CN114364016B - AOA and RSSI-based indoor positioning method, device and system based on weighted fingerprint feature matching - Google Patents

AOA and RSSI-based indoor positioning method, device and system based on weighted fingerprint feature matching Download PDF

Info

Publication number
CN114364016B
CN114364016B CN202111568603.8A CN202111568603A CN114364016B CN 114364016 B CN114364016 B CN 114364016B CN 202111568603 A CN202111568603 A CN 202111568603A CN 114364016 B CN114364016 B CN 114364016B
Authority
CN
China
Prior art keywords
reference point
fingerprint
vector
aoa
rssi
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
Application number
CN202111568603.8A
Other languages
Chinese (zh)
Other versions
CN114364016A (en
Inventor
高迎宾
孙达
夏玮玮
张亦农
燕锋
沈连丰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing Xijueshuo Information Technology Co ltd
Original Assignee
Nanjing Xijueshuo Information Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Nanjing Xijueshuo Information Technology Co ltd filed Critical Nanjing Xijueshuo Information Technology Co ltd
Priority to CN202111568603.8A priority Critical patent/CN114364016B/en
Publication of CN114364016A publication Critical patent/CN114364016A/en
Application granted granted Critical
Publication of CN114364016B publication Critical patent/CN114364016B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W64/00—Locating users or terminals or network equipment for network management purposes, e.g. mobility management
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
    • G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
    • G01S5/0252—Radio frequency fingerprinting
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
    • G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
    • G01S5/0295—Proximity-based methods, e.g. position inferred from reception of particular signals
    • G01S5/02955—Proximity-based methods, e.g. position inferred from reception of particular signals by computing a weighted average of the positions of the signal transmitters
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
    • G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
    • G01S5/04—Position of source determined by a plurality of spaced direction-finders
    • 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
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mathematical Physics (AREA)
  • Theoretical Computer Science (AREA)
  • Collating Specific Patterns (AREA)

Abstract

The invention discloses an AOA and RSSI weighted fingerprint feature matching indoor positioning method, device and system, wherein the positioning method trains each feature in a reference fingerprint vector corresponding to each reference point in a fingerprint library by using a random forest machine learning method to obtain the weight of each feature so as to form a reference point weight vector; the reference fingerprint vector corresponding to each reference point is obtained based on signal strength RSSI and arrival angle AOA of the reference point; vector processing is carried out on the signal strength RSSI and the arrival angle AOA of the to-be-detected point, and an online fingerprint vector is obtained by utilizing a natural breakpoint classification method; based on the online fingerprint vector and the reference point weight vector, a matching algorithm is adopted to obtain a matched reference point; and adding the coordinates corresponding to the obtained reference points, and averaging to obtain the predicted coordinates of the to-be-measured points. The invention improves the precision of the fingerprint method, reduces the number of base stations used, and improves the positioning accuracy.

Description

AOA and RSSI-based indoor positioning method, device and system based on weighted fingerprint feature matching
Technical Field
The invention belongs to the field of Bluetooth indoor fingerprint positioning, and particularly relates to an AOA and RSSI weighted fingerprint feature matching indoor positioning method, device and system.
Background
In recent years, the development of positioning technology is very popular, and particularly with the rapid development of the internet of things and artificial intelligence, the market demand and desire for positioning technology also rise. The indoor space which is used as a space for a long time for human beings is complex in electromagnetic environment and geographic environment, and the requirements on positioning technology are relatively high. Bluetooth has the characteristics of high precision, high concurrency, low power consumption, low cost, high compatibility and the like, and is one of important carriers of indoor positioning technology. In an indoor environment, the method for acquiring more accurate position information through an indoor positioning technology with high precision and high reliability is a current research hot spot.
Traditional bluetooth fingerprint indoor location is through gathering the RSSI information of a plurality of bluetooth beacons of arranging in indoor environment, after handling, generates the fingerprint map, and the matching stage compares with the fingerprint in the fingerprint map, obtains the positional information of matching. However, this method only uses RSSI values as an index of fingerprint matching, generally requires at least 4 bluetooth beacons, and if the accuracy is to be improved, a large number of bluetooth beacons need to be arranged, and the accuracy is only in the order of meters. An AOA & AOD method is proposed in the Bluetooth 5.1 version, and by monitoring that carrier phases reaching different antennas are different, a receiving end can detect an included angle between an arrival signal and a normal of the receiving end, although in principle, only two or more transmitting ends are needed, the position of the receiving end can be calculated through calculation. But still face several challenges at present, including: signal reflection interference, antenna array errors, complex environments, and antenna directivity disturbances, several challenges require much effort from the microwave radio frequency and algorithmic parts.
Disclosure of Invention
Aiming at the problems, the invention provides an AOA and RSSI weighted fingerprint feature matching indoor positioning method, device and system, which improve the precision of a fingerprint method, reduce the number of used base stations and improve the positioning accuracy.
In order to achieve the technical purpose and achieve the technical effect, the invention is realized by the following technical scheme:
in a first aspect, the present invention provides an indoor positioning method based on AOA and RSSI weighted fingerprint feature matching, including:
training each feature in the reference fingerprint vectors corresponding to each reference point in the fingerprint library by using a machine learning method of a random forest to obtain the weight of each feature, thereby forming a reference point weight vector; the reference fingerprint vector corresponding to each reference point is obtained based on signal strength RSSI and arrival angle AOA of the reference point;
vector processing is carried out on the signal strength RSSI and the arrival angle AOA of the to-be-detected point, and an online fingerprint vector is obtained by utilizing a natural breakpoint classification method;
based on the online fingerprint vector and the reference point weight vector, a matching algorithm is adopted to obtain a matched reference point;
and adding the coordinates corresponding to the obtained reference points, and averaging to obtain the predicted coordinates of the to-be-measured points.
Optionally, the reference fingerprint vector corresponding to each reference point is obtained by the following method:
vector sorting is carried out on the RSSI and the AOA of the signal intensity collected from different base stations at the reference point;
and obtaining a reference fingerprint vector corresponding to the reference point by utilizing a natural breakpoint classification method according to the vector-sorted data.
Optionally, the expression of the data of vector finishing is:
wherein a is i (n) represents an nth angle of arrival AOA value acquired by an ith base station at a kth reference point, r i (N) represents an nth signal strength, RSSI, value acquired by an ith base station at an kth reference point, i=1, 2.
Optionally, the method for obtaining the reference fingerprint vector corresponding to each reference point includes:
carrying out amplitude limiting treatment on the collected arrival angle AOA, and then obtaining a treated AOA vector and an RSSI vector through a 2 delta criterion together with the signal strength RSSI;
carrying out iterative classification on an arrival angle AOA or a signal strength RSSI acquired by each base station at each reference point by using a natural breakpoint classification method, calculating intra-class variance, overall variance and the ratio of the intra-class variance and the overall variance, and selecting a classification condition with the minimum ratio;
counting the number of each category, calculating the weight of the base station, carrying out weighted calculation by using the boundary average value of the category to obtain the AOA and RSSI reference fingerprint values of the base station at the reference point, and sorting the AOA and RSSI reference fingerprint values into reference fingerprint vectors corresponding to the reference point.
Optionally, the method for calculating the reference point weight vector includes:
adding corresponding label categories to the data in the fingerprint database to obtain a training data set;
training the weight of each feature through a random forest, and forming a reference point weight vector based on the weight of each feature; the characteristics refer to the angle of arrival AOA and signal strength RSSI of each base station.
Optionally, the calculation formula of the reference point weight vector is:
VIM＝[VIM 1 ,…,VIM j ,…,VIM 2B ]
VIM jm ＝GI m -GI l -GI r
wherein VIM is a reference point weight vector, VIM j Is the characteristic x j Weights of (2), VIM jt Is the characteristic x j Weights at t-th tree, VIM jm GI, which is the difference between Gini indexes before and after node m branching l And GI r Gini index, GI, respectively representing left and right subtrees branched by node m m Is given by the index of Gini,is an estimate of the probability that the sample at the mth node belongs to the kth class, corresponding to the kth reference point.
Optionally, the method for acquiring the matched reference point includes:
calculating a weighted Euclidean distance based on the online fingerprint vector and a reference point weight vector;
sorting the weighted Euclidean distances, and setting a threshold value to obtain a first Euclidean distance set smaller than or equal to the threshold value;
and selecting a second Euclidean distance set which is smaller than or equal to the average value of the first Euclidean distance set from the first Euclidean distance set, and further obtaining a reference point corresponding to each distance in the second Euclidean distance set.
Optionally, the calculation formula of the weighted euclidean distance is as follows:
d＝[x 1 ,x 2 ,…,x 2B-1 ,x 2B ]
wherein d k The fingerprint vector of the kth reference point is VIM, the reference point weight vector is VIM, d is the data of the online fingerprint vector, and x is VIM 1 ,x 2 ,…,x 2B-1 ,x 2B Is the fingerprint value of the point to be measured.
Optionally, the formulas of the first euclidean distance set, the second euclidean distance set and the average value of the first euclidean distance set are as follows:
wherein epsilon is a set threshold, phi is a first set of Euclidean distances, phi is a second set of Euclidean distances,for the first Euclidean distance set average value, U is the Euclidean distance in the first Euclidean distance setNumber of separation dis k For weighting euclidean distances.
Optionally, the method for calculating the predicted coordinates of the point to be measured includes:
and based on the coordinates corresponding to the second Euclidean distance set, averaging the coordinates, and calculating the predicted coordinates of the to-be-measured point.
Optionally, the calculation formula of the predicted coordinates of the point to be measured is:
wherein,and->Respectively the abscissa and the ordinate of the predicted coordinates of the to-be-measured point, Z is the number of elements in the second Euclidean distance set phi, Z epsilon (1, Z) and x z And y z The abscissa of the z-th element is indicated.
In a second aspect, the present invention provides an indoor positioning device based on AOA and RSSI weighted fingerprint feature matching, including:
the reference point weight vector obtaining module is used for training each feature in the reference fingerprint vectors corresponding to the reference points in the fingerprint library by utilizing a machine learning method of a random forest to obtain the weight of each feature so as to form a reference point weight vector; the reference fingerprint vector corresponding to each reference point is obtained based on signal strength RSSI and arrival angle AOA of the reference point;
the online fingerprint vector acquisition module is used for carrying out vector processing on the signal strength RSSI and the arrival angle AOA of the to-be-detected point and obtaining an online fingerprint vector by utilizing a natural breakpoint classification method;
the reference point acquisition module is used for acquiring a matched reference point by adopting a matching algorithm based on the online fingerprint vector and the reference point weight vector;
and the prediction module is used for averaging after adding the coordinates corresponding to the obtained reference points to obtain the predicted coordinates of the to-be-measured points.
In a third aspect, the invention provides an AOA and RSSI weighted fingerprint feature-based indoor positioning system, which comprises a storage medium and a processor;
the storage medium is used for storing instructions;
the processor is configured to operate in accordance with the instructions to perform the method according to any one of the first aspects.
Compared with the prior art, the invention has the beneficial effects that:
the invention expands the data type from single RSSI value to RSSI value and AOA value based on fingerprint matching method, enriches the information category of fingerprint; secondly, training each feature by using a random forest to obtain importance scores of the features; then, a data processing method is provided based on a natural breakpoint classification method, and a fingerprint vector is obtained; in the online matching stage, the KNN method is improved, the matched reference points are obtained by calculating the weighted Euclidean distance, and finally the predicted coordinates are calculated. The method greatly improves the precision of the fingerprint method, reduces the number of base stations used, and improves the positioning accuracy.
Drawings
In order that the invention may be more readily understood, a more particular description of the invention will be rendered by reference to specific embodiments that are illustrated in the appended drawings, in which:
FIG. 1 is a flow chart of an indoor positioning method based on AOA and RSSI weighted fingerprint feature matching of the invention;
FIG. 2 is a schematic view of an indoor positioning method based on AOA and RSSI weighted fingerprint feature matching in the present invention;
figure 3 is a model of the importance of random forest training features.
Detailed Description
The present invention will be described in further detail with reference to the following examples in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the detailed description and specific examples are intended for purposes of illustration only and are not intended to limit the scope of the invention.
The principle of application of the invention is described in detail below with reference to the accompanying drawings.
Example 1
The embodiment of the invention provides an AOA and RSSI weighted fingerprint feature matching indoor positioning method, which comprises the following steps:
(1) Training each feature in the reference fingerprint vectors corresponding to each reference point in the fingerprint library by using a machine learning method of a random forest to obtain the weight of each feature, thereby forming a reference point weight vector; the reference fingerprint vector corresponding to each reference point is calculated based on the signal strength (Received Signal Strength Indication, RSSI) and the Angle Of Arrival (AOA) Of the reference point;
(2) Vector processing is carried out on the signal strength RSSI and the arrival angle AOA of the to-be-detected point, and an online fingerprint vector is obtained by utilizing a natural breakpoint classification method;
(3) Based on the online fingerprint vector and the reference point weight vector, a matching algorithm is adopted to obtain a matched reference point;
(4) Adding the coordinates corresponding to the obtained reference points, and averaging to obtain predicted coordinates of the to-be-measured points
Fig. 1 is a flow chart of an indoor positioning method based on AOA and RSSI weighted fingerprint feature matching, and in the practical application process, the whole method is divided into two stages: the method comprises an offline stage and an online stage, wherein the offline stage comprises four parts of data acquisition, data processing, fingerprint library establishment and random forest training, and the online stage comprises four parts of data acquisition, data processing, data matching and calculation of predicted coordinates;
fig. 2 is a schematic view of an indoor positioning method based on AOA and RSSI weighted fingerprint feature matching of the present invention, first we arrange base stations (M1, M2, M3, M4) at the vertices of a rectangular area, and the base stations are connected to a computer. And selecting reference points at certain intervals in the area, placing sub-equipment (S) at the reference points, constructing a chain with the base station in sequence, transmitting signals, collecting AOA and RSSI values by the base station, storing the AOA and RSSI values to a computer, and establishing a database and training weights in an off-line stage at the computer. During an online stage, randomly selecting one point in each small square as an online point (namely a to-be-measured point), sequentially acquiring data, carrying out data processing at a computer end, then carrying out matching, and calculating a predicted coordinate;
the off-line stage specifically comprises the following steps:
and a data acquisition step: assuming that K reference points and B base stations are selected in the area to be examined, the N AOA and RSSI value vectors acquired at each reference point are sorted into the following form:
wherein a is i (n) represents an nth angle of arrival AOA value acquired by an ith base station at a kth reference point, r i (N) represents an nth signal strength, RSSI, value acquired by an ith base station at an kth reference point, i=1, 2.
And a data processing step: the invention provides a filtering method based on a natural breakpoint classification method to determine fingerprints of a traditional fingerprint library, and the following 3 steps of processing all data (namely signal strength RSSI and arrival angle AOA) acquired from different base stations at reference points: firstly, carrying out amplitude limiting treatment on an AOA value, and then removing about 5% of extreme values by using a 2 delta criterion on the AOA and RSSI; then using natural breakpoint classification to divide the data into several classes which are different as far as possible; and finally, weighting the data to obtain a representative numerical value serving as a reference fingerprint value. The method comprises the following specific steps:
(1) For the collected AOA value, the AOA value is limited to-90 degrees to 90 degrees, and then 95% of effective data is screened out through a 2 delta criterion together with the RSSI value, wherein the specific formula is as follows:
where j=1, 2,..2B, and j is an integer, x can be expressed as an AOA or RSSI value of any one base station at any one reference point, N is the number of acquired data, μ is an average value, and σ is a standard;
obtaining the processed AOA vector a i ＝[a i (1),…,a i (n),…,a i (N)] T RSSI vector r i ＝[r i (1),…,r i (n),…,r i (N)] T 。
(2) Natural fracture classification is a clustering algorithm that will a i And r i The elements of (2) are divided into G groups. For convenience, the present invention defines x j ＝a i ∪r i . First, calculate x j Is denoted as a; then, for x j Ordering all elements of (a) and dividing into different groups in an iterative manner; the Mean Square Error (MSE) of each group is calculated and their sum is obtained, denoted S, with the following specific calculation formula:
wherein,is x in group g j Average value of (N), g=1, 2,.. g Is x in group g j Number of (n), x g (q) is the q-th value in the g-th group.
Finally, the evaluation index E in the formula (5) is calculated, the optimal grouping of the maximum E is determined, and the boundary of the G group is obtained as b= { b 1 ,…,b g ,…,b G+1 }。
(3) After data are divided into G groups by a natural breakpoint classification method, a filtered fingerprint value x is calculated f The method comprises the following specific steps:
i. get the boundary b= { b of group G 1 ,…,b g ,…,b G+1 }, wherein b 1 ,…,b g ,…,b G+1 Is the boundary value of each packet;
according to x j (n)∈[b g ,b g+1 ) Calculating the number of data belonging to each group to obtain c= (c) 1 ,…,c g ,…,c G ) Wherein g=g, x j (n)＝b G+1 Belonging to group G, wherein c 1 ,…,c g ,…,c G Representing the number of groups to which the data belongs;
calculating the weight w g Average mu g Fingerprint value x f The formula is as follows:
the final fingerprint vector for the kth reference point is as follows
Wherein a is fi Is a i Fingerprint value of vector, r fi Is r i The fingerprint values of the vectors can be collectively expressed as x fj J=1, 2,..2B, and j is an integer.
Fingerprint library establishment: after deriving the fingerprint vectors for all reference points, the entire fingerprint library can be expressed as:
D＝[d 1 ,d 2 ,…,d K-1 ,d K ] T (8)
Random forest training weights: as shown in fig. 3, equation (2) is first converted into the following form:
wherein f k (n) represents all x at the kth reference point j (n),j∈[1,2B]. Random forest from f k (n) repeatedly randomly selecting the subsampled set S 1 ,…,S t ,…,S T . The classification trees are then generated to form a random forest, where T is the number of trees, t=1, 2.
The Gini index calculation formula is:
according to Gini index, at each node of the random forest, a variable that minimizes the base index is selected as a branch variable and split into left and right nodes. Wherein,is an estimate of the probability that the sample at the mth node belongs to the kth class, corresponding to the kth reference point. Feature x j Score statistics expressed as VIM j (variable importance measure). The difference between the Gini indices before and after node m branching is denoted as VIM jm ：
VIM jm ＝GI m -GI l -GI r (11)
Wherein GI is l And GI r The Gini indexes of the left and right subtrees branched by the node m are respectively represented.
If the characteristic x j M times in the t-th tree, then feature x j The weights (i.e., importance) at the t-th tree are:
feature x j The Gini weight, i.e. the normalized value of importance over all nodes, of the entire random forest is defined as:
and calculating the Gini index of each feature, and giving the corresponding importance to obtain the required feature importance vector:
VIM＝[VIM 1 ,…,VIM j ,…,VIM 2B ](14)
The online stage specifically comprises the following steps:
on-line matching: processing the online collected data by using the same data processing mode as the offline stage to obtain the data d= [ x ] of the online fingerprint vector 1 ,x 2 ,…,x 2B-1 ,x 2B ]。
The weights of the conventional KNN and WKNN methods on all fingerprint values are equal, and the weight vector VIM of the reference points of all data trained by the random forest is used to obtain a weighted Euclidean distance formula as follows:
and after the Euclidean distances of all the reference points are ordered, setting a threshold epsilon, and obtaining an Euclidean distance set phi smaller than or equal to the threshold. Assuming that there are U Euclidean distances in φ, the average Euclidean distance is calculated and expressed asThen selecting less than or equal to the average value +.>The formula is as follows:
calculating predicted coordinates: and calculating the predicted coordinates of the to-be-measured point by using the coordinate sets corresponding to the Euclidean distance values, wherein the formula is as follows:
wherein Z is the number of elements in the second Euclidean distance set phi, Z is E (1, Z), x z And y z Respectively representing the abscissa of the z-th element,and->Respectively the abscissa and the ordinate of the predicted coordinates of the point to be measured.
Example 2
Based on the same inventive concept as embodiment 1, an embodiment of the present invention provides an indoor positioning device based on AOA and RSSI weighted fingerprint feature matching, including:
the reference point weight vector obtaining module is used for training each feature in the reference fingerprint vectors corresponding to the reference points in the fingerprint library by utilizing a machine learning method of a random forest to obtain the weight of each feature so as to form a reference point weight vector; the reference fingerprint vector corresponding to each reference point is obtained based on signal strength RSSI and arrival angle AOA of the reference point;
the online fingerprint vector acquisition module is used for carrying out vector processing on the signal strength RSSI and the arrival angle AOA of the to-be-detected point and obtaining an online fingerprint vector by utilizing a natural breakpoint classification method;
the reference point acquisition module is used for acquiring a matched reference point by adopting a matching algorithm based on the online fingerprint vector and the reference point weight vector;
and the prediction module is used for averaging after adding the coordinates corresponding to the obtained reference points to obtain the predicted coordinates of the to-be-measured points.
The remainder was the same as in example 1.
Example 3
The embodiment of the invention provides an AOA and RSSI weighted fingerprint feature-based indoor positioning system, which comprises a storage medium and a processor, wherein the storage medium is used for storing the received signals;
the storage medium is used for storing instructions;
the processor is configured to operate in accordance with the instructions to perform the method according to any one of embodiment 1.
The foregoing has shown and described the basic principles and main features of the present invention and the advantages of the present invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, and that the above embodiments and descriptions are merely illustrative of the principles of the present invention, and various changes and modifications may be made without departing from the spirit and scope of the invention, which is defined in the appended claims. The scope of the invention is defined by the appended claims and equivalents thereof.

Claims (11)

1. An indoor positioning method based on AOA and RSSI weighted fingerprint feature matching is characterized by comprising the following steps:
training each feature in the reference fingerprint vectors corresponding to each reference point in the fingerprint library by using a machine learning method of a random forest to obtain the weight of each feature, thereby forming a reference point weight vector; the reference fingerprint vector corresponding to each reference point is obtained based on signal strength RSSI and arrival angle AOA of the reference point;
vector processing is carried out on the signal strength RSSI and the arrival angle AOA of the to-be-detected point, and an online fingerprint vector is obtained by utilizing a natural breakpoint classification method;
based on the online fingerprint vector and the reference point weight vector, a matching algorithm is adopted to obtain a matched reference point;
adding the coordinates corresponding to the obtained reference points, and averaging to obtain predicted coordinates of the to-be-measured points;
the reference fingerprint vector corresponding to each reference point is obtained by the following method:
vector sorting is carried out on the RSSI and the AOA of the signal intensity collected from different base stations at the reference point;
for the data subjected to vector arrangement, a natural breakpoint classification method is utilized to obtain a reference fingerprint vector corresponding to a reference point;
the method for obtaining the reference fingerprint vector corresponding to each reference point comprises the following steps:
carrying out amplitude limiting treatment on the collected arrival angle AOA, and then obtaining a treated AOA vector and an RSSI vector through a 2 delta criterion together with the signal strength RSSI;
carrying out iterative classification on an arrival angle AOA or a signal strength RSSI acquired by each base station at each reference point by using a natural breakpoint classification method, calculating intra-class variance, overall variance and the ratio of the intra-class variance and the overall variance, and selecting a classification condition with the minimum ratio;
counting the number of each category, calculating the weight of the base station, carrying out weighted calculation by using the boundary average value of the category to obtain the AOA and RSSI reference fingerprint values of the base station at the reference point, and sorting the AOA and RSSI reference fingerprint values into reference fingerprint vectors corresponding to the reference point.
2. The indoor positioning method based on AOA and RSSI weighted fingerprint feature matching as set forth in claim 1, wherein the method comprises the steps of: the expression of the vector-sorted data is as follows:
wherein a is i (n) represents an nth angle of arrival AOA value acquired by an ith base station at a kth reference point, r i (N) represents an nth signal strength, RSSI, value acquired by an ith base station at an kth reference point, i=1, 2.
3. The indoor positioning method based on AOA and RSSI weighted fingerprint feature matching as set forth in claim 1, wherein the method comprises the steps of: the calculation method of the reference point weight vector comprises the following steps:
adding corresponding label categories to the data in the fingerprint database to obtain a training data set;
training the weight of each feature through a random forest, and forming a reference point weight vector based on the weight of each feature; the characteristics refer to the angle of arrival AOA and signal strength RSSI of each base station.
4. The indoor positioning method based on AOA and RSSI weighted fingerprint feature matching as claimed in claim 3, wherein the calculation formula of the reference point weight vector is:
VIM＝[VIM 1 ,...,VIM j ,...,VIM 2B ]
VIM jm ＝GI m -GI l -GI r
wherein VIM is a reference point weight vector, VIM j Is the characteristic x j Weights of (2), VIM jt Is the characteristic x j Weights at t-th tree, VIM jm GI, which is the difference between Gini indexes before and after node m branching l And GI r Gini index, GI, respectively representing left and right subtrees branched by node m m Is given by the index of Gini,is an estimate of the probability that the sample at the mth node belongs to the kth class, corresponding to the kth reference point.
5. The indoor positioning method based on AOA and RSSI weighted fingerprint feature matching according to claim 1, wherein the method for acquiring the matched reference point comprises the following steps:
calculating a weighted Euclidean distance based on the online fingerprint vector and a reference point weight vector;
sorting the weighted Euclidean distances, and setting a threshold value to obtain a first Euclidean distance set smaller than or equal to the threshold value;
and selecting a second Euclidean distance set which is smaller than or equal to the average value of the first Euclidean distance set from the first Euclidean distance set, and further obtaining a reference point corresponding to each distance in the second Euclidean distance set.
6. The indoor positioning method based on AOA and RSSI weighted fingerprint feature matching as set forth in claim 5, wherein: the calculation formula of the weighted Euclidean distance is as follows:
d＝[x 1 ,x 2 ,...,x 2B-1 ,x 2B ]
wherein d k The fingerprint vector of the kth reference point is VIM, the reference point weight vector is VIM, d is the data of the online fingerprint vector, and x is VIM 1 ,x 2 ,...,x 2B-1 ,x 2B Is the fingerprint value of the point to be measured.
7. The indoor positioning method based on AOA and RSSI weighted fingerprint feature matching as set forth in claim 6, wherein: the formula of the average values of the first Euclidean distance set, the second Euclidean distance set and the first Euclidean distance set is as follows:
wherein epsilon is a set threshold, phi is a first set of Euclidean distances, phi is a second set of Euclidean distances,for the first Euclidean distance set levelThe average value, U, is the number of Euclidean distances in the first Euclidean distance set, dis k For weighting euclidean distances.
8. The indoor positioning method based on AOA and RSSI weighted fingerprint feature matching as set forth in claim 7, wherein: the calculation method of the predicted coordinates of the point to be measured comprises the following steps:
and based on the coordinates corresponding to the second Euclidean distance set, averaging the coordinates, and calculating the predicted coordinates of the to-be-measured point.
9. The indoor positioning method based on AOA and RSSI weighted fingerprint feature matching as set forth in claim 8, wherein: the calculation formula of the predicted coordinates of the point to be measured is as follows:
wherein,and->Respectively the abscissa and the ordinate of the predicted coordinates of the to-be-measured point, Z is the number of elements in the second Euclidean distance set phi, Z epsilon (1, Z) and x z And y z The abscissa of the z-th element is indicated.
10. An indoor positioner of weighted fingerprint feature matching based on AOA and RSSI, characterized by comprising:
the reference point weight vector obtaining module is used for training each feature in the reference fingerprint vectors corresponding to the reference points in the fingerprint library by utilizing a machine learning method of a random forest to obtain the weight of each feature so as to form a reference point weight vector; the reference fingerprint vector corresponding to each reference point is obtained based on signal strength RSSI and arrival angle AOA of the reference point;
the online fingerprint vector acquisition module is used for carrying out vector processing on the signal strength RSSI and the arrival angle AOA of the to-be-detected point and obtaining an online fingerprint vector by utilizing a natural breakpoint classification method;
the reference point acquisition module is used for acquiring a matched reference point by adopting a matching algorithm based on the online fingerprint vector and the reference point weight vector;
the prediction module is used for averaging after adding the coordinates corresponding to the obtained reference points to obtain predicted coordinates of the to-be-measured points;
the reference fingerprint vector corresponding to each reference point is obtained by the following method:
vector sorting is carried out on the RSSI and the AOA of the signal intensity collected from different base stations at the reference point;
for the data subjected to vector arrangement, a natural breakpoint classification method is utilized to obtain a reference fingerprint vector corresponding to a reference point;
the method for obtaining the reference fingerprint vector corresponding to each reference point comprises the following steps:
carrying out amplitude limiting treatment on the collected arrival angle AOA, and then obtaining a treated AOA vector and an RSSI vector through a 2 delta criterion together with the signal strength RSSI;
carrying out iterative classification on an arrival angle AOA or a signal strength RSSI acquired by each base station at each reference point by using a natural breakpoint classification method, calculating intra-class variance, overall variance and the ratio of the intra-class variance and the overall variance, and selecting a classification condition with the minimum ratio;
counting the number of each category, calculating the weight of the base station, carrying out weighted calculation by using the boundary average value of the category to obtain the AOA and RSSI reference fingerprint values of the base station at the reference point, and sorting the AOA and RSSI reference fingerprint values into reference fingerprint vectors corresponding to the reference point.
11. An indoor positioning system based on AOA and RSSI weighted fingerprint feature matching is characterized in that: including a storage medium and a processor;
the storage medium is used for storing instructions;
the processor is operative to perform the method according to any one of claims 1-9, in accordance with the instructions.
CN202111568603.8A 2021-12-21 2021-12-21 AOA and RSSI-based indoor positioning method, device and system based on weighted fingerprint feature matching Active CN114364016B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111568603.8A CN114364016B (en) 2021-12-21 2021-12-21 AOA and RSSI-based indoor positioning method, device and system based on weighted fingerprint feature matching

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111568603.8A CN114364016B (en) 2021-12-21 2021-12-21 AOA and RSSI-based indoor positioning method, device and system based on weighted fingerprint feature matching

Publications (2)

Publication Number Publication Date
CN114364016A CN114364016A (en) 2022-04-15
CN114364016B true CN114364016B (en) 2023-12-22

Family

ID=81100493

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111568603.8A Active CN114364016B (en) 2021-12-21 2021-12-21 AOA and RSSI-based indoor positioning method, device and system based on weighted fingerprint feature matching

Country Status (1)

Country Link
CN (1) CN114364016B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2000069198A1 (en) * 1999-05-06 2000-11-16 Cell-Loc Inc. Wireless location system
CN108770057A (en) * 2018-03-30 2018-11-06 博彦科技股份有限公司 The determination method and fingerprint positioning method in predetermined fingerprint library, device, storage medium
CN112996106A (en) * 2021-03-07 2021-06-18 中国人民解放军陆军工程大学 Honeycomb-removing large-scale MIMO system positioning method
CN113259837A (en) * 2021-04-29 2021-08-13 西安电子科技大学 Indoor positioning method based on angle estimation and fingerprint positioning algorithm
CN113747360A (en) * 2021-07-28 2021-12-03 南京邮电大学 Indoor positioning method based on mixed visible light and BLE

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2000069198A1 (en) * 1999-05-06 2000-11-16 Cell-Loc Inc. Wireless location system
CN108770057A (en) * 2018-03-30 2018-11-06 博彦科技股份有限公司 The determination method and fingerprint positioning method in predetermined fingerprint library, device, storage medium
CN112996106A (en) * 2021-03-07 2021-06-18 中国人民解放军陆军工程大学 Honeycomb-removing large-scale MIMO system positioning method
CN113259837A (en) * 2021-04-29 2021-08-13 西安电子科技大学 Indoor positioning method based on angle estimation and fingerprint positioning algorithm
CN113747360A (en) * 2021-07-28 2021-12-03 南京邮电大学 Indoor positioning method based on mixed visible light and BLE

Also Published As

Publication number Publication date
CN114364016A (en) 2022-04-15

Similar Documents

Publication Publication Date Title
CN108989976A (en) Fingerprint positioning method and system in a kind of wisdom classroom
CN109951807A (en) Fusion RSS and CSI indoor orientation method based on WiFi signal
CN103889051B (en) Indoor WLAN fingerprint positioning methods based on AP ID filterings and Kalman filtering
CN109115205A (en) A kind of indoor fingerprint positioning method and system based on geomagnetic sensor array
CN106646338A (en) Rapidly accurate indoor location method
CN107360548A (en) Indoor article localization method and system based on RFID
CN109029429B (en) WiFi and geomagnetic fingerprint based multi-classifier global dynamic fusion positioning method
CN104185275A (en) Indoor positioning method based on WLAN
CN109348416B (en) Fingerprint indoor positioning method based on binary k-means
CN113259837B (en) Indoor positioning method based on angle estimation and fingerprint positioning algorithm
CN111366820A (en) Pattern recognition method, device, equipment and storage medium for partial discharge signal
CN111586605B (en) KNN indoor target positioning method based on adjacent weighted self-adaptive k value
Gomes et al. Clustering and hierarchical classification for high-precision RFID indoor location systems
Siyang et al. WKNN indoor Wi-Fi localization method using k-means clustering based radio mapping
CN111465093A (en) Fingerprint positioning method and device, electronic equipment and computer storage medium
CN112350745B (en) Sorting method of frequency hopping communication radio station
CN106250787B (en) A kind of unbundling formula activity recognition method of low-cost high robust
CN113271539B (en) Indoor target positioning method based on improved CNN model
CN110858289B (en) Grain moisture detection method and system based on deep learning
CN111757257B (en) Dynamic fuzzy matching indoor positioning method for overcoming equipment difference
CN114364016B (en) AOA and RSSI-based indoor positioning method, device and system based on weighted fingerprint feature matching
CN117098067A (en) Multi-mode deep learning indoor positioning method based on gradient fusion
CN111191475A (en) Passive behavior identification method based on UHF RFID
CN114679683B (en) Indoor intelligent positioning method based on derived fingerprint migration
CN116028885A (en) Fall detection method, system, equipment and medium based on channel state information

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