Passive positioning method based on Wi-Fi beam forming feedback information

Abstract

The invention provides a passive positioning method based on a wireless fidelity (Wi-Fi) beam forming feedback information (BFI). The passive positioning method comprises the following steps of: firstly, carrying out passive positioning on the basis of the Wi-Fi beam forming feedback information (BFI); the method comprises the following steps: firstly, pre-determining the position of a transceiver, and respectively extracting BFIs of a direct path and a target reflection path of the transceiver; compared with extraction of channel state information (CSI), the BFI is inherent information in a Wi-Fi communication protocol, special acquisition hardware does not need to be additionally deployed, the BFI can be directly extracted from the communication process, the hardware cost and the deployment threshold of a positioning system are greatly reduced, and large-scale popularization and application are easier to realize. And secondly, by constructing a BFI covariance matrix, high-dimensional matrix operation in a traditional MUSIC algorithm is converted into low-dimensional matrix processing, and the calculation complexity is remarkably reduced. And finally, constructing a spectral function to respectively obtain an angle of departure (AOD) of a direct path of the transceiver and a target reflection path and a flight time difference of the two paths, substituting the parameters into the elliptic constraint model for calculation, and finally solving to obtain an estimated position of the target. According to the method designed by the invention, the real-time performance and the deployment flexibility of the system are greatly improved under the condition that the calculation complexity is remarkably reduced and the support of specific hardware is not needed.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of realizing perception by using commercial WiFi equipment, and particularly relates to a passive positioning method based on Wi-Fi beamforming feedback information (BFI), which realizes positioning of a target by processing beamforming feedback information. BACKGROUND

[0002] With the popularity of the Internet of Things and intelligent devices, precise positioning technology has become an indispensable key support in smart city and industrial monitoring scenarios. In the field of smart city, precise positioning technology can realize real-time monitoring and intelligent scheduling of traffic flow, and relieve urban congestion. In the industrial monitoring scenario, it can realize precise positioning and state tracking of production equipment, and guarantee production safety and efficiency. Traditional positioning schemes rely on channel state information (CSI), which can achieve high-precision positioning to a certain extent, but the high computational complexity and hardware dependency make the device cost high, the power consumption large, and it is difficult to meet the demand of large-scale deployment.

[0003] In recent years, beamforming feedback information (BFI) has attracted widespread attention from the industry and academia due to its standardization compatibility and lightweight characteristics. Existing research has successfully applied BFI to perception tasks such as respiratory monitoring and activity recognition. For example, in the smart home scenario, BFI technology is used to realize non-contact monitoring of human respiratory frequency and activity state, providing data support for health management and security monitoring. However, the potential of BFI in the positioning field has not been fully tapped. BFI extracts core channel information through a compression feedback mechanism, which can significantly reduce data volume and computational overhead compared to traditional CSI, providing a new idea for low-cost positioning. SUMMARY

[0004] The purpose of the present application is to provide a passive positioning method based on Wi-Fi beamforming feedback information in a wireless communication system. This method directly uses the BFI inherent in the existing communication protocol, without relying on special hardware to collect CSI. By extracting BFI directly, the computational complexity is significantly reduced, the system real-time performance is improved, and no additional hardware cost is added.

[0005] The passive positioning method based on Wi-Fi beamforming feedback information comprises the following steps:

[0006] Step one: assuming that M antennas constitute a uniform linear array to sample the spatial signal transmitted by a single receiving antenna, under far-field conditions, L signal propagation paths are incident to the uniform linear array; assuming that N subcarriers of beamforming feedback information (BFI) can be collected on each antenna, a two-dimensional matrix is constructed using information from different antennas and different subcarriers where 1≤m≤M, 1≤n≤N, represents the BFI value obtained on the Mth antenna and the Nth subcarrier;

[0007] Step two: the two-dimensional matrix is smoothed along the subcarrier dimension and the antenna dimension in turn to obtain an enhanced matrix where K is the smoothing window length of the subcarrier dimension, and the parameter P is the smoothing window length of the antenna dimension;

[0008] Step three: calculate V e the autocorrelation matrix of

[0009] and perform eigenvalue decomposition to solve the noise subspace and the signal subspace;

[0010] Step four: construct a noise eigenvector matrix U n from the eigenvectors in the noise subspace, and calculate the spectral function formed by U n and the vector a(θ,τ):

[0011]

[0012] where (·) H is the conjugate transpose symbol, and a(θ,τ) is the direction vector:

[0013]

[0014] where a(θ)=[1 e j2πfdsinθ / c … e j2πf(P-1)dsinθ / c ] T , b(τ)=[1 e j2πΔfτ … e j2π(K-1)Δfτ ] T , θ and τ are parameters to be estimated, which are the departure angle and the relative time of flight of the transmitted signal, respectively;

[0015] Step five: peak searching of the spectrum function P(θ,τ); in ideal condition, the estimated θ should be the real angle of departure (AOD) of each path; and the estimated τ should be the real time of flight (TOF) of each path plus an offset, and the offset is consistent for each path;

[0016] Step six: when the estimated target is in the non-straight path of the transceiver, and the transceiver position is determined in advance, first, the parameters are estimated by the above algorithm to obtain the AOD of the straight path of the transceiver and the reflection path of the target, and the time delay difference Δτ of the two paths, wherein Δτ=|τ1-τ2|, τ1 is the relative time of flight of the reflection path, and τ2 is the relative time of flight of the straight path; then, the parameters are substituted into the elliptical constraint model for calculation, and finally the specific position of the estimated target is solved.

[0017] The passive positioning method based on Wi-Fi beamforming feedback information according to the claim, the step two, the received two-dimensional V matrix is converted into an enhanced matrix Comprising the following steps:

[0018] The matrix V is represented as:

[0019]

[0020] Firstly, the BFI is smoothed along the subcarrier dimension, and the smoothed matrix is:

[0021]

[0022] Wherein, The smoothed matrix of the subcarrier of the mth antenna is represented as K, and K is the smoothing window length of the subcarrier dimension;

[0023] Then, the BFI matrix is smoothed along the antenna dimension, and the smoothed matrix is:

[0024]

[0025] Wherein, the parameter P is the smoothing window length of the antenna dimension;

[0026] The passive positioning method based on Wi-Fi beamforming feedback information according to the claim, the step two, solving the noise subspace and the signal subspace, comprising the following steps:

[0027] Under single fast snap, the autocorrelation matrix of the enhanced BFI data matrix is calculated as:

[0028]

[0029] Then, the autocorrelation matrix is subjected to Eigenvalue Decomposition (EVD):

[0030]

[0031] In the formula, Σ s , Σ n respectively represent diagonal matrices composed of eigenvalues of the signal subspace and the noise subspace, U s , U n respectively represent eigenvectors corresponding to the eigenvalues of the signal subspace and the noise subspace; the eigenvalues are arranged in descending order as follows:

[0032] λ1≥λ2≥…≥λ L > λ L+1 =…= λ M = σ 2

[0033] In the formula, λ1, λ2, …, λ L are eigenvalues of the signal subspace, λ L+1 , …; λ M are eigenvalues of the noise subspace.

[0034] Advantages

[0035] The application provides a passive positioning method based on Wi-Fi beamforming feedback information. Firstly, the positions of transceivers are determined in advance, and BFI of direct paths and reflection paths of targets of the transceivers is extracted respectively. Compared with CSI, BFI is inherent information naturally existing in a Wi-Fi communication protocol, and can be directly extracted from a communication process without deploying special collection hardware, so that the hardware cost and deployment threshold of the positioning system are greatly reduced, and large-scale popularization and application are more easily realized. Secondly, by constructing a BFI covariance matrix, high-dimensional matrix operation in a traditional MUSIC algorithm is converted into low-dimensional matrix processing, so that the calculation complexity is significantly reduced. Finally, a spectrum function is constructed to obtain AOD of the direct paths and the reflection paths of the transceivers and the time delay difference Δτ of the two paths, and these parameters are substituted into an elliptical constraint model for calculation, and finally the specific position of the target is solved.

[0036] The method designed in the application greatly improves the real-time performance and deployment flexibility of the system under the condition of significantly reducing the calculation complexity and without specific hardware support. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 A block diagram of the passive positioning method based on Wi-Fi beamforming feedback information of the application DETAILED DESCRIPTION

[0038] The application aims to provide a passive positioning method based on Wi-Fi beamforming feedback information in a wireless communication system.

[0039] The passive positioning method based on Wi-Fi beamforming feedback information comprises the following steps:

[0040] Step one: assuming that a uniform linear array composed of M antennas samples the spatial signal transmitted by a single receiving antenna, under far-field conditions, L signal propagation paths are incident to the uniform linear array; assuming that N beamforming feedback information (BFI) of subcarriers can be collected on each antenna, a two-dimensional matrix is constructed by using the information of different antennas and different subcarriers wherein 1≤m≤M, 1≤n≤N, represents the BFI value obtained on the Mth antenna and the Nth subcarrier;

[0041] Step two: the two-dimensional matrix is smoothed along the subcarrier dimension and the antenna dimension in sequence to obtain an enhanced matrix wherein K is the smoothing window length of the subcarrier dimension, and P is the smoothing window length of the antenna dimension;

[0042] Step three: calculating V e the autocorrelation matrix of V and performing eigenvalue decomposition to solve the noise subspace and the signal subspace;

[0043] Step four: constructing a noise eigenvector matrix U n from the eigenvectors in the noise subspace, and calculating the spectral function formed by U n and the vector a(θ,τ):

[0044]

[0045] wherein (·) H is the conjugate transpose symbol, and a(θ,τ) is a direction vector:

[0046]

[0047] wherein a(θ)=[1 e j2πfdsinθ / c … e j2πf(P-1)dsinθ / c ] T , b(τ)=[1 e j2πΔfτ … e j2π(K-1)Δfτ ]T θ, τ are the parameters to be estimated, which are the angle of departure and the relative time of flight of the transmitted signal, respectively;

[0048] Step five: peak searching of the spectrum function P(θ, τ); under ideal conditions, the estimated θ should be the real angle of departure (AOD) of each path, and the estimated τ should be the real time of flight (TOF) of each path plus an offset, which is consistent for each path;

[0049] Step six: when the estimation target is in the non-straight path of the transceiver, and the position of the transceiver is predetermined, first, the parameters are estimated by the above algorithm to obtain the AOD of the straight path of the transceiver and the reflection path of the target, and the time delay difference Δτ of the two paths, wherein Δτ = |τ1-τ2|, τ1 is the relative time of flight of the reflection path, and τ2 is the relative time of flight of the straight path; then, the parameters are substituted into the elliptical constraint model for calculation, and finally, the specific position of the estimation target is solved.

[0050] The passive positioning method based on Wi-Fi beamforming feedback information according to the claim, the step two, the received two-dimensional V matrix is converted into an enhanced matrix Comprising the following steps:

[0051] The matrix V is represented as:

[0052]

[0053] Firstly, the BFI is smoothed along the subcarrier dimension, and the smoothed matrix is:

[0054]

[0055] Wherein, represents the smoothed matrix of the subcarrier on the mth antenna, and K is the smoothing window length of the subcarrier dimension;

[0056] Then, the BFI matrix is smoothed along the antenna dimension, and the smoothed matrix is:

[0057]

[0058] Wherein, the parameter P is the smoothing window length of the antenna dimension;

[0059] The passive positioning method based on Wi-Fi beamforming feedback information according to the claim, the step two, solving the noise subspace and the signal subspace, comprising the following steps:

[0060] Under single fast snap, the enhanced BFI data matrix is calculated Autocorrelation matrix of the signal:

[0061]

[0062] Then, the autocorrelation matrix is subjected to Eigenvalue Decomposition (EVD):

[0063]

[0064] wherein, Σ s , Σ n respectively represent diagonal matrices composed of eigenvalues of the signal subspace and the noise subspace, U s , U n respectively represent eigenvectors corresponding to the eigenvalues of the signal subspace and the noise subspace; the eigenvalues are arranged in descending order as:

[0065] λ1≥λ2≥…≥λ L ≥λ L+1 =…=λ M =σ 2

[0066] wherein, λ1, λ2, …, λ L are eigenvalues of the signal subspace, λ L+1 , …; λ M are eigenvalues of the noise subspace.

Claims

1. A passive location method based on Wi-Fi beamforming feedback information, comprising the following steps: Step one: assuming that M antennas constitute a uniform linear array to sample the spatial signal transmitted by a single receiving antenna, under far-field conditions, L signal propagation paths are incident to the uniform linear array; assuming that N subcarriers of beamforming feedback information (BFI) can be collected on each antenna, a two-dimensional matrix is constructed using information of different antennas and different subcarriers where 1≤m≤M, 1≤n≤N, represents the BFI value obtained on the Mth antenna and the Nth subcarrier; Step two: processing the two-dimensional matrix The smoothing processing is performed along the subcarrier dimension and the antenna dimension successively to obtain an enhanced matrix wherein K is the smoothing window length of the subcarrier dimension, and P is the smoothing window length of the antenna dimension. Step three: Compute V e autocorrelation matrix of and perform eigen decomposition to solve the noise subspace and signal subspace; Step four: Construct a noise eigenvector matrix U from the eigenvectors in the noise subspace n , compute U n and the spectral function with vector a(θ,τ): wherein (·) H is the conjugate transpose symbol, a(θ,τ) is the steering vector: where a(θ) = [1 e j2πfdsinθ / c … e j2πf(P-1)dsinθ / c ] T , b(τ) = [1 e j2πΔfτ … e j2π(K-1)Δfτ ] T , θ, τ are parameters to be estimated, respectively, the departure angle of the transmitted signal, the relative time of flight; Step 5: Peak searching is performed on the spectrum function P (θ, τ) ; in an ideal condition, the estimated θ should be the real angle of departure (AOD) of each path, and the estimated τ should be the real time of flight (TOF) of each path plus an offset, and the offset is consistent for each path; Step 6: When the estimated target is in a non-straight path of the transceiver, and the position of the transceiver is determined in advance, firstly, parameter estimation is performed by using the above algorithm to obtain the AOD of the straight path of the transceiver and the reflection path of the target, and the time delay difference Δτ of the two paths, wherein Δτ = |τ1-τ2|, τ1 is the relative time of flight of the reflection path, and τ2 is the relative time of flight of the straight path; and then, the parameters are substituted into an elliptical constraint model for calculation, and finally, the specific position of the estimated target is solved.

2. The passive location method based on Wi-Fi beamforming feedback information according to claim 1, wherein the step two, converting the received two-dimensional V matrix into an enhanced matrix comprising the steps of: The matrix V is expressed as: Firstly, smoothing is performed on the BFI along the sub-carrier dimension, and a smoothed matrix is obtained as follows: wherein, represents the smoothed matrix on the mthantenna, and K is the smoothing window length of the subcarrier dimension. Then, smoothing is performed on the BFI matrix along the antenna dimension, and a smoothed matrix is obtained as follows: Wherein, P is the length of the antenna dimension smoothing window. 3.The passive location method based on Wi-Fi beamforming feedback information according to claim 1, wherein the step 2 of solving the noise subspace and the signal subspace comprises the following steps: Under a single shot, calculate the autocorrelation matrix of the enhanced BFI data matrix: ​ Then, eigenvalue decomposition (EVD) is performed on the autocorrelation matrix: In the formula, Σ s , Σ n respectively represent diagonal matrices composed of eigenvalues of the signal subspace and the noise subspace, U s , U n respectively represent eigenvectors corresponding to the eigenvalues of the signal subspace and the noise subspace; the eigenvalues are arranged in descending order: λ1≥ λ2≥... ≥ λ L > λ L+1 == λ M = σ 2 where λ1, λ2, …, λ L are the signal subspace eigenvalues, λ L+1 ,…; λ M are the noise subspace eigenvalues.

Images