CN111885700A - Mobile terminal positioning method and device combined with support vector machine - Google Patents
Mobile terminal positioning method and device combined with support vector machine Download PDFInfo
- Publication number
- CN111885700A CN111885700A CN202010515649.2A CN202010515649A CN111885700A CN 111885700 A CN111885700 A CN 111885700A CN 202010515649 A CN202010515649 A CN 202010515649A CN 111885700 A CN111885700 A CN 111885700A
- Authority
- CN
- China
- Prior art keywords
- support vector
- target
- vector machine
- division line
- reference signal
- 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.)
- Granted
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W64/00—Locating users or terminals or network equipment for network management purposes, e.g. mobility management
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/02—Services making use of location information
- H04W4/021—Services related to particular areas, e.g. point of interest [POI] services, venue services or geofences
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/30—Services specially adapted for particular environments, situations or purposes
- H04W4/33—Services specially adapted for particular environments, situations or purposes for indoor environments, e.g. buildings
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Position Fixing By Use Of Radio Waves (AREA)
Abstract
The invention discloses a mobile terminal positioning method and device combined with a support vector machine. In the off-line training stage, each reference position on a map and reference signal characteristics corresponding to the reference position are used as RF fingerprint data, an RF fingerprint database is built, corresponding training sets are obtained from the RF fingerprint database according to each division line defined on the map to train a corresponding support vector machine, in the on-line positioning stage, target signal characteristics obtained by a terminal to be positioned are respectively input into all the trained support vector machines, a target area of the terminal to be positioned is determined according to all output results, then the RF fingerprint data in the target area are extracted from the RF fingerprint database, the target signal characteristics are compared with the RF fingerprint data in the target area according to a KNN algorithm, the target position of the terminal to be positioned is determined, and therefore accurate positioning of the mobile terminal is achieved. The invention can effectively improve the indoor positioning precision of the mobile terminal.
Description
Technical Field
The invention relates to the technical field of wireless positioning, in particular to a mobile terminal positioning method and device combined with a support vector machine.
Background
In recent years, with the rapid increase in the popularity of mobile terminals, various mobile applications have been developed in a well-spraying manner, wherein Location Based Services (LBS) applications range from medical care, logistics management, security, navigation, location based information delivery, location based network security, location based user participation in games, and the like.
Currently, the mainstream wireless positioning method is based on an RF fingerprint system for positioning, that is, in an off-line stage, features of wireless signals at different indoor positions are extracted by the RF fingerprint system, and a correspondence between the positions and the features of the wireless signals is stored in an RF fingerprint database, and in an on-line stage, the currently acquired features of the wireless signals are compared with the features of the wireless signals in the RF fingerprint database by a mobile terminal, so as to determine the position of the mobile terminal. In fact, an indoor propagation environment is very complex and is influenced by multipath propagation of wireless signals, shielding of indoor objects and the like, the wireless signals have small-scale fading, and the signal intensity between two positions with a very close distance may be greatly different.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a mobile terminal positioning method and device combined with a support vector machine, which can effectively improve the indoor positioning accuracy of a mobile terminal.
In order to solve the above technical problem, in a first aspect, an embodiment of the present invention provides a method for positioning a mobile terminal by using a support vector machine, including:
s1, an off-line training stage, which specifically comprises:
taking each reference position on the map and the reference signal characteristic corresponding to the reference position as RF fingerprint data to construct an RF fingerprint database; wherein the reference signal characteristic is a signal strength of a wireless signal measured at the reference location;
according to each division line defined on the map, acquiring a corresponding training set from the RF fingerprint database, and training a corresponding support vector machine through the training set; wherein the division lines comprise horizontal division lines and/or vertical division lines;
s2, an online positioning stage, which specifically comprises:
respectively inputting the target signal characteristics acquired by the terminal to be positioned into all the trained support vector machines, and determining a target area of the terminal to be positioned according to all the output results; wherein the target signal characteristic is the signal strength of the wireless signal measured by the terminal to be positioned;
and extracting RF fingerprint data in the target area from the RF fingerprint database, comparing the target signal characteristics with the RF fingerprint data in the target area according to a KNN algorithm, and determining the target position of the terminal to be positioned.
Further, the acquiring a corresponding training set from the RF fingerprint database according to each division line defined on the map and training a corresponding support vector machine through the training set, includes:
defining n of said horizontal dividing lines on said map; wherein n is more than or equal to 1;
according to each horizontal division line, taking the position relation between each reference signal feature in the RF fingerprint database and the reference position to which the reference signal feature belongs and the horizontal division line as training data to obtain the corresponding training set;
and training a corresponding support vector machine through each training set, so that the support vector machine outputs the position relation between the reference position to which the reference signal characteristic belongs and the horizontal division line according to the input reference signal characteristic.
Further, the acquiring a corresponding training set from the RF fingerprint database according to each division line defined on the map and training a corresponding support vector machine through the training set, includes:
defining m of said vertical dividing lines on said map; wherein m is more than or equal to 1;
according to each vertical division line, taking the position relation between each reference signal feature in the RF fingerprint database and the reference position to which the reference signal feature belongs and the vertical division line as training data to obtain the corresponding training set;
and training a corresponding support vector machine through each training set, so that the support vector machine outputs the position relation between the reference position to which the reference signal characteristic belongs and the vertical division line according to the input reference signal characteristic.
Further, the step of respectively inputting the target signal characteristics acquired by the terminal to be positioned into all the trained support vector machines, and determining the target area of the terminal to be positioned according to all the output results specifically includes:
and respectively inputting the target signal characteristics into all the trained support vector machines, enabling each support vector machine to output the position relationship between the position of the target signal characteristics and the division line according to the target signal characteristics, and determining the target area according to all the position relationships.
Further, the extracting the RF fingerprint data in the target area from the RF fingerprint database, comparing the target signal feature with the RF fingerprint data in the target area according to a KNN algorithm, and determining the target position of the terminal to be positioned specifically is:
extracting RF fingerprint data within the target area from the RF fingerprint database;
calculating the Euclidean distance between the reference signal feature and the target signal feature in each target region according to a KNN algorithm;
and screening a plurality of reference signal characteristics according to the sequence of the Euclidean distances from small to large, and determining the target position according to the reference positions to which all the screened reference signal characteristics belong.
In a second aspect, an embodiment of the present invention provides a mobile terminal positioning apparatus incorporating a support vector machine, including:
the offline training module specifically comprises:
the RF fingerprint database construction unit is used for constructing an RF fingerprint database by taking each reference position on a map and the reference signal characteristics corresponding to the reference position as RF fingerprint data; wherein the reference signal characteristic is a signal strength of a wireless signal measured at the reference location;
a support vector machine training unit, configured to obtain a corresponding training set from the RF fingerprint database according to each division line defined on the map, and train a corresponding support vector machine through the training set; wherein the division lines comprise horizontal division lines and/or vertical division lines;
the online positioning module specifically comprises:
the target area determining unit is used for respectively inputting the target signal characteristics acquired by the terminal to be positioned into all the trained support vector machines and determining the target area of the terminal to be positioned according to all the output results; wherein the target signal characteristic is the signal strength of the wireless signal measured by the terminal to be positioned;
and the target position determining unit is used for extracting the RF fingerprint data in the target area from the RF fingerprint database, comparing the target signal characteristics with the RF fingerprint data in the target area according to a KNN algorithm, and determining the target position of the terminal to be positioned.
Further, the acquiring a corresponding training set from the RF fingerprint database according to each division line defined on the map and training a corresponding support vector machine through the training set, includes:
defining n of said horizontal dividing lines on said map; wherein n is more than or equal to 1;
according to each horizontal division line, taking the position relation between each reference signal feature in the RF fingerprint database and the reference position to which the reference signal feature belongs and the horizontal division line as training data to obtain the corresponding training set;
and training a corresponding support vector machine through each training set, so that the support vector machine outputs the position relation between the reference position to which the reference signal characteristic belongs and the horizontal division line according to the input reference signal characteristic.
Further, the acquiring a corresponding training set from the RF fingerprint database according to each division line defined on the map and training a corresponding support vector machine through the training set, includes:
defining m of said vertical dividing lines on said map; wherein m is more than or equal to 1;
according to each vertical division line, taking the position relation between each reference signal feature in the RF fingerprint database and the reference position to which the reference signal feature belongs and the vertical division line as training data to obtain the corresponding training set;
and training a corresponding support vector machine through each training set, so that the support vector machine outputs the position relation between the reference position to which the reference signal characteristic belongs and the vertical division line according to the input reference signal characteristic.
Further, the step of respectively inputting the target signal characteristics acquired by the terminal to be positioned into all the trained support vector machines, and determining the target area of the terminal to be positioned according to all the output results specifically includes:
and respectively inputting the target signal characteristics into all the trained support vector machines, enabling each support vector machine to output the position relationship between the position of the target signal characteristics and the division line according to the target signal characteristics, and determining the target area according to all the position relationships.
Further, the extracting the RF fingerprint data in the target area from the RF fingerprint database, comparing the target signal feature with the RF fingerprint data in the target area according to a KNN algorithm, and determining the target position of the terminal to be positioned specifically is:
extracting RF fingerprint data within the target area from the RF fingerprint database;
calculating the Euclidean distance between the reference signal feature and the target signal feature in each target region according to a KNN algorithm;
and screening a plurality of reference signal characteristics according to the sequence of the Euclidean distances from small to large, and determining the target position according to the reference positions to which all the screened reference signal characteristics belong.
The embodiment of the invention has the following beneficial effects:
in an off-line training stage, taking each reference position on a map and a reference signal characteristic corresponding to the reference position, namely the signal intensity of a wireless signal measured at the reference position as RF fingerprint data, constructing an RF fingerprint database, acquiring a corresponding training set from the RF fingerprint database according to each division line defined on the map to train a corresponding support vector machine, in an on-line positioning stage, respectively inputting a target signal characteristic acquired by a terminal to be positioned, namely the signal intensity of the wireless signal measured by the terminal to be positioned, into all the support vector machines after training, determining a target area of the terminal to be positioned according to all output results, further extracting the RF fingerprint data in the target area from the RF fingerprint database, comparing the target signal characteristic with the RF fingerprint data in the target area according to a KNN algorithm, and determining the target position of the terminal to be positioned, therefore, accurate positioning of the mobile terminal is achieved. Compared with the prior art, the embodiment of the invention combines the RF fingerprint database and the support vector machine, utilizes the trained support vector machine to perform coarse positioning on the terminal to be positioned, determines the target area, further performs fine positioning on the target area based on the RF fingerprint database, determines the target position, and can effectively improve the positioning precision of the mobile terminal indoors.
Drawings
Fig. 1 is a schematic flowchart of a mobile terminal positioning method in combination with a support vector machine according to a first embodiment of the present invention;
FIG. 2 is a schematic flow chart of an offline training phase according to a first embodiment of the present invention;
FIG. 3 is a flow chart illustrating an online positioning phase according to a first embodiment of the present invention;
fig. 4 is another schematic flow chart of a method for positioning a mobile terminal in combination with a support vector machine according to a first embodiment of the present invention;
FIG. 5 is a schematic diagram of the layout of reference positions in the second embodiment of the present invention;
fig. 6 is a schematic structural diagram of a mobile terminal positioning apparatus incorporating a support vector machine according to a third embodiment of the present invention.
Detailed Description
The technical solutions in the present invention will be described clearly and completely with reference to the accompanying drawings, and it is obvious that the described embodiments are only some embodiments of the present invention, 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 should be noted that, the step numbers in the text are only for convenience of explanation of the specific embodiments, and do not serve to limit the execution sequence of the steps.
As shown in fig. 1-4, a first embodiment provides a method for positioning a mobile terminal in conjunction with a support vector machine, including steps S1-S2:
s1, an off-line training stage, which specifically comprises the following steps S11-S12:
s11, taking each reference position on the map and the reference signal characteristic corresponding to the reference position as RF fingerprint data to construct an RF fingerprint database; wherein the reference signal characteristic is a signal strength of the wireless signal measured at the reference location.
S12, acquiring a corresponding training set from the RF fingerprint database according to each division line defined on the map, and training a corresponding support vector machine through the training set; wherein the division lines comprise horizontal division lines and/or vertical division lines.
S2, an online positioning stage, which specifically comprises the steps S21-S22:
s21, respectively inputting the target signal characteristics acquired by the terminal to be positioned into all trained support vector machines, and determining the target area of the terminal to be positioned according to all output results; the target signal characteristic is the signal strength of the wireless signal measured by the terminal to be positioned.
And S22, extracting the RF fingerprint data in the target area from the RF fingerprint database, comparing the target signal characteristics with the RF fingerprint data in the target area according to a KNN algorithm, and determining the target position of the terminal to be positioned.
As an example, in the process of constructing the RF fingerprint database in the offline training stage, the map is divided into a plurality of grids, a known Reference position (RP: Reference Point) is taken from each grid to obtain p Reference positions, then the Signal strength (RSS: Received Signal strength) of the wireless Signal measured at each Reference position is used as a Reference Signal feature, and each Reference position and the Reference Signal feature corresponding to the Reference position are used as one RF fingerprint data to construct the RF fingerprint database.
For q Access Points (AP) in the environment, the reference signal characteristic corresponding to the s-th reference position may be recorded as a vector.
In the process of training a plurality of support vector machines in an off-line training stage, if n horizontal division lines are defined on a map, namely the map is divided into a plurality of transverse areas, a corresponding training set is obtained from an RF fingerprint database according to each horizontal division line, and the corresponding support vector machine f is trained through the training setnMake the support vector machine fnOutputting the position relation of the reference position to which the reference signal characteristic belongs and the horizontal division line according to the input reference signal characteristic, wherein for example, output 1 indicates that the reference position is on the south side of the horizontal division line, and output-1 indicates that the reference position is on the north side of the horizontal division line; if m vertical division lines are defined on the map, which is equivalent to dividing the map into a plurality of longitudinal areas, acquiring a corresponding training set from the RF fingerprint database according to each vertical division line, and training a support vector machine g through the training setmMake the support vector machine gmAnd outputting the position relation of the reference position to which the reference signal feature belongs and the vertical division line according to the input reference signal feature, wherein for example, output 1 indicates that the reference position is on the east side of the vertical division line, and output-1 indicates that the reference position is on the west side of the vertical division line.
Wherein, when n is 0, it is equivalent to not defining a horizontal division line on the map; when m is 0, it is equivalent to not defining a vertical division line on the map.
In the process of determining the target area of the terminal to be positioned in the on-line positioning stage, the signal intensity of a wireless signal measured by the terminal to be positioned is taken as a target signal characteristic, and the target signal characteristic is respectively input into all trained support vector machines fnAnd/or gmFrom a support vector machine fnAnd/or gmAnd outputting the position relation between the position of the target signal characteristic and the division line according to the target signal characteristic, and determining a target area of the terminal to be positioned according to all the position relations.
In the process of determining the target position of the terminal to be positioned in the online positioning stage, firstly, extracting RF fingerprint data in a target area from an RF fingerprint database, then comparing the characteristics of a target signal with the RF fingerprint data in the target area according to a KNN algorithm, taking a plurality of RF fingerprint data which are most matched for calculation, and determining the target position of the terminal to be positioned.
In the embodiment, in an off-line training stage, each reference position on a map and a reference signal feature corresponding to the reference position, that is, the signal intensity of a wireless signal measured at the reference position are used as RF fingerprint data to construct an RF fingerprint database, and according to each division line defined on the map, a corresponding training set is obtained from the RF fingerprint database to train a corresponding support vector machine, in an on-line positioning stage, a target signal feature obtained by a terminal to be positioned, that is, the signal intensity of a wireless signal measured by the terminal to be positioned, is respectively input to all the support vector machines after training, a target area of the terminal to be positioned is determined according to all output results, further RF fingerprint data in the target area is extracted from the RF fingerprint database, and according to a KNN algorithm, the target signal feature is compared with the RF fingerprint data in the target area to determine the target position of the terminal to be positioned, therefore, accurate positioning of the mobile terminal is achieved. In the embodiment, the RF fingerprint database and the support vector machine are combined, the trained support vector machine is utilized to perform coarse positioning on the terminal to be positioned, the target area is determined, the target area is further precisely positioned based on the RF fingerprint database, the target position is determined, and the indoor positioning accuracy of the mobile terminal can be effectively improved.
In a preferred embodiment, the acquiring a corresponding training set from the RF fingerprint database according to each division line defined on the map and training a corresponding support vector machine through the training set, includes: defining n horizontal division lines on a map; wherein n is more than or equal to 1; according to each horizontal division line, taking the position relation between each reference signal characteristic and the reference position to which the reference signal characteristic belongs in the RF fingerprint database and the horizontal division line as training data to obtain a corresponding training set; and training the corresponding support vector machine through each training set, so that the support vector machine outputs the position relation between the reference position to which the reference signal characteristic belongs and the horizontal division line according to the input reference signal characteristic.
As an example, if only at least one horizontal division line is defined on the map, which is equivalent to dividing the map into a plurality of horizontal areas, according to each horizontal division line, the reference signal feature in the RF fingerprint database and the position relationship between the reference position to which the reference signal feature belongs and the horizontal division line are used as a training data, a corresponding training set is obtained, and the corresponding support vector machine f is trained through the training setnMake the support vector machine fnAnd outputting the position relation of the reference position to which the reference signal characteristic belongs and the horizontal division line according to the input reference signal characteristic, wherein for example, output 1 indicates that the reference position is on the south side of the horizontal division line, and output-1 indicates that the reference position is on the north side of the horizontal division line.
Taking the r-th horizontal division line as an example, the specific training method is as follows:
the first step is as follows: construction of training set T of support vector machine by RSS vectors in RF fingerprint databaser＝{(RSS1,l1),(RSS2,l2),...,(RSSp,lp)}，lkE γ { -1,1}, k 1kThe value 1 indicates that the kth reference position is south of the horizontal division line, lkThe value-1 indicates that the kth reference position is to the north side of the horizontal division line.
The second step is that: and constructing and solving a convex quadratic programming problem.
Wherein the content of the first and second substances,
is a Gaussian kernel function, parameter C in formula (2)>0 is a penalty parameter, sigma is a noise variance, and can be selected according to experience, and the convex optimization problem can be solved
The solving method is a conventional method in the art, and an existing solving tool is available, which is not described herein.
The third step: selecting alpha in the open interval (0, C)*Component (b) of
Calculation of b*。
The fourth step: constructing a decision function fr(RSS)。
Wherein the content of the first and second substances,
as a sign function, when the decision function frWhen the (RSS) value is 1, the reference position is judged to be on the south side of the horizontal division line, and when the decision function frWhen the value of (RSS) is-1, the reference position is judged to be at the north side of the horizontal division line, and when the decision function frWhen the value of (RSS) is 0, the horizontal division line does not work, and the horizontal division line is considered to be not set, so that the continuous operation of the positioning method is not influenced.
Considering that the positioning problem is not necessarily a linear separable problem, if a linear support vector machine is adopted, an additional error is introduced, and the support vector machine adopting the gaussian kernel in the embodiment is not limited by the linear separable problem, so that the positioning accuracy of the mobile terminal can be further improved.
In a preferred embodiment, the acquiring a corresponding training set from the RF fingerprint database according to each division line defined on the map and training a corresponding support vector machine through the training set, includes: defining m vertical division lines on a map; wherein m is more than or equal to 1; according to each vertical division line, taking the position relation between each reference signal characteristic and the reference position to which the reference signal characteristic belongs in the RF fingerprint database and the vertical division line as training data to obtain a corresponding training set; and training the corresponding support vector machine through each training set, so that the support vector machine outputs the position relation between the reference position to which the reference signal characteristic belongs and the vertical division line according to the input reference signal characteristic.
As an example, if only at least one vertical division line is defined on the map, which is equivalent to dividing the map into a plurality of longitudinal areas, according to each vertical division line, the position relationship between each reference signal feature in the RF fingerprint database and the reference position to which the reference signal feature belongs and the vertical division line is used as a training data, a corresponding training set is obtained, and a support vector machine g is trained through the training setmMake the support vector machine gmOutputting the position relation between the reference position to which the reference signal characteristic belongs and the vertical division line according to the input reference signal characteristic, for example, outputting 1 to indicate that the reference position is inOn the east side of the vertical dashed line, output-1 indicates that the reference position is on the west side of the vertical dashed line.
Taking the r-th vertical division line as an example, the specific training method is as follows:
the first step is as follows: construction of training set T of support vector machine by RSS vectors in RF fingerprint databaser＝{(RSS1,l1),(RSS2,l2),...,(RSSp,lp)}，lkE γ { -1,1}, k 1kThe value 1 indicates that the kth reference position is on the east side of the vertical dividing line, lkThe value-1 indicates that the kth reference position is west of the vertical dividing line.
The second step is that: and constructing and solving a convex quadratic programming problem.
Wherein the content of the first and second substances,
is a Gaussian kernel function, parameter C in equation (7)>0 is a penalty parameter, sigma is a noise variance, and can be selected according to experience, and the convex optimization problem can be solved
The solving method is a conventional method in the art, and an existing solving tool is available, which is not described herein.
The third step: selecting alpha in the open interval (0, C)*Component (b) of
Calculation of b*。
The fourth step: constructing a decision function fr(RSS)。
Wherein the content of the first and second substances,
as a sign function, when the decision function frWhen the value of (RSS) is 1, the reference position is judged to be on the east side of the vertical division line, and when the decision function frWhen the value of (RSS) is-1, the reference position is judged to be on the west side of the vertical division line, and when the decision function frWhen the value of (RSS) is 0, the vertical division line does not work, and the vertical division line is considered to be not set, so that the continuous operation of the positioning method is not influenced.
Considering that the positioning problem is not necessarily a linear separable problem, if a linear support vector machine is adopted, an additional error is introduced, and the support vector machine adopting the gaussian kernel in the embodiment is not limited by the linear separable problem, so that the positioning accuracy of the mobile terminal can be further improved.
As an example, if at least one horizontal division line and at least one vertical division line are defined on the map, which is equivalent to dividing the map into a plurality of grid areas, a corresponding training set is obtained from the RF fingerprint database according to each division line defined on the map, and the corresponding support vector machine is trained through the training set.
In a preferred embodiment, the step of respectively inputting the target signal characteristics acquired by the terminal to be positioned into all trained support vector machines, and determining the target region of the terminal to be positioned according to all output results specifically includes: and respectively inputting the target signal characteristics into all trained support vector machines, enabling each support vector machine to output the position relation between the position of the target signal characteristics and the division line according to the target signal characteristics, and determining a target area according to all the position relations.
By way of example, radio measured by a terminal to be positionedSignal strength of a signal as a target signal characteristic RSSolCharacterizing the target signal by RSSolRespectively inputting all trained support vector machines fnAnd/or gmFrom a support vector machine fnAnd/or gmBased on target signal characteristics RSSolOutputting the position relation f of the position of the target signal characteristic and the division line1(RSSol),f2(RSSol),...,fn(RSSol)，g1(RSSol),g2(RSSol),...,gm(RSSol) Then according to all the position relations f1(RSSol),f2(RSSol),...,fn(RSSol)，g1(RSSol),g2(RSSol),...,gm(RSSol) And roughly positioning the position of the terminal to be positioned in an area surrounded by the horizontal division line and/or the vertical division line, and determining a target area of the terminal to be positioned.
In the embodiment, the plurality of support vector machines are trained in the offline training stage, and the trained support vector machines are utilized to perform coarse positioning on the terminal to be positioned, so that the positioning accuracy of the mobile terminal is improved.
In a preferred embodiment, the extracting RF fingerprint data in the target area from the RF fingerprint database, and comparing the target signal characteristic with the RF fingerprint data in the target area according to the KNN algorithm to determine the target position of the terminal to be located specifically includes: extracting RF fingerprint data in the target area from the RF fingerprint database; calculating the Euclidean distance between the reference signal characteristic and the target signal characteristic in each target area according to the KNN algorithm; and screening a plurality of reference signal characteristics according to the sequence of the Euclidean distance from small to large, and determining the target position according to the reference positions to which all the screened reference signal characteristics belong.
As an example, the RF fingerprint data in the target area is extracted from the RF fingerprint database, then the target signal characteristics are compared with the RF fingerprint data in the target area according to the KNN algorithm, several RF fingerprint data which are most matched are taken for calculation, and the target position of the terminal to be positioned is determined. The method comprises the following specific steps:
the first step is as follows: taking RF fingerprint data in the target area in the RF fingerprint database, and recording as RSSrough＝{RSSx1,RSSx2,...,RSSxt}。
The second step is that: calculating target signal characteristics and RSSroughThe euclidean distance between.
di＝||RSSol-RSSxi||2,i＝1,2,...,t (11)
The third step: selecting t minimum distances from the distances calculated in the second step, and recording the corresponding reference position as P1,...,PtLet the coordinates corresponding to these reference positions be (x) respectively1,y1),(x2,y2),...,(xt,yt)。
Estimating the target position from these coordinates:
wherein the content of the first and second substances,
as a weighting coefficient when
The estimated coordinates of the target position are then the average of the coordinates of the corresponding reference positions.
As shown in fig. 5, a second embodiment based on the first embodiment provides a method for positioning a mobile terminal by using a support vector machine, which includes steps S1-S2:
s1, an off-line training stage, which specifically comprises:
the acquisition of RF fingerprint data is performed in an office area of about 30 m x 20 m, 6 access points AP are deployed in the environment, 252 reference positions RP are taken within the office area, and the map is divided into 3 lateral areas with 2 horizontal dividing lines.
From top to bottom, the reference positions from left to right are respectively denoted as RP1,RP2,...,RP252The signal strength RSS of the wireless signal is used as the reference signal characteristic, so that the RF fingerprint data can be recorded as RSSs＝{RSSs1,RSSs2,...,RSSs6}，s＝1,2,...,252。
In the present embodiment, the reference position of the north of the horizontal division line a is RP1,RP2,...,RP54The south reference position of the horizontal division line A is RP55,RP56,...,RP252From these, a support vector machine f can be constructedAThe training set of (2): t isA＝{(RSS1,-1),...,(RSS54,-1),(RSS55,1),...,(RSS252,1)}。
Similarly, the horizontal division line B has the north reference position as RP1,RP2,...,RP54The south reference position of the horizontal division line B is RP55,RP56,...,RP252From these, a support vector machine f can be constructedBThe training set of (2): t isB＝{(RSS1,-1),...,(RSS216,-1),(RSS217,1),...,(RSS252,1)}。
Corresponding support vector machine f can be constructed according to the formulas (1) to (5)A、fB。
S2, an online positioning stage, which specifically comprises:
the terminal to be positioned is in the middle area of the two horizontal division lines, and the RSS vector measured by the terminal to be positioned is RSSol. Will RSSolSubstituting the trained support vector machine fA、fBGet fA(RSSol)＝1，fB(RSSol) And therefore, the target region is determined to be the middle region.
Taking RF fingerprint data in the middle region RSS55,RSS56,...,RSS216And RSSolMatching, calculating vector distance according to formula (11), and taking 3 RPs with minimum vector distance and recording as RPo1,RPo2,RPo3Provided with themRespectively is (x)o1,yo1),(xo2,yo2),(xo3,yo3). And finally, estimating the target position:
thereby completing the position estimation of the target to be positioned.
In this embodiment, a plurality of target positions are randomly selected on a map, and the above steps are repeated to perform a plurality of tests, which proves that the average positioning accuracy of the method can be improved by more than 20% compared with that of the conventional KNN algorithm.
As shown in fig. 6, a third embodiment provides a mobile terminal positioning apparatus incorporating a support vector machine, including: the offline training module 21 specifically includes: an RF fingerprint database construction unit 211, configured to construct an RF fingerprint database by using each reference location on the map and the reference signal feature corresponding to the reference location as an RF fingerprint data; wherein the reference signal characteristic is a signal strength of the wireless signal measured at the reference location; a support vector machine training unit 212, configured to obtain a corresponding training set from the RF fingerprint database according to each division line defined on the map, and train a corresponding support vector machine through the training set; wherein the division lines comprise horizontal division lines and/or vertical division lines; the online positioning module 22 specifically includes: the target area determining unit 221 is configured to input the target signal characteristics acquired by the terminal to be positioned into all trained support vector machines, and determine the target area of the terminal to be positioned according to all output results; the target signal is characterized by the signal intensity of a wireless signal measured by a terminal to be positioned; and the target position determining unit 222 is configured to extract RF fingerprint data in the target area from the RF fingerprint database, compare the target signal characteristics with the RF fingerprint data in the target area according to a KNN algorithm, and determine a target position of the terminal to be positioned.
Illustratively, the RF fingerprint database construction unit 211 divides the map into a plurality of grids, obtains p Reference positions from each grid by taking a known Reference position (RP: Reference Point), and constructs an RF fingerprint database by taking the Signal strength (RSS: Received Signal strength) of the wireless Signal measured at each Reference position as a Reference Signal feature and taking each Reference position and the Reference Signal feature corresponding to the Reference position as one RF fingerprint data.
For q Access Points (AP) in the environment, the reference signal characteristic corresponding to the s-th reference position may be recorded as a vector.
If n horizontal division lines are defined on the map by the support vector machine training unit 212, which is equivalent to dividing the map into a plurality of horizontal areas, a corresponding training set is obtained from the RF fingerprint database according to each horizontal division line, and the corresponding support vector machine f is trained by the training setnMake the support vector machine fnOutputting the position relation of the reference position to which the reference signal characteristic belongs and the horizontal division line according to the input reference signal characteristic, wherein for example, output 1 indicates that the reference position is on the south side of the horizontal division line, and output-1 indicates that the reference position is on the north side of the horizontal division line; if m vertical division lines are defined on the map, which is equivalent to dividing the map into a plurality of longitudinal areas, acquiring a corresponding training set from the RF fingerprint database according to each vertical division line, and training a support vector machine g through the training setmMake the support vector machine gmAnd outputting the position relation of the reference position to which the reference signal feature belongs and the vertical division line according to the input reference signal feature, wherein for example, output 1 indicates that the reference position is on the east side of the vertical division line, and output-1 indicates that the reference position is on the west side of the vertical division line.
Wherein, when n is 0, it is equivalent to not defining a horizontal division line on the map; when m is 0, it is equivalent to not defining a vertical division line on the map.
By the target area determining unit 221, the signal strength of the wireless signal measured by the terminal to be positioned is taken as the target signal characteristics, and the target signal characteristics are respectively input into all trained support vector machines fnAnd/or gmFrom a support vector machine fnAnd/or gmOutputting the target signal according to the target signal characteristicsAnd determining a target area of the terminal to be positioned according to the position relation between the position of the characteristic and the position of the division line.
Through the target position determining unit 222, the RF fingerprint data in the target area is extracted from the RF fingerprint database, and then the target signal characteristics are compared with the RF fingerprint data in the target area according to the KNN algorithm, and a plurality of RF fingerprint data that are most matched are selected for calculation, so as to determine the target position of the terminal to be positioned.
In the embodiment, each reference position on the map and the reference signal characteristic corresponding to the reference position, namely, the signal intensity of the wireless signal measured at the reference position are taken as RF fingerprint data by the offline training module 21, an RF fingerprint database is constructed, a corresponding training set is obtained from the RF fingerprint database according to each division line defined on the map to train a corresponding support vector machine, the target signal characteristic obtained by the terminal to be positioned, namely, the signal intensity of the wireless signal measured by the terminal to be positioned is respectively input into all the support vector machines after training by the online positioning module 22, the target area of the terminal to be positioned is determined according to all the output results, further, the RF fingerprint data in the target area is extracted from the RF fingerprint database, and the target signal characteristic is compared with the RF fingerprint data in the target area according to the KNN algorithm, and determining the target position of the terminal to be positioned, thereby realizing the accurate positioning of the mobile terminal. In the embodiment, the RF fingerprint database and the support vector machine are combined, the trained support vector machine is utilized to perform coarse positioning on the terminal to be positioned, the target area is determined, the target area is further precisely positioned based on the RF fingerprint database, the target position is determined, and the indoor positioning accuracy of the mobile terminal can be effectively improved.
In a preferred embodiment, the acquiring a corresponding training set from the RF fingerprint database according to each division line defined on the map and training a corresponding support vector machine through the training set, includes: defining n horizontal division lines on a map; wherein n is more than or equal to 1; according to each horizontal division line, taking the position relation between each reference signal characteristic and the reference position to which the reference signal characteristic belongs in the RF fingerprint database and the horizontal division line as training data to obtain a corresponding training set; and training the corresponding support vector machine through each training set, so that the support vector machine outputs the position relation between the reference position to which the reference signal characteristic belongs and the horizontal division line according to the input reference signal characteristic.
As an example, if only at least one horizontal division line is defined on the map, which is equivalent to dividing the map into a plurality of horizontal areas, according to each horizontal division line, the reference signal feature in the RF fingerprint database and the position relationship between the reference position to which the reference signal feature belongs and the horizontal division line are used as a training data, a corresponding training set is obtained, and the corresponding support vector machine f is trained through the training setnMake the support vector machine fnAnd outputting the position relation of the reference position to which the reference signal characteristic belongs and the horizontal division line according to the input reference signal characteristic, wherein for example, output 1 indicates that the reference position is on the south side of the horizontal division line, and output-1 indicates that the reference position is on the north side of the horizontal division line.
Taking the r-th horizontal division line as an example, the specific training method is as follows:
the first step is as follows: construction of training set T of support vector machine by RSS vectors in RF fingerprint databaser＝{(RSS1,l1),(RSS2,l2),...,(RSSp,lp)}，lkE γ { -1,1}, k 1kThe value 1 indicates that the kth reference position is south of the horizontal division line, lkThe value-1 indicates that the kth reference position is to the north side of the horizontal division line.
The second step is that: and constructing and solving a convex quadratic programming problem.
Wherein the content of the first and second substances,
is a Gaussian kernel function, parameter C in formula (2)>0 is a penalty parameter, sigma is a noise variance, and can be selected according to experience, and the convex optimization problem can be solved
The solving method is a conventional method in the art, and an existing solving tool is available, which is not described herein.
The third step: selecting alpha in the open interval (0, C)*Component (b) of
Calculation of b*。
The fourth step: constructing a decision function fr(RSS)。
Wherein the content of the first and second substances,
as a sign function, when the decision function frWhen the (RSS) value is 1, the reference position is judged to be on the south side of the horizontal division line, and when the decision function frWhen the value of (RSS) is-1, the reference position is judged to be at the north side of the horizontal division line, and when the decision function frWhen the value of (RSS) is 0, the horizontal division line does not work, and the horizontal division line is considered to be not set, so that the continuous operation of the positioning method is not influenced.
Considering that the positioning problem is not necessarily a linear separable problem, if a linear support vector machine is adopted, an additional error is introduced, and the support vector machine adopting the gaussian kernel in the embodiment is not limited by the linear separable problem, so that the positioning accuracy of the mobile terminal can be further improved.
In a preferred embodiment, the acquiring a corresponding training set from the RF fingerprint database according to each division line defined on the map and training a corresponding support vector machine through the training set, includes: defining m vertical division lines on a map; wherein m is more than or equal to 1; according to each vertical division line, taking the position relation between each reference signal characteristic and the reference position to which the reference signal characteristic belongs in the RF fingerprint database and the vertical division line as training data to obtain a corresponding training set; and training the corresponding support vector machine through each training set, so that the support vector machine outputs the position relation between the reference position to which the reference signal characteristic belongs and the vertical division line according to the input reference signal characteristic.
As an example, if only at least one vertical division line is defined on the map, which is equivalent to dividing the map into a plurality of longitudinal areas, according to each vertical division line, the position relationship between each reference signal feature in the RF fingerprint database and the reference position to which the reference signal feature belongs and the vertical division line is used as a training data, a corresponding training set is obtained, and a support vector machine g is trained through the training setmMake the support vector machine gmAnd outputting the position relation of the reference position to which the reference signal feature belongs and the vertical division line according to the input reference signal feature, wherein for example, output 1 indicates that the reference position is on the east side of the vertical division line, and output-1 indicates that the reference position is on the west side of the vertical division line.
Taking the r-th vertical division line as an example, the specific training method is as follows:
the first step is as follows: construction of training set T of support vector machine by RSS vectors in RF fingerprint databaser＝{(RSS1,l1),(RSS2,l2),...,(RSSp,lp)}，lkE γ { -1,1}, k 1kThe value 1 indicates that the kth reference position is on the east side of the vertical dividing line, lkThe value-1 indicates that the kth reference position is west of the vertical dividing line.
The second step is that: and constructing and solving a convex quadratic programming problem.
Wherein the content of the first and second substances,
is a Gaussian kernel function, parameter C in equation (7)>0 is a penalty parameter, sigma is a noise variance, and can be selected according to experience, and the convex optimization problem can be solved
The solving method is a conventional method in the art, and an existing solving tool is available, which is not described herein.
The third step: selecting alpha in the open interval (0, C)*Component (b) of
Calculation of b*。
The fourth step: constructing a decision function fr(RSS)。
Wherein the content of the first and second substances,
as a sign function, when the decision function frWhen the value of (RSS) is 1, the reference position is judged to be on the east side of the vertical division line, and when the decision function frWhen the value of (RSS) is-1, the reference position is judged to be on the west side of the vertical division line, and when the decision function frWhen the value of (RSS) is 0, the vertical division line does not work, the vertical division line is not considered to be arranged,the continuous operation of the positioning method is not influenced.
Considering that the positioning problem is not necessarily a linear separable problem, if a linear support vector machine is adopted, an additional error is introduced, and the support vector machine adopting the gaussian kernel in the embodiment is not limited by the linear separable problem, so that the positioning accuracy of the mobile terminal can be further improved.
As an example, if at least one horizontal division line and at least one vertical division line are defined on the map, which is equivalent to dividing the map into a plurality of grid areas, a corresponding training set is obtained from the RF fingerprint database according to each division line defined on the map, and the corresponding support vector machine is trained through the training set.
In a preferred embodiment, the step of respectively inputting the target signal characteristics acquired by the terminal to be positioned into all trained support vector machines, and determining the target region of the terminal to be positioned according to all output results specifically includes: and respectively inputting the target signal characteristics into all trained support vector machines, enabling each support vector machine to output the position relation between the position of the target signal characteristics and the division line according to the target signal characteristics, and determining a target area according to all the position relations.
Illustratively, the RSS is a target signal characteristic based on the signal strength of a wireless signal measured by a terminal to be positionedolCharacterizing the target signal by RSSolRespectively inputting all trained support vector machines fnAnd/or gmFrom a support vector machine fnAnd/or gmBased on target signal characteristics RSSolOutputting the position relation f of the position of the target signal characteristic and the division line1(RSSol),f2(RSSol),...,fn(RSSol)，g1(RSSol),g2(RSSol),...,gm(RSSol) Then according to all the position relations f1(RSSol),f2(RSSol),...,fn(RSSol)，g1(RSSol),g2(RSSol),...,gm(RSSol) The position of the terminal to be positioned is roughAnd positioning the terminal in the area enclosed by the horizontal division line and/or the vertical division line, and determining the target area of the terminal to be positioned.
In the embodiment, the plurality of support vector machines are trained in the offline training stage, and the trained support vector machines are utilized to perform coarse positioning on the terminal to be positioned, so that the positioning accuracy of the mobile terminal is improved.
In a preferred embodiment, the extracting RF fingerprint data in the target area from the RF fingerprint database, and comparing the target signal characteristic with the RF fingerprint data in the target area according to the KNN algorithm to determine the target position of the terminal to be located specifically includes: extracting RF fingerprint data in the target area from the RF fingerprint database; calculating the Euclidean distance between the reference signal characteristic and the target signal characteristic in each target area according to the KNN algorithm; and screening a plurality of reference signal characteristics according to the sequence of the Euclidean distance from small to large, and determining the target position according to the reference positions to which all the screened reference signal characteristics belong.
As an example, the RF fingerprint data in the target area is extracted from the RF fingerprint database, then the target signal characteristics are compared with the RF fingerprint data in the target area according to the KNN algorithm, several RF fingerprint data which are most matched are taken for calculation, and the target position of the terminal to be positioned is determined. The method comprises the following specific steps:
the first step is as follows: taking RF fingerprint data in the target area in the RF fingerprint database, and recording as RSSrough＝{RSSx1,RSSx2,...,RSSxt}。
The second step is that: calculating target signal characteristics and RSSroughThe euclidean distance between.
di＝||RSSol-RSSxi||2,i＝1,2,...,t (11)
The third step: selecting t minimum distances from the distances calculated in the second step, and recording the corresponding reference position as P1,...,PtLet the coordinates corresponding to these reference positions be (x) respectively1,y1),(x2,y2),...,(xt,yt)。
Estimating the target position from these coordinates:
wherein the content of the first and second substances,
as a weighting coefficient when
The estimated coordinates of the target position are then the average of the coordinates of the corresponding reference positions.
In summary, the embodiment of the present invention has the following advantages:
in an off-line training stage, taking each reference position on a map and a reference signal characteristic corresponding to the reference position, namely the signal intensity of a wireless signal measured at the reference position as RF fingerprint data, constructing an RF fingerprint database, acquiring a corresponding training set from the RF fingerprint database according to each division line defined on the map to train a corresponding support vector machine, in an on-line positioning stage, respectively inputting a target signal characteristic acquired by a terminal to be positioned, namely the signal intensity of the wireless signal measured by the terminal to be positioned, into all the support vector machines after training, determining a target area of the terminal to be positioned according to all output results, further extracting the RF fingerprint data in the target area from the RF fingerprint database, comparing the target signal characteristic with the RF fingerprint data in the target area according to a KNN algorithm, and determining the target position of the terminal to be positioned, therefore, accurate positioning of the mobile terminal is achieved. According to the embodiment of the invention, the RF fingerprint database and the support vector machine are combined, the trained support vector machine is utilized to perform coarse positioning on the terminal to be positioned, the target area is determined, the target area is further precisely positioned based on the RF fingerprint database, the target position is determined, and the positioning precision of the mobile terminal in a room can be effectively improved.
While the foregoing is directed to the preferred embodiment of the present invention, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention.
It will be understood by those skilled in the art that all or part of the processes of the above embodiments may be implemented by hardware related to instructions of a computer program, and the computer program may be stored in a computer readable storage medium, and when executed, may include the processes of the above embodiments. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.
Claims (10)
1. A mobile terminal positioning method combined with a support vector machine is characterized by comprising the following steps:
s1, an off-line training stage, which specifically comprises:
taking each reference position on the map and the reference signal characteristic corresponding to the reference position as RF fingerprint data to construct an RF fingerprint database; wherein the reference signal characteristic is a signal strength of a wireless signal measured at the reference location;
according to each division line defined on the map, acquiring a corresponding training set from the RF fingerprint database, and training a corresponding support vector machine through the training set; wherein the division lines comprise horizontal division lines and/or vertical division lines;
s2, an online positioning stage, which specifically comprises:
respectively inputting the target signal characteristics acquired by the terminal to be positioned into all the trained support vector machines, and determining a target area of the terminal to be positioned according to all the output results; wherein the target signal characteristic is the signal strength of the wireless signal measured by the terminal to be positioned;
and extracting RF fingerprint data in the target area from the RF fingerprint database, comparing the target signal characteristics with the RF fingerprint data in the target area according to a KNN algorithm, and determining the target position of the terminal to be positioned.
2. The method of claim 1, wherein the obtaining a corresponding training set from the RF fingerprint database according to each division line defined on the map and training the corresponding support vector machine through the training set comprises:
defining n of said horizontal dividing lines on said map; wherein n is more than or equal to 1;
according to each horizontal division line, taking the position relation between each reference signal feature in the RF fingerprint database and the reference position to which the reference signal feature belongs and the horizontal division line as training data to obtain the corresponding training set;
and training a corresponding support vector machine through each training set, so that the support vector machine outputs the position relation between the reference position to which the reference signal characteristic belongs and the horizontal division line according to the input reference signal characteristic.
3. The method as claimed in claim 1 or 2, wherein the obtaining a corresponding training set from the RF fingerprint database according to each division line defined on the map and training the corresponding support vector machine through the training set comprises:
defining m of said vertical dividing lines on said map; wherein m is more than or equal to 1;
according to each vertical division line, taking the position relation between each reference signal feature in the RF fingerprint database and the reference position to which the reference signal feature belongs and the vertical division line as training data to obtain the corresponding training set;
and training a corresponding support vector machine through each training set, so that the support vector machine outputs the position relation between the reference position to which the reference signal characteristic belongs and the vertical division line according to the input reference signal characteristic.
4. The method according to claim 1, wherein the method for positioning a mobile terminal using a support vector machine includes the steps of inputting the target signal characteristics obtained by the terminal to be positioned into all the trained support vector machines, and determining the target area of the terminal to be positioned according to all the output results, specifically:
and respectively inputting the target signal characteristics into all the trained support vector machines, enabling each support vector machine to output the position relationship between the position of the target signal characteristics and the division line according to the target signal characteristics, and determining the target area according to all the position relationships.
5. The method according to claim 1, wherein the step of extracting the RF fingerprint data in the target area from the RF fingerprint database, and comparing the target signal feature with the RF fingerprint data in the target area according to a KNN algorithm to determine the target position of the terminal to be positioned specifically comprises:
extracting RF fingerprint data within the target area from the RF fingerprint database;
calculating the Euclidean distance between the reference signal feature and the target signal feature in each target region according to a KNN algorithm;
and screening a plurality of reference signal characteristics according to the sequence of the Euclidean distances from small to large, and determining the target position according to the reference positions to which all the screened reference signal characteristics belong.
6. A mobile terminal positioning device incorporating a support vector machine, comprising:
the offline training module specifically comprises:
the RF fingerprint database construction unit is used for constructing an RF fingerprint database by taking each reference position on a map and the reference signal characteristics corresponding to the reference position as RF fingerprint data; wherein the reference signal characteristic is a signal strength of a wireless signal measured at the reference location;
a support vector machine training unit, configured to obtain a corresponding training set from the RF fingerprint database according to each division line defined on the map, and train a corresponding support vector machine through the training set; wherein the division lines comprise horizontal division lines and/or vertical division lines;
the online positioning module specifically comprises:
the target area determining unit is used for respectively inputting the target signal characteristics acquired by the terminal to be positioned into all the trained support vector machines and determining the target area of the terminal to be positioned according to all the output results; wherein the target signal characteristic is the signal strength of the wireless signal measured by the terminal to be positioned;
and the target position determining unit is used for extracting the RF fingerprint data in the target area from the RF fingerprint database, comparing the target signal characteristics with the RF fingerprint data in the target area according to a KNN algorithm, and determining the target position of the terminal to be positioned.
7. The apparatus of claim 6, wherein the means for obtaining a corresponding training set from the RF fingerprint database according to each division line defined on the map and training the corresponding support vector machine through the training set comprises:
defining n of said horizontal dividing lines on said map; wherein n is more than or equal to 1;
according to each horizontal division line, taking the position relation between each reference signal feature in the RF fingerprint database and the reference position to which the reference signal feature belongs and the horizontal division line as training data to obtain the corresponding training set;
and training a corresponding support vector machine through each training set, so that the support vector machine outputs the position relation between the reference position to which the reference signal characteristic belongs and the horizontal division line according to the input reference signal characteristic.
8. The apparatus of claim 6 or 7, wherein the means for obtaining a corresponding training set from the RF fingerprint database according to each division line defined on the map and training the corresponding support vector machine through the training set comprises:
defining m of said vertical dividing lines on said map; wherein m is more than or equal to 1;
according to each vertical division line, taking the position relation between each reference signal feature in the RF fingerprint database and the reference position to which the reference signal feature belongs and the vertical division line as training data to obtain the corresponding training set;
and training a corresponding support vector machine through each training set, so that the support vector machine outputs the position relation between the reference position to which the reference signal characteristic belongs and the vertical division line according to the input reference signal characteristic.
9. The mobile terminal positioning device combined with a support vector machine according to claim 6, wherein the target signal characteristics obtained by the terminal to be positioned are respectively input to all the trained support vector machines, and the target area of the terminal to be positioned is determined according to all the output results, specifically:
and respectively inputting the target signal characteristics into all the trained support vector machines, enabling each support vector machine to output the position relationship between the position of the target signal characteristics and the division line according to the target signal characteristics, and determining the target area according to all the position relationships.
10. The device according to claim 6, wherein the RF fingerprint data in the target area is extracted from the RF fingerprint database, and the target signal feature is compared with the RF fingerprint data in the target area according to a KNN algorithm to determine the target position of the terminal to be positioned, specifically:
extracting RF fingerprint data within the target area from the RF fingerprint database;
calculating the Euclidean distance between the reference signal feature and the target signal feature in each target region according to a KNN algorithm;
and screening a plurality of reference signal characteristics according to the sequence of the Euclidean distances from small to large, and determining the target position according to the reference positions to which all the screened reference signal characteristics belong.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010515649.2A CN111885700B (en) | 2020-06-08 | 2020-06-08 | Mobile terminal positioning method and device combined with support vector machine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010515649.2A CN111885700B (en) | 2020-06-08 | 2020-06-08 | Mobile terminal positioning method and device combined with support vector machine |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111885700A true CN111885700A (en) | 2020-11-03 |
| CN111885700B CN111885700B (en) | 2022-04-12 |
Family
ID=73156486
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010515649.2A Active CN111885700B (en) | 2020-06-08 | 2020-06-08 | Mobile terminal positioning method and device combined with support vector machine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111885700B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115022795A (en) * | 2021-02-18 | 2022-09-06 | 诺基亚技术有限公司 | RF fingerprinting mapping update |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101651951A (en) * | 2009-09-15 | 2010-02-17 | 哈尔滨工业大学 | Establishing method and positioning method of indoor positioning network of support vector machine based on WLAN |
| EP2798366A1 (en) * | 2011-12-29 | 2014-11-05 | Khalifa University of Science, Technology and Research | Method and system for localization |
| CN104469676A (en) * | 2014-11-21 | 2015-03-25 | 北京拓明科技有限公司 | Method and system for locating mobile terminal |
| CN105467382A (en) * | 2015-12-31 | 2016-04-06 | 南京信息工程大学 | SVM (Support Vector Machine)-based multi-sensor target tracking data fusion algorithm and system thereof |
| CN105528912A (en) * | 2016-01-20 | 2016-04-27 | 中山大学 | Indoor parking stall recommendation method and system |
| CN106604229A (en) * | 2016-12-27 | 2017-04-26 | 东南大学 | Indoor positioning method based on manifold learning and improved support vector machine |
| CN107333238A (en) * | 2017-07-03 | 2017-11-07 | 杭州电子科技大学 | A kind of indoor fingerprint method for rapidly positioning based on support vector regression |
| CN108989976A (en) * | 2018-06-04 | 2018-12-11 | 华中师范大学 | Fingerprint positioning method and system in a kind of wisdom classroom |
| CN109029429A (en) * | 2018-09-04 | 2018-12-18 | 电子科技大学 | Multi-categorizer overall situation dynamic fusion localization method based on WiFi and earth magnetism fingerprint |
| CN110012428A (en) * | 2019-05-22 | 2019-07-12 | 合肥工业大学 | A kind of indoor orientation method based on WiFi |
-
2020
- 2020-06-08 CN CN202010515649.2A patent/CN111885700B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101651951A (en) * | 2009-09-15 | 2010-02-17 | 哈尔滨工业大学 | Establishing method and positioning method of indoor positioning network of support vector machine based on WLAN |
| EP2798366A1 (en) * | 2011-12-29 | 2014-11-05 | Khalifa University of Science, Technology and Research | Method and system for localization |
| CN104469676A (en) * | 2014-11-21 | 2015-03-25 | 北京拓明科技有限公司 | Method and system for locating mobile terminal |
| CN105467382A (en) * | 2015-12-31 | 2016-04-06 | 南京信息工程大学 | SVM (Support Vector Machine)-based multi-sensor target tracking data fusion algorithm and system thereof |
| CN105528912A (en) * | 2016-01-20 | 2016-04-27 | 中山大学 | Indoor parking stall recommendation method and system |
| CN106604229A (en) * | 2016-12-27 | 2017-04-26 | 东南大学 | Indoor positioning method based on manifold learning and improved support vector machine |
| CN107333238A (en) * | 2017-07-03 | 2017-11-07 | 杭州电子科技大学 | A kind of indoor fingerprint method for rapidly positioning based on support vector regression |
| CN108989976A (en) * | 2018-06-04 | 2018-12-11 | 华中师范大学 | Fingerprint positioning method and system in a kind of wisdom classroom |
| US20200142045A1 (en) * | 2018-06-04 | 2020-05-07 | Central China Normal University | Fingerprint positioning method and system in smart classroom |
| CN109029429A (en) * | 2018-09-04 | 2018-12-18 | 电子科技大学 | Multi-categorizer overall situation dynamic fusion localization method based on WiFi and earth magnetism fingerprint |
| CN110012428A (en) * | 2019-05-22 | 2019-07-12 | 合肥工业大学 | A kind of indoor orientation method based on WiFi |
Non-Patent Citations (2)
| Title |
|---|
| 周锦等: "基于KNN-SVM算法的室内定位系统设计", 《华中科技大学学报(自然科学版)》 * |
| 施涛涛等: "基于支持向量机的混合相似度室内指纹定位算法", 《计算机工程》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115022795A (en) * | 2021-02-18 | 2022-09-06 | 诺基亚技术有限公司 | RF fingerprinting mapping update |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111885700B (en) | 2022-04-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109242844B (en) | Pancreatic cancer tumor automatic identification system based on deep learning, computer equipment and storage medium | |
| CN107333238B (en) | Indoor fingerprint rapid positioning method based on support vector regression | |
| CN102821465B (en) | WLAN (Wireless Local Area Network) indoor positioning method based on subregion information entropy gain | |
| WO2019062734A1 (en) | Indoor positioning method and device based on wi-fi hot spots | |
| CN111327377B (en) | Method, device, equipment and storage medium for field intensity prediction | |
| CN110536245A (en) | A kind of indoor wireless positioning method and system based on deep learning | |
| CN111757464B (en) | Region contour extraction method and device | |
| Suroso et al. | Location fingerprint technique using Fuzzy C-Means clustering algorithm for indoor localization | |
| CN113015093B (en) | Indoor wireless positioning method based on three-dimensional depth residual error neural network | |
| CN111885700B (en) | Mobile terminal positioning method and device combined with support vector machine | |
| Caso et al. | Frequentist inference for WiFi fingerprinting 3D indoor positioning | |
| CN111698695A (en) | LTE fingerprint type positioning method based on neural network | |
| Torres-Sospedra et al. | Scalable and efficient clustering for fingerprint-based positioning | |
| CN105657653B (en) | Indoor positioning method based on fingerprint data compression | |
| Lei et al. | DeFLoc: Deep learning assisted indoor vehicle localization atop FM fingerprint map | |
| CN108111976B (en) | WiFi signal fingerprint data optimization method and device | |
| CN111757257B (en) | Dynamic fuzzy matching indoor positioning method for overcoming equipment difference | |
| CN106488554B (en) | Fingerprint database establishing method and system | |
| CN111669702A (en) | Coarse and fine matching positioning method based on WIFI address and field intensity | |
| CN108445443B (en) | KNN-based fingerprint point clustering method | |
| CN111639712A (en) | Positioning method and system based on density peak clustering and gradient lifting algorithm | |
| US20150287176A1 (en) | Method and appratus for hybrid test pattern generation for opc model calibration | |
| Shen et al. | A sparse manifold learning approach to robust indoor positioning based on Wi-Fi RSS fingerprinting | |
| CN108055638A (en) | Obtain method, apparatus, computer-readable medium and the equipment of target location | |
| US11894880B2 (en) | Automatic fine-grained radio map construction and adaption |
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 |