Indoor carrier phase positioning model construction method suitable for WiFi
阅读说明：本技术 一种适用于WiFi的室内载波相位定位模型构建方法 (Indoor carrier phase positioning model construction method suitable for WiFi ) 是由 何维 岳紫颖 田增山 李泽 王中春 于 2019-12-03 设计创作，主要内容包括：本发明提出一种适用于WiFi的室内载波相位定位模型构建方法。首先,提取CSI(Channel State Information)信息并进行预处理,获取较为稳定的载波相位观测值；然后,利用二维MUSIC(Multiple Signal Classification)算法,建立基于WiFi的目标信号超分辨参数估计模型,提取TOA(Time of Arrival)构建伪距观测值；接着,基于定位接收机获得直达径的伪距和载波相位观测值,构建载波相位测距方程；最后,基于多个定位接收机,利用载波相位测距方程构造TDOA(Time Difference of Arrival)差分定位模型。该发明方法有效地构建了载波相位定位模型,为载波相位精确定位提供模型依据。(The invention provides a construction method of an indoor carrier phase positioning model suitable for WiFi. Firstly, extracting CSI (channel State information) information and preprocessing the CSI information to obtain a relatively stable carrier phase observation value; then, a target Signal super-resolution parameter estimation model based on WiFi is established by using a two-dimensional MUSIC (multiple Signal Classification) algorithm, and TOA (time of arrival) is extracted to construct a pseudo-range observation value; secondly, a carrier phase ranging equation is constructed based on a pseudo range and a carrier phase observation value of the direct path obtained by the positioning receiver; finally, a TDOA (time Difference of arrival) differential location model is constructed based on the plurality of location receivers by using a carrier phase ranging equation. The method effectively constructs a carrier phase positioning model and provides a model basis for accurate positioning of the carrier phase.)
1. A method for constructing an indoor carrier phase positioning model suitable for WiFi is characterized by comprising the following steps:
a) extracting CSI (channel State information) information and preprocessing the CSI information to obtain a stable carrier phase observation value;
b) establishing a target Signal super-resolution parameter estimation model based on WiFi by using a two-dimensional MUSIC (multiple Signal Classification) algorithm, and extracting TOA (time of arrival) to construct a pseudo-range observation value;
c) obtaining a pseudo range of a direct path and a carrier phase observation value based on a positioning receiver, and constructing a carrier phase ranging equation;
d) a tdoa (time Difference of arrival) differential location model is constructed using carrier-phase ranging equations based on a plurality of location receivers.
2. The method for constructing the phase-location model of the indoor carrier wave suitable for the WiFi according to claim 1, characterized in that the phase of the subcarrier wave of the center frequency point is obtained based on the CSI information, i (i ═ 1,2, …, N) receivers are used for location, each receiver uses a uniform linear array composed of N (N > 3) array elements, the signal uses an Orthogonal Frequency Division Multiplexing (OFDM) modulation method, the number of the subcarrier waves is M, the WiFi outputs the CSI information of each subcarrier wave, which corresponds to the phase value of the carrier wave, and 30 subcarrier waves are selected from the CSI information;
because signal sampling and hardware can bring linear error, the phase needs to be corrected, the OFDM modulation mode can know that the center frequency point subcarrier is not affected by the linear error, because the phase of the center frequency point subcarrier transmits a direct current signal, the phase of the center frequency point subcarrier cannot be directly obtained, in order to obtain the phase of the center frequency point subcarrier, firstly, the CSI phase of the subcarrier is uncoiled, and then, the phase value of the signal center frequency point subcarrier, namely the carrier phase, is obtained by utilizing cubic spline interpolation;
in order to improve the extraction precision of the phase, a more accurate phase value can be obtained by measuring multiple groups of data and performing statistical analysis, firstly, in order to eliminate the influence of the periodicity of the CSI phase, the phase of the subcarrier of the central frequency point is compensated: calculating the mean of the phases of all data packets
wherein phi isnFor carrier phase observations targeted at receiver n,
3. The method for constructing the indoor carrier phase positioning model applicable to WiFi of claim 1, wherein the pseudorange information of the direct path is separated based on super-resolution estimation, and for any receiver, the Channel Frequency Response (CFR) can be expressed as:
wherein, γlAnd τlRespectively representing attenuation and tof (time of flight) of the ith path, wherein L is the number of paths, and f is a carrier frequency, and CSI of all subcarriers can be obtained by performing discrete sampling on CFR, which can be represented as:
H＝[h1,1,…,h1,M,…,hN,1,…,hN,M]T(3)
establishing an angle and time delay estimation model, wherein H can be expressed as:
H＝Z(θ,τ)S(γ)+N (4)
where S (γ) is an L × 1 attenuation vector, N is a NM × 1 noise vector, and Z (θ, τ) is a NM × L directional matrix, which can be expressed as:
wherein the content of the first and second substances,
a(θ)＝[a1(θ),…,aN(θ)]T(6)
b(τ)＝[b1(τ),…,bM(τ)]T(7)
wherein, an(θ)＝e-j2πf(n-1)dsinθ/cIs the nth element in a (theta), theta is the AOA of the path, d is the spacing between the antennas is half wavelength, c is the propagation speed of the electromagnetic wave in vacuum, bm(τ)＝e-j2πΔf(m-1)τFor the mth element in b (τ), Δ f is the subcarrier frequency interval, and the AOA and TOF of the multipath signal can be jointly estimated by using the two-dimensional MUSIC algorithm for equation (4), and further, the pseudo-range observed value of the direct path is separated:
ρn＝τn·c (8)
where ρ isnIs the pseudorange observation, tau, of the target to the receiver nnThe direct path delay from the target to the receiver, c is the propagation speed of the electromagnetic wave in vacuum.
4. The method as claimed in claim 1, wherein the relationship between the measured value and the theoretical value is constructed, i (i ═ 1,2, …, n) receivers are used for positioning, and based on the carrier phase measured value of the direct path, the carrier phase ranging equation is expressed as:
where λ is the wavelength of the carrier wave, NnThe carrier phase integer part of the target and receiver n,
the time of arrival TOA of the direct path is used to perform coarse estimation of the distance from the transmitter to the receiver, and the distance is expressed as:
wherein r isnTarget Euclidean distance to receiver, δ t is target clock error, δ tnFor the clock error of the receiver n to be,
5. The method as claimed in claim 1, wherein the linear model is constructed by using a weighted least square method based on newton iteration, and the target coordinate is P ═ x, y, z]The receiver coordinate is represented as Ri＝[xi,yi,zi]1,2, …, n, comprising in particular the steps of:
the carrier phase localization model is constructed by equations (9) and (10):
the clock error is eliminated by adopting a TDOA difference mode in the formula (11):
due to strict clock synchronization between receivers, equation (12) can be abbreviated as:
wherein the content of the first and second substances,
solving the system of equations (13) using a least squares method based on newton's iteration can be divided into the following steps:
a) prepare data and initial solutions for all connectionsReceiver RnAnd measuring the pseudo range rho from the receiver n to the targetnAnd carrier phase phinBefore Newton iteration is started, an initial estimated value X of the current target position is given0＝[0,0,0]From the initial value of the differential integer ambiguity Δ Ni,1＝0(i＝1,2,…,n)；
b) Linearizing a nonlinear equation system, assuming that k is the number of Newton iterations in the current solution process, then k-1 is the number of iterations completed in the current solution, and solving the last iteration of the formula (13) to obtain Xk-1＝[xk-1，yk-1，zk-1]And (3) linearizing and ignoring observation errors:
wherein the content of the first and second substances,
Δxk,k-1＝xk-xk-1
Δyk,k-1＝yk-yk-1
Δzk,k-1＝zk-zk-1
ΔNn,1＝Nn-N1
equation (14) is abbreviated:
wherein the content of the first and second substances,
ΔX＝[Δxk,k-1Δyk,k-1Δzk,k-1]T
ΔN＝[ΔN2,1ΔN3,1… ΔNn,1]T
c) solving a system of linear equations, solving equation (15) using a weighted least squares algorithm:
wherein, W is a weight matrix which is a diagonal matrix formed by the reciprocal of the sum of the standard deviations of the errors of all the measured values;
d) updating the solution and covariance matrices of the system of equations:
Xk＝Xk-1+ΔX＝Xk-1+[Δxk,k-1Δyk,k-1Δzk,k-1]T(17)
e) judging the convergence of Newton iteration:
if the Newton iteration has converged to the required precision, the Newton iteration method can terminate the loop operation and calculate the updated value (namely X) after the current iteration is calculatedk) As a final result; otherwise, increasing the k value by 1, and returning to b) to repeat Newton iteration calculation again.
Technical Field
The invention belongs to an indoor positioning technology, and discloses a method for constructing an indoor carrier phase positioning model suitable for WiFi.
Background
In recent years, services based on location information are closely related to life of people, such as sharing a bicycle, taking a car by a mobile phone service, taking out services, express information monitoring, intelligent care for old people and children, and the like. The rapid development of location-based applications has led to an increasing demand for precision and reliability in positioning technology. Meanwhile, the high-precision positioning service plays an important role in the fields of national security, economic development, social livelihood and the like. The satellite-based outdoor positioning technology can achieve centimeter-level positioning accuracy, the mobile communication network-based outdoor positioning technology can achieve dozens of meters of positioning accuracy, and the satellite-based outdoor positioning technology plays an important role in the fields of precision agriculture, intelligent transportation, Internet of things and the like. The indoor environment is the most active area of people and is closely related to the production and life of people, so that the indoor position information plays an important role in many applications, such as museums, warehouse asset management, smart medical treatment, smart cities and the like. Since satellite signals hardly penetrate buildings and positioning accuracy based on a mobile communication network cannot meet the requirements of indoor users, an indoor high-accuracy positioning technology has become one of research hotspots and difficulties.
At present, WiFi networks are very popular in life, so that indoor positioning services can be provided by using existing WiFi networks. For positioning, the positioning parameters determine the final positioning accuracy to some extent, and the commonly used positioning parameters mainly include: received Signal Strength (RSS), Angle of Arrival (AOA), and Time of Flight (TOF). RSS is a coarse-grained positioning parameter, so the positioning accuracy based on RSS is low, generally about 2-4 m. The positioning method based on RSS mainly comprises the following steps: firstly, a trilateral localization method based on propagation model ranging; and secondly, a positioning method based on fingerprints. For the first approach, the propagation model needs to be trained in the off-line phase, and the propagation model depends mainly on the attenuation factor. Therefore, the accuracy of the attenuation factor affects the final ranging result. Due to indoor multipath propagation and environmental changes, the trained attenuation factor is often poor in accuracy. In comparison, the fingerprint-based positioning method can provide higher positioning accuracy. However, the construction of the fingerprint database requires a lot of manpower and material resources, and the fingerprint database needs to be updated when the environment changes. Compared with RSS, AOA and TOF are two geometric parameters, and therefore, higher positioning accuracy can be obtained. AOA-based positioning is performed by measuring the angle between the object and multiple receivers and then solving for the object's position using trilateration. The estimation of AOA usually requires the use of array antennas, and the number of antennas in the array is often required to be larger than the number of paths in order to suppress the interference of multipath signals. However, calibration of the radio frequency channels, antenna coupling, multipath interference, and installation calibration all limit the application of AOA positioning methods. TOF-based positioning is achieved by measuring the propagation delay of the target to multiple receivers, thus requiring strict clock synchronization between the transmitter and receiver, otherwise a lot of phase errors are introduced. In addition to clock synchronization, the signal bandwidth also determines the resolution of multipath on TOF. For example, for WiFi signals of 20M and 40M bandwidth, the resolution of TOF is only 15M and 7.5M. Time Difference of Arrival (TDOA) is another Time observation measurement that requires strict clock synchronization between receivers. Since WiFi can provide rich Channel State Information (CSI), the resolution of the contained carrier phase Information is high, so the carrier phase location model has practical significance. However, the single carrier phase information is not accurate, so the toa (time of arrival) information needs to be combined to construct a joint positioning model.
The invention provides an indoor carrier phase positioning model construction method suitable for WiFi, which is used for positioning a target by utilizing carrier phase information and TOA information of a direct signal. Compared with the existing distance-based positioning method, the method has the advantages that the resolution of observation data can be improved by utilizing the carrier phase, and the high-precision positioning of the target is realized. Firstly, extracting CSI information and preprocessing the CSI information to obtain a relatively stable carrier phase observation value; then, establishing a target signal super-resolution parameter estimation model based on WiFi by using a two-dimensional MUSIC (multiple Signal Classification) algorithm, and extracting TOA to construct a pseudo-range observation value; secondly, a carrier phase ranging equation is constructed based on a pseudo range and a carrier phase observation value of the direct path obtained by the positioning receiver; and finally, constructing a TDOA differential positioning model by utilizing a carrier phase ranging equation based on the plurality of positioning receivers. The method effectively constructs a carrier phase positioning model and provides a model basis for accurate positioning of the carrier phase.
Disclosure of Invention
The invention aims to provide an indoor carrier phase positioning model construction method suitable for WiFi, which can effectively utilize carrier phases to position targets.
The invention discloses a method for constructing an indoor carrier phase positioning model suitable for WiFi, which comprises the following steps of:
step one, extracting CSI information and preprocessing the CSI information to obtain a relatively stable carrier phase observation value;
and step two, establishing a target signal super-resolution parameter estimation model based on WiFi by using a two-dimensional MUSIC algorithm, and extracting TOA to construct a pseudo-range observation value.
Thirdly, obtaining a pseudo range and a carrier phase observation value of the direct path based on the positioning receiver, and constructing a carrier phase ranging equation;
and fourthly, constructing a TDOA differential positioning model by utilizing a carrier phase ranging equation based on a plurality of positioning receivers.
Advantageous effects
The invention discloses a method for constructing an indoor carrier phase positioning model suitable for WiFi, which has the following advantages:
1. the distance measurement resolution based on the carrier phase depends on the observed quantity, the distance measurement resolution is the carrier wavelength, the resolution is better than 30cm for signals in an L wave band and above, and the resolution is obviously improved compared with the traditional positioning parameters.
2. The TDOA difference principle is utilized to eliminate the influence of clock difference on positioning, and the method can be applied to the existing commercial wireless local area network and mobile network;
3. effectively constructs a carrier phase positioning model and provides a model basis for accurate positioning of carrier phases.
Drawings
FIG. 1 is a flow chart of the present invention;
fig. 2 is a schematic view of an indoor scene.
Detailed description of the preferred embodiments
The invention is described in further detail below with reference to the accompanying drawings:
FIG. 1 is a flow chart of the present invention, comprising the following steps:
step one, presume under the indoor environmentThe room size is L multiplied by W multiplied by Am3As shown in fig. 2, wherein L is the room length, W is the room width, and a is the room height. Positioning by using i (i ═ 1,2, …, N) receivers, wherein each receiver adopts a uniform linear array consisting of N (N > 3) array elements, signals adopt an Orthogonal Frequency Division Multiplexing (OFDM) modulation mode, and the number of subcarriers is M. And the WiFi outputs CSI information of each subcarrier, the CSI information corresponds to the carrier phase value, and 30 subcarriers are selected from the carrier phase value.
Since signal sampling and hardware introduce linearity errors, phase correction is required. According to the OFDM modulation mode, the center frequency point subcarrier is not influenced by the linear error. Since the phase of the center frequency point subcarrier transmits a direct current signal, the phase thereof cannot be directly obtained. In order to obtain the phase of the subcarrier of the central frequency point, firstly, the CSI phase of the subcarrier is unwrapped, and then, a phase value of the subcarrier of the signal central frequency point, namely, the carrier phase is obtained by utilizing cubic spline interpolation.
In order to improve the extraction precision of the phase, more accurate phase values can be obtained by measuring multiple groups of data and carrying out statistical analysis. Firstly, in order to eliminate the influence of the phase periodicity of the CSI, the phase of the center frequency subcarrier is compensated: calculating the mean of the phases of all data packets
wherein phi isnFor carrier phase observations targeted at receiver n,and λ is the wavelength of the carrier wave based on the statistical mean of the carrier phases of the CSI.
Step two, for any receiver, the Channel Frequency Response (CFR) can be expressed as:
wherein, γlAnd τlRespectively representing attenuation and TOF of the ith path, wherein L is the number of paths and f is the carrier frequency. The CFR is discretely sampled to obtain CSI matrices of all subcarriers, which can be expressed as:
H＝[h1，1，…，h1，M，…，hN，1，…，hN，M]T(3)
establishing an angle and time delay estimation model, wherein H can be expressed as:
H＝Z(θ，τ)S(γ)+N (4)
where S (γ) is an L × 1 attenuation vector and N is a NM × 1 noise vector. Z (θ, τ) is a NM × L directional matrix, which can be expressed as:
wherein the content of the first and second substances,
a(θ)＝[a1(θ)，…，aN(θ)]T(6)
b(τ)＝[b1(τ)，…，bM(τ)]T(7)
wherein, an(θ)＝e-j2πf(n-1)dsinθ/cIs the nth element in a (theta), theta is the AOA of the path, d is the distance between the antennas which is half wavelength, and c is the propagation speed of the electromagnetic wave in vacuum. bm(τ)＝e-j2πΔf(m-1)τIs the mth element in b (τ) and Δ f is the subcarrier frequency spacing. By using the two-dimensional MUSIC algorithm for equation (4), AOA and TOF of the multipath signal can be jointly estimated. Further, separating pseudo-range observations of the direct path:
ρn＝τn·c (8)
where ρ isnIs the pseudorange observation, tau, of the target to the receiver nnThe direct path delay from the target to the receiver, c is the propagation speed of the electromagnetic wave in vacuum.
Step three, based on the carrier phase measured value of the direct path, the carrier phase ranging equation is expressed as:
where λ is the wavelength of the carrier wave, NnThe carrier phase integer part of the target and receiver n,
The time of arrival TOA of the direct path is used to perform coarse estimation of the distance from the transmitter to the receiver, and the distance is expressed as:
wherein r isnTarget Euclidean distance to receiver, δ t is target clock error, δ tnFor the clock error of the receiver n to be,
Step four, as shown in figure 2, for the sake of intuitionRepresenting the establishment of a three-dimensional coordinate system within a room. P, R for target and receiver positions respectivelynTo mark, the coordinates of the target are P ═ x, y, z]The receiver coordinate is represented as Ri＝[xi,yi,zi]I is 1,2, …, n. The carrier phase localization model is constructed by equations (9) and (10):
the clock error is eliminated by adopting a TDOA difference mode in the formula (11):
due to strict clock synchronization between receivers, equation (12) can be abbreviated as:
wherein the content of the first and second substances,
solving the system of equations (13) using a least squares method based on newton's iteration can be divided into the following steps:
4a, preparing data and an initial solution. For all receivers RnAnd measuring the pseudo range rho from the receiver n to the targetnAnd carrier phase phin. Before Newton iteration is started, an initial estimated value X of the current target position is given0＝[0,0,0]From the initial value of the differential integer ambiguity Δ Ni,1＝0(i＝1,2,…,n)。
And 4b, linearizing the nonlinear equation system. Assuming that k is the number of newton iterations in progress in the current solution, k-1 is the number of iterations that have been completed in the current solution. Solving the formula (13) in the last iteration to obtain the solution Xk-1＝[xk-1，yk-1，zk-1]And (3) linearizing and ignoring observation errors:
wherein the content of the first and second substances,
Δxk,k-1＝xk-xk-1
Δyk,k-1＝yk-yk-1
Δzk,k-1＝zk-zk-1
ΔNn,1＝Nn-N1
equation (14) is abbreviated:
wherein the content of the first and second substances,
ΔX＝[Δxk,k-1Δyk,k-1Δzk,k-1]T
ΔN＝[ΔN2,1ΔN3,1… ΔNn,1]T
and 4c, solving a linear equation system. Solving equation (15) using a weighted least squares algorithm:
wherein, W is a weight matrix which is a diagonal matrix formed by inverses of the sum of standard deviations of errors of various measured values.
And 4d, updating the solution of the equation set and the covariance matrix.
Xk＝Xk-1+ΔX＝Xk-1+[Δxk,k-1Δyk,k-1Δzk,k-1]T(17)
And 4e, judging the convergence of Newton iteration.
If the Newton iteration has converged to the required precision, the Newton iteration method can terminate the loop operation and calculate the updated value (namely X) after the current iteration is calculatedk) As a final result; otherwise, the k value is increased by 1, and the Newton iteration calculation is repeated again after returning to 3 b.