CN112616184B - Mobile equipment position estimation method based on multi-base station channel state information fusion - Google Patents
Mobile equipment position estimation method based on multi-base station channel state information fusion Download PDFInfo
- Publication number
- CN112616184B CN112616184B CN202011460684.5A CN202011460684A CN112616184B CN 112616184 B CN112616184 B CN 112616184B CN 202011460684 A CN202011460684 A CN 202011460684A CN 112616184 B CN112616184 B CN 112616184B
- Authority
- CN
- China
- Prior art keywords
- base station
- channel
- state information
- sample
- channel state
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- 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
- H04W64/003—Locating users or terminals or network equipment for network management purposes, e.g. mobility management locating network equipment
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L25/00—Baseband systems
- H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
- H04L25/0202—Channel estimation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L25/00—Baseband systems
- H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
- H04L25/0202—Channel estimation
- H04L25/024—Channel estimation channel estimation algorithms
- H04L25/0242—Channel estimation channel estimation algorithms using matrix methods
Abstract
The invention discloses a mobile equipment position estimation method based on multi-base station channel state information fusion, and belongs to the technical field of radio mobile positioning. Including a base station collecting non-real time channel state information samples from a mobile device; extracting angular domain power spectrum channel characteristics; calculating the difference between the samples according to the angular domain power spectrum channel characteristics to obtain a single base station sample difference matrix; the data fusion node fuses the multiple single base station sample difference degree matrixes to obtain a global sample difference degree matrix; the data fusion node constructs a channel map for the sample data set according to the global sample difference degree matrix and calibrates the channel map; and the base station collects the real-time channel state information from the mobile equipment, performs matching positioning by combining the sample data set and the channel map, and acquires the real-time position information of the mobile equipment. The invention does not need to have accurate time synchronization among base stations and on-site survey, reduces the manual participation and realizes the automatic operation of the mobile positioning system.
Description
Technical Field
The invention belongs to the technical field of radio mobile positioning, and particularly relates to a mobile equipment position estimation method based on multi-base station channel state information fusion.
Background
The mobile position information acquisition technology is a key support for building a smart city, and can provide basic data for application scenes such as automatic management of wireless network spectrum resources, automatic driving, lighting of smart cities, smart logistics, smart security, mobile internet social contact, social situation perception and the like by sensing and acquiring space-time position information of crowds, logistics, vehicles and other various internet-of-things devices.
With the application of technologies such as large-scale multiple input multiple output (massive MIMO), ultra-dense networking, ultra wide band and the like in a 5G mobile network, the 5G base station has stronger radio frequency sensing capability and mobile positioning capability, and practical possibility is provided for high-precision user positioning and tracking by using a cellular network. On one hand, the existing TDOA-based cellular network mobile positioning method needs higher time synchronization precision among base stations, and has poor working effect in complex environments such as dense urban areas, indoor environments and the like; on the other hand, the traditional positioning method based on radio frequency fingerprint identification needs to acquire a large amount of radio frequency fingerprint training data with position tags, and needs to perform a large amount of manual site survey and data collection work to realize seamless coverage positioning in a large area.
Disclosure of Invention
The technical problem is as follows: aiming at the problems that the existing mobile equipment positioning method needs higher time synchronization precision among base stations and needs manual site survey to realize effective positioning, the invention provides a mobile equipment position estimation method based on multi-base-station channel state information fusion, which can realize mobile positioning and tracking under the conditions that the base stations are not high in time synchronization precision and unsupervised.
The technical scheme is as follows: the invention discloses a mobile equipment position estimation method based on multi-base station channel state information fusion, which comprises the following steps:
step S1: a base station collects non-real-time channel state information samples from mobile equipment and constructs a wireless channel state information sample data set;
step S2: extracting angular domain power spectrum channel characteristics from the channel state information samples;
step S3: calculating the difference between the samples according to the angular domain power spectrum channel characteristics to obtain a single base station sample difference matrix;
step S4: the data fusion node fuses the multiple single base station sample difference degree matrixes to obtain a global sample difference degree matrix;
step S5: the data fusion node performs manifold learning on the sample data set according to the global sample difference degree matrix to obtain a two-dimensional channel map, and then calibrates the channel map according to the base station position and the base station antenna orientation information;
Step S6: and the base station collects the real-time channel state information from the mobile equipment, performs matching positioning by combining the sample data set and the channel map, and acquires the real-time position information of the mobile equipment.
Further, the channel is an uplink channel from the mobile device to the base station, and the channel state information includes a channel covariance matrix estimated by the base station from the pilot signal, signal-to-noise ratio information, a mobile device ID, and a timestamp.
Further, step S1 includes:
step S11: estimating the signal-to-noise ratio, extracting the mobile equipment ID and the timestamp corresponding to the sample, and determining a unique sample label according to the mobile equipment number, the ID and the timestamp information;
step S12: the base station estimates channel response according to the pilot signal;
step S13: the channel covariance matrix is estimated in the time or frequency domain.
Further, step S2 includes:
step S21: performing eigenvalue decomposition on the channel covariance matrix;
step S22: estimating the number of main multipath components according to the distribution of the characteristic values, so that the sum of the power of the main multipath is larger than the set percentage of the total power of the channel;
step S23: and calculating the power and the arrival angle of the main multipath component, pairing the power and the arrival angle, and constructing the power spectrum characteristic of the angular domain.
Further, in step S23, when calculating the power and the arrival angle of the main multipath component, a multi-signal classification algorithm, or a space alternative general expectation maximization algorithm, or a rotation invariant subspace algorithm is used.
Further, step S3 includes:
step S31: calculating Cartesian coordinates of the mirror image emission points according to the angular domain power spectrum characteristics;
step S32: performing cluster analysis on all generated mirror image emission points;
step S33: and calculating the difference between the channel samples according to the clustering result, and constructing a single-base-station channel sample data difference matrix.
Further, step S4 includes:
step S41: the base station reports channel sample data, a sample label, a channel sample data difference matrix and signal-to-noise ratio information to the data fusion node;
step S42: the data fusion node fuses the diversity factor matrixes reported by the base stations by taking the signal-to-noise ratio as weight;
step S43: and constructing a global sample difference degree matrix.
Further, step S5 includes:
step S51: performing manifold learning on the channel samples according to the global sample difference degree matrix to obtain a two-dimensional channel map;
step S52: and calibrating the channel map according to the position of the base station and the orientation information of the base station antenna.
Further, step S6 includes:
step S61: acquiring channel state information of the mobile equipment in real time;
step S62: calculating the difference between the real-time channel sample and the historical sample, and selecting a plurality of most similar historical samples from the sample data set;
step S63: the center of gravity of the plurality of historical samples is calculated as a mobile device position estimation result.
Further, the base station is a multi-antenna base station.
Has the advantages that: compared with the prior art, the invention has the following advantages:
(1) reducing the requirement of mobile positioning on time synchronization of multiple base stations
Compared with the traditional TDOA-based positioning technology in a cellular network, the method for positioning the position of the mobile equipment needs accurate time synchronization among base stations, only needs the adjacent base stations to know that a channel information data sample is a pilot signal sent by the same equipment at the same position, and can be determined through the identification number of the mobile equipment and the time stamp of the data, thereby effectively reducing the requirement of the mobile positioning on the time synchronization of multiple base stations.
(2) Supporting non-user cooperative positioning
In the method of the invention, as long as the mobile equipment has communication requirements, the network end can realize the positioning and tracking of the mobile equipment by passively collecting channel state information data, and the process does not need the cooperation of users, thereby effectively reducing the energy consumption required by equipment positioning and supporting the non-user cooperative positioning.
(3) Operable under non-line-of-sight channel conditions
The mobile equipment position positioning method provided by the invention is based on the channel characteristic difference degree measurement, and even under the non-line-of-sight condition, the channel characteristics corresponding to the close positions are similar, so that the method can work under the non-line-of-sight channel condition.
(4) No need for manual site survey:
the invention realizes position estimation by using unsupervised learning, does not need a data sample with a position label, can greatly reduce the manual participation and realizes the automatic operation of the mobile positioning system.
Drawings
FIG. 1 is a schematic diagram of a multi-base-station mobile positioning scenario according to the present invention;
FIG. 2 is a flow chart of a mobile device position estimation method based on multi-base station channel state information fusion according to the present invention;
FIG. 3 is a diagram illustrating channel map scaling in an embodiment of the invention;
FIG. 4 is a diagram of a simulation verification scenario in an embodiment of the present invention;
FIG. 5 is an uncorrected channel map generated during simulation verification in an embodiment of the present invention;
FIG. 6 is a corrected channel map generated during simulation verification in an embodiment of the present invention;
FIG. 7 is a diagram illustrating a result of a mobile location estimation performed during simulation verification according to an embodiment of the present invention;
Fig. 8 is a graph of the average positioning accuracy of one hundred random points obtained during simulation verification in the embodiment of the present invention.
Detailed Description
The invention is further described below with reference to the following examples and the drawings.
FIG. 1 illustrates a multi-base station mobile positioning scenario in which the present invention is applicable; the number of the base stations is B, the base stations are provided with a plurality of antennas, and the serial number of the base stations is B ∈ {1,2, …, B }; each base station is configured with linear arrays with the same number of array elements, and the number of the array elements is N; the mobile device has a single antenna and is an omnidirectional antenna; in the operation process of a wireless network, a mobile device sends a pilot signal with fixed power to a base station for network uplink channel estimation; the pilot signal contains device Identification (ID) information and its transmission time information.
Fig. 2 shows a specific flow of an embodiment of the present invention, which includes the following specific steps:
step S1: the base station collects non-real-time channel state information samples from the mobile equipment and constructs a wireless channel state information sample set.
Wherein, the channel is an uplink channel from the mobile device to the base station, and the channel state information comprises a channel covariance matrix estimated by the base station according to the pilot signal, signal-to-noise ratio information, a mobile device ID and a timestamp. Specifically, step S1 includes the steps of:
Step S11: and estimating the signal-to-noise ratio, extracting the mobile equipment ID and the time stamp corresponding to the sample, and determining a unique sample label according to the mobile equipment number, the ID and the time stamp information. Specifically, a pilot signal transmitted by the mobile device can be received by a plurality of base stations at the same time, and the base station receiving the pilot signal records the mobile device ID information and the timestamp corresponding to the pilot signal; meanwhile, the ratio of Reference Signal Received Power (RSRP) to system noise power is used as the estimated value of the signal-to-noise ratio gamma and is recorded as
Wherein sigma 2 Is the noise power.
Step S12: the base station estimates the channel response from the pilot signal. Specifically, the result is recorded as a complex vector h of size N × 1.
Step S13: the covariance matrix of the multi-antenna channel is estimated in the time or frequency domain. Specifically, a set of channel response vectors { h } of a single frequency point over a continuous period of time is collected i :i＝t 1 ,t 2 ,...,t T Or a set of channel response vectors (h) at different frequency points at the same time j :j＝f 1 ,f 2 ,...,f F }; since the base station has multiple antennas, the channel covariance matrix Y can be estimated as
Or
The number of the channel response vectors is equal to or larger than T and equal to N or equal to F and equal to N.
Each base station estimates a channel covariance matrix as a channel sample according to a pilot signal transmitted by a certain mobile device at a certain time, and a channel sample data set collected by the base station b can be expressed as 
Where i is a sample label uniquely determined by the mobile device ID and the sample collection time, S b The set of labels for the samples collected by base station b, the corresponding set of signal-to-noise ratios collected by base station b is noted
Through the identification number of the mobile equipment and the time stamp of the data, the adjacent base stations know that the channel information data sample is from the pilot signal transmitted by the same equipment at the same position, so that the accurate time synchronization among the base stations is not required like the TDOA-based positioning technology in the traditional cellular network, and the requirement of mobile positioning on the time synchronization among multiple base stations is reduced.
Step S2: extracting angular domain power spectrum channel characteristics from the channel state information samples; specifically, in the embodiment of the present invention, the method includes the following steps:
step S21: and carrying out eigenvalue decomposition on the channel covariance matrix.
Specifically, the channel covariance matrix Y ═ U × diag (λ) 1 ,λ 2 ,...,λ N )×U H Wherein λ is 1 ,λ 2 ,...,λ N The eigenvalues are arranged from large to small, where H represents the transpose of the matrix.
Step S22: the number of dominant multipath components is estimated based on the distribution of the eigenvalues such that the sum of the power of the dominant multipath is greater than a set percentage of the total power of the channel. In bookIn an embodiment of the invention, the set percentage is selected to be 80%. Specifically, the magnitude of the eigenvalue is positively correlated with the power of the multipath, when the number L of the multipath is smaller than the number N of the antennas, L of the eigenvalues have larger values, and other eigenvalues have smaller values; in a practical system, the number of multipaths is unknown, and it is likely that the number is larger than the number of antennas; selecting at least M larger characteristic values lambda from the characteristic values 1 ,λ 2 ,…,λ M So that
Step S23: and calculating the power and the arrival angle of the main multipath component, pairing the power and the arrival angle, and constructing an angular domain power spectrum characteristic.
Specifically, in the present embodiment, a multiple signal classification algorithm is adopted to select the sum of the eigenvalues λ 1 ,λ 2 ,…,λ M The corresponding eigenvectors constitute the signal subspace U s ＝[u 1 ,u 2 ,…,u M ]The remaining feature vectors constitute the noise subspace U n ＝[u M+1 ,…,u N ](ii) a The pseudo-power spectrum of the angular domain is calculated as follows:
wherein a is BS (phi) is the base station antenna array response vector; searching M peak values and arrival angles phi corresponding to the M peak values from p (phi) 1 ,φ 2 ,...,φ M Constructing angular domain power spectrum channel characteristics
By the method, the base station b can receive the data from
In-process acquisition angular domain power spectrum channel feature set
Angular domain power spectrum channel characteristics
The power and angle of arrival of the M channel paths in a sample are described, corresponding to the M mirror emission points (if the channel contains a direct-view path, then the true emission point).
It should be noted that, in this step, not only the multiple signal classification algorithm may be used to extract the power spectrum feature of the angular domain of the channel, but in other embodiments, other algorithms, such as a space-alternative general expectation-maximization algorithm or a rotation-invariant subspace algorithm, may also be used.
Step S3: and calculating the difference between the samples according to the angular domain power spectrum channel characteristics to obtain a single base station sample difference matrix. Specifically, in this embodiment, the following steps are included:
Step S31: and calculating the Cartesian coordinates of the mirror image emission points according to the angular domain power spectrum characteristics. According to the characteristics of angular domain power spectrum
Calculating the Cartesian coordinates of the mirror image emission points, wherein the specific formula is as follows:
P＝Ψ(F)＝[p 1 ,p 2 ,...,p M ]，
wherein
ρ(λ m ) Is a non-linear scaling function to compensate for the scattering loss of multipath.
By the method, the base station b can receive the data from
In-process acquisition of mirror image emission point cloud
Step S32: and performing cluster analysis on all generated mirror image emission points. In particular, in embodiments of the present invention, density-based clustering is employedAlgorithm, in the present embodiment, DBSCAN algorithm pair is adopted
The points in (1) are clustered, and a clustering label is marked on each point, and the points with the same clustering label belong to the same cluster.
In other embodiments, all the generated mirror emission points may be clustered by using a K-means hierarchical clustering algorithm.
Step S33: and calculating the difference between the channel samples according to the clustering result, and constructing a single-base-station channel sample data difference matrix. In particular, for a sample set
Two samples of
And
set the mirror image emission point sets as
And
if it is
And
if the mirror image emission points belong to the same cluster, a pair of points (p) is searched in the two point sets l ,q k ) So that p is l ,q k Belong to the same cluster, and (p) l ,q k ) The value for which the corresponding two path power values are smaller is largest among all possible combinations,
and
has a degree of difference of Δ (Y) n ,Y m )＝||p l -q k L. If it is
And
if no point belongs to the same cluster, then
And
has a degree of difference of
Using the above method, the base station b aligns its sample set
Calculating the difference between any two different samples to obtain a single base station sample difference matrix D b The elements of which are composed of
Is composed of (i) n, m ∈ S b 。
Step S4: the data fusion node fuses a plurality of single base station sample difference degree matrixes to obtain a global sample difference degree matrix, and specifically comprises the following steps:
step S41: the base station reports channel sample data, a sample label, a channel sample data difference matrix and signal-to-noise ratio information to the data fusion node, and the fusion node obtains a group of difference matrixes
Signal to noise ratio
And set of sample tag sets S b } b∈{1,2,...,B} 。
Step S42: and the data fusion node fuses the diversity factor matrixes reported by the base stations by taking the signal-to-noise ratio as weight. Specifically, first pair { S b } b∈{1,2,...,B} Combining to obtain S ═ S 1 ∪S 2 ...∪S B (ii) a Taking a sample p, q belongs to S, and calculating the global sample difference:
wherein 1(A) is an indicative function,
the weight of the difference degree between the sample p and the sample q observed by the base station b in fusion; if one of p and q is not in the acceptance range of the base station b, the weight is 0; otherwise, the smaller value of the signal-to-noise ratios corresponding to the p and q samples is taken as the weight.
Step S43: and constructing a global sample difference degree matrix. Specifically, for all p, q ∈ S, D is calculated according to the method step 42 p,q Constructing a global sample disparity matrix D with elements D p,q 。
In the embodiment of the invention, the single base station channel sample data difference degree matrix and the global sample difference degree matrix are respectively constructed through the steps S3 and S4, so that the measurement is carried out by utilizing the channel characteristic difference degree, and the channel characteristics corresponding to the close positions are similar even under the non-line-of-sight condition, so that the method can work under the non-line-of-sight channel condition.
Step S5: and the data fusion node performs manifold learning on the sample data set according to the global sample difference degree matrix to obtain a two-dimensional channel map, and then calibrates the channel map according to the base station position and the base station antenna orientation information. Specifically, the method comprises the following steps:
step S51: performing manifold learning on the channel samples according to the sample global difference matrix D to obtain a two-dimensional channelMap { z k } k∈S Wherein z is k And (3) two-dimensional map coordinates corresponding to a sample label k, wherein in the embodiment of the invention, a t-distribution random proximity embedding (t-SNE) algorithm is adopted in the manifold learning method, and of course, in other embodiments, a Sammon mapping or a laplacian feature mapping algorithm may also be adopted.
Step S52: channel map { z ] according to base station position and base station antenna orientation information k } k∈S Calibrating to obtain corrected two-dimensional channel map
Specifically, as shown in fig. 3, a plurality of samples having direct-view paths to two or more base stations are selected as anchor points, and the geographic positions of the anchor points are calculated by using the DOA positioning method and combining the base station positions and the antenna orientation information. In other embodiments, the position information of the anchor point may also be obtained according to satellite positioning or other positioning methods. Then from { z k } k∈S Finding out the subset corresponding to the anchor Point, solving a translation parameter and a rotation parameter by an Iterative Closest Point algorithm (ICP), and carrying out the { z + operation on the translation parameter and the rotation parameter k } k∈S All the points are translated and rotated to obtain corrected two-dimensional channel map
And after the two-dimensional channel map is obtained and the calibration is finished, matching positioning can be implemented.
Step S6: and the base station collects the real-time channel state information from the mobile equipment, performs matching positioning by combining the sample data set and the channel map, and acquires the real-time position information of the mobile equipment. The method specifically comprises the following steps:
step S61: acquiring channel state information of the mobile equipment in real time; specifically, the mobile device transmits a pilot signal, and the base station receiving the pilot signal collects channel data Ω b 。
Step S62: calculating a real-time channel sample omega b And historical sample data set
Degree of difference of (2)
Selecting the most similar K (K) from the sample data set>0) A history sample
Step S63: according to the channel map
Calculating the gravity center of the K historical samples
As a result of the mobile device position estimation, the position estimation of the mobile device is thereby completed.
It can be seen from step S6 that, as long as the mobile device has a communication requirement, the network end can realize the positioning and tracking of the mobile device by passively collecting channel state information data, and this process does not require cooperation of a user, thereby effectively reducing energy consumption required by device positioning and supporting non-user cooperative positioning. And when the position of the mobile equipment is positioned, the position estimation is realized by using unsupervised learning, a data sample with a position label is not needed, the manual participation degree can be greatly reduced, and the automatic operation of the mobile positioning system is realized.
The effects of the embodiments of the present invention are described below in conjunction with simulation verification. FIG. 4 is a diagram of a simulation verification scenario, including urban building geographical location information, building material information, street information; the number of base stations in a scene is 10, and the number of antennas of each base station is 16; the mobile device has a single antenna; the position information of 2000 samples is randomly generated according to the street information, and the channel response from the sample points to the base station is generated by adopting a ray tracing algorithm.
Fig. 5 is an uncorrected channel map generated during simulation verification, and fig. 6 is a corrected channel map generated, and it can be seen that the uncorrected channel map contains relative position information and topology information between samples in a sample set, and the corrected channel map gives absolute position information of the samples.
Fig. 7 shows a moving position estimation result obtained by real-time matching positioning during simulation verification, where the number K of adjacent samples is 20, and it can be seen that the gravity center of the 20 samples is used as the position estimation result and the actual position error is less than 4 meters.
FIG. 8 is a graph of average positioning accuracy of one hundred random points obtained with the present embodiment under different K values in a simulation verification scenario; it can be seen that the method provided by the embodiment can achieve the positioning accuracy of a meter level; when K is larger than 10, the average positioning precision is stabilized to be about 3 meters.
Through simulation verification, the method can be used for accurately positioning the mobile equipment. Compared with the traditional multi-base-station TDOA (time difference of arrival) positioning method, the mobile equipment positioning method does not need accurate time synchronization among the base stations and can work under the conditions of multipath and non-line-of-sight; compared with the traditional radio frequency fingerprint positioning method, the method does not need on-site survey, can greatly reduce the manual participation and realize the automatic operation of the mobile positioning system.
The above examples are only preferred embodiments of the present invention, it should be noted that: it will be apparent to those skilled in the art that various modifications and equivalents can be made without departing from the spirit of the invention, and it is intended that all such modifications and equivalents fall within the scope of the invention as defined in the claims.
Claims (10)
1. The mobile equipment position estimation method based on the multi-base station channel state information fusion is characterized by comprising the following steps of:
step S1: a base station collects non-real-time channel state information samples from mobile equipment and constructs a wireless channel state information sample data set;
step S2: extracting angular domain power spectrum channel characteristics from the channel state information samples;
step S3: calculating the difference between the samples according to the angular domain power spectrum channel characteristics to obtain a single base station sample difference matrix;
step S4: the data fusion node fuses the multiple single base station sample difference degree matrixes to obtain a global sample difference degree matrix;
step S5: the data fusion node performs manifold learning on the sample data set according to the global sample difference degree matrix to obtain a two-dimensional channel map, and then calibrates the channel map according to the base station position and the base station antenna orientation information;
Step S6: and the base station collects the real-time channel state information from the mobile equipment, performs matching positioning by combining the sample data set and the channel map, and acquires the real-time position information of the mobile equipment.
2. The method of claim 1, wherein the channel is an uplink channel from the mobile device to the base station, and the channel state information includes a channel covariance matrix estimated by the base station from the pilot signal, signal-to-noise ratio information, a mobile device ID, and a timestamp.
3. The method for estimating the position of a mobile device based on the fusion of multi-base station channel state information as claimed in claim 2, wherein step S1 comprises:
step S11: estimating the signal-to-noise ratio, extracting the mobile equipment ID and the timestamp corresponding to the sample, and determining a unique sample label according to the mobile equipment number, the ID and the timestamp information;
step S12: the base station estimates channel response according to the pilot signal;
step S13: the channel covariance matrix is estimated in the time or frequency domain.
4. The method for estimating the position of a mobile device based on the fusion of multi-base station channel state information as claimed in claim 3, wherein the step S2 comprises:
Step S21: performing eigenvalue decomposition on the channel covariance matrix;
step S22: estimating the number of main multipath components according to the distribution of the characteristic values, so that the sum of the power of the main multipath is larger than the set percentage of the total power of the channel;
step S23: and calculating the power and the arrival angle of the main multipath component, pairing the power and the arrival angle, and constructing the power spectrum characteristic of the angular domain.
5. The method for estimating the position of a mobile device based on the fusion of multi-base station channel state information as claimed in claim 4, wherein in step S23, when the power and the arrival angle of the main multipath component are calculated, a multi-signal classification algorithm, or a space alternative general expectation maximization algorithm, or a rotation invariant subspace method is adopted.
6. The method for estimating the position of a mobile device based on the fusion of multi-base station channel state information as claimed in claim 4, wherein the step S3 comprises:
step S31: calculating Cartesian coordinates of the mirror image emission points according to the angular domain power spectrum characteristics;
step S32: performing cluster analysis on all generated mirror image emission points;
step S33: and calculating the difference between the channel samples according to the clustering result, and constructing a single-base-station channel sample data difference matrix.
7. The method for estimating the position of a mobile device based on the fusion of multi-base station channel state information as claimed in claim 6, wherein step S4 comprises:
step S41: the base station reports channel sample data, a sample label, a channel sample data difference matrix and signal-to-noise ratio information to the data fusion node;
step S42: the data fusion node fuses the diversity factor matrixes reported by the base stations by taking the signal-to-noise ratio as weight;
step S43: and constructing a global sample difference degree matrix.
8. The method for estimating the position of a mobile device based on the fusion of multi-base station channel state information as claimed in claim 7, wherein step S5 comprises:
step S51: performing manifold learning on the channel samples according to the global sample difference degree matrix to obtain a two-dimensional channel map;
step S52: and calibrating the channel map according to the position of the base station and the orientation information of the base station antenna.
9. The method for estimating the position of a mobile device based on the fusion of multi-base station channel state information as claimed in claim 8, wherein step S6 comprises:
step S61: acquiring channel state information of the mobile equipment in real time;
step S62: calculating the difference between the real-time channel sample and the historical sample, and selecting a plurality of most similar historical samples from the sample data set;
Step S63: the center of gravity of the plurality of historical samples is calculated as a mobile device position estimation result.
10. The method for estimating the position of a mobile device based on the fusion of the channel state information of multiple base stations as claimed in any one of claims 1 to 9, wherein the base station is a multi-antenna base station.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011460684.5A CN112616184B (en) | 2020-12-11 | 2020-12-11 | Mobile equipment position estimation method based on multi-base station channel state information fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011460684.5A CN112616184B (en) | 2020-12-11 | 2020-12-11 | Mobile equipment position estimation method based on multi-base station channel state information fusion |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112616184A CN112616184A (en) | 2021-04-06 |
| CN112616184B true CN112616184B (en) | 2022-07-29 |
Family
ID=75233565
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011460684.5A Active CN112616184B (en) | 2020-12-11 | 2020-12-11 | Mobile equipment position estimation method based on multi-base station channel state information fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112616184B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113595711B (en) * | 2021-08-05 | 2023-03-14 | 东南大学 | Indoor communication and positioning integrated method and system based on wireless local area network |
| CN115808701B (en) * | 2023-02-09 | 2023-05-26 | 南京航空航天大学 | Multi-machine collaborative wide area space channel map mapping device and method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2007003680A1 (en) * | 2005-06-30 | 2007-01-11 | Nokia Corporation | A method and device for pilot power measurement in a receiver |
| CN106656376A (en) * | 2016-12-30 | 2017-05-10 | 北京邮电大学 | Cooperative spectrum sensing method based on feature value consistent estimation |
| CN109082984A (en) * | 2018-06-29 | 2018-12-25 | 浙江大学城市学院 | A kind of road abnormality detection model based on window division and dynamic time warping |
| CN110809904A (en) * | 2017-07-01 | 2020-02-18 | 英特尔公司 | Method and apparatus for vehicle radio communication |
-
2020
- 2020-12-11 CN CN202011460684.5A patent/CN112616184B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2007003680A1 (en) * | 2005-06-30 | 2007-01-11 | Nokia Corporation | A method and device for pilot power measurement in a receiver |
| CN106656376A (en) * | 2016-12-30 | 2017-05-10 | 北京邮电大学 | Cooperative spectrum sensing method based on feature value consistent estimation |
| CN110809904A (en) * | 2017-07-01 | 2020-02-18 | 英特尔公司 | Method and apparatus for vehicle radio communication |
| CN109082984A (en) * | 2018-06-29 | 2018-12-25 | 浙江大学城市学院 | A kind of road abnormality detection model based on window division and dynamic time warping |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112616184A (en) | 2021-04-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ye et al. | Neural-network-assisted UE localization using radio-channel fingerprints in LTE networks | |
| CN111479231B (en) | Indoor fingerprint positioning method for millimeter wave large-scale MIMO system | |
| CN108696932B (en) | Outdoor fingerprint positioning method using CSI multipath and machine learning | |
| CN107318084B (en) | Fingerprint positioning method and device based on optimal similarity | |
| US11089563B2 (en) | 5G-signal-based DOA fingerprint-based positioning method | |
| CN109672973B (en) | Indoor positioning fusion method based on strongest AP | |
| CN106851571B (en) | Decision tree-based rapid KNN indoor WiFi positioning method | |
| CN103874118B (en) | Radio Map bearing calibrations in WiFi indoor positionings based on Bayesian regression | |
| CN108717175B (en) | Indoor fingerprint positioning method based on region division and sparse support vector regression | |
| CN107015198B (en) | Indoor positioning method based on irregular arrangement of antennas | |
| CN112616184B (en) | Mobile equipment position estimation method based on multi-base station channel state information fusion | |
| US11419088B2 (en) | Radiation source positioning method based on radio spectrum monitoring big data processing | |
| CN110381436A (en) | A kind of quick fingerprint positioning method based on the mono- station system of extensive MIMO | |
| CN109068267B (en) | Indoor positioning method based on LoRa SX1280 | |
| CN103501538A (en) | Indoor positioning method based on multipath energy fingerprints | |
| CN107727095A (en) | 3D indoor orientation methods based on spectral clustering and weighting reverse transmittance nerve network | |
| CN107305246B (en) | Positioning method and device based on received signal strength indication | |
| Adege et al. | Applying Deep Neural Network (DNN) for large-scale indoor localization using feed-forward neural network (FFNN) algorithm | |
| CN106255203B (en) | The localization method of terminal RSRP disparity compensation based on MDS | |
| CN113207089A (en) | Position fingerprint positioning method based on CSI and crowdsourcing migration self-calibration updating | |
| Du et al. | KF-KNN: Low-cost and high-accurate FM-based indoor localization model via fingerprint technology | |
| Li et al. | Location estimation in large indoor multi-floor buildings using hybrid networks | |
| Mukhtar et al. | Machine learning-enabled localization in 5g using lidar and rss data | |
| Ma et al. | LTE user equipment RSRP difference elimination method using multidimensional scaling for LTE fingerprint-based positioning system | |
| Mondal et al. | An efficient grid-based RF fingerprint positioning algorithm for user location estimation in heterogeneous small cell networks |
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 |