoor positioning and tracking using Wi-Fi access points | |
| Li et al. | Location estimation in large indoor multi-floor buildings using hybrid networks | |
| He et al. | WiFi iLocate: WiFi based indoor localization for smartphone | |
| CN105044659B (en) | Indoor positioning device and method based on ambient light spectrum fingerprint | |
| CN106358233B (en) | A kind of RSS data smoothing method based on Multidimensional Scaling algorithm | |
| CN109212464B (en) | Method and equipment for estimating terminal distance and position planning | |
| Wang et al. | Study on improvement of fingerprint matching algorithm in wireless LAN based indoor positioning system | |
| CN111836188B (en) | Online cooperative positioning and system based on Wi-Fi RSS | |
| Pirzada et al. | Location fingerprinting technique for WLAN device-free indoor localization system | |
| Imran et al. | A Novel Indoor Positioning System Using Kernel Local Discriminant Analysis in Internet‐of‐Things | |
| CN110471030A (en) | Based on the radio frequency tomography passive type Position-Solving method for improving conjugate gradient |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |