CN105611627A - Method for estimating AOA of WLAN access point based on double antennas - Google Patents

Abstract

The invention relates to a method for estimating the AOA of a WLAN access point based on double antennas. By adopting a common double-antenna WLAN access point, the time of flight TOF from a transmitting end to a receiving end and an angle of arrival are subjected to joint estimation by utilizing received CSI information at the receiving end. According to the invention, influence of an indoor complex multi-path environment on angle estimation is sufficiently considered; the average error of the estimated angle is within 5 DEG; and direct transmitting radical angle estimation requirements under indoor multi-path signals can be satisfied.

Description

The method of estimation of WLAN access point AOA based on double antenna

Technical field

The present invention relates to signal processing technology, wireless location technology field, is specifically related to based on double antennaThe method of estimation of WLAN access point AOA, is applicable to the navigation system of indoor environment.

Background technology

At moving communicating field, people are to location-based service (LocationBasedServices, LBS)Demand constantly increase, and along with WLAN (WirelessLocalAreaNetworks,WLAN) universal, it is emerging that the indoor locating system based on WLAN has catered to of this demand justStudy hotspot.

AOA (angleofarrival, ripple the reaches angle) information of obtaining incoming wave under indoor environment can be for indoorNavigation system provides crucial positional parameter, thereby realizes the hi-Fix under indoor environment. Meanwhile, and orderThe front location technology based on RSSI is compared, and the location technology based on AOA does not need to carry out location fingerprint storehouse structureBuild, saved a large amount of human costs. Meanwhile, the location technology based on AOA is a kind of Passive Positioning technologyThe network side location that can realize easily base station, does not need any unnecessary software of user installation. Therefore, howThe angle of arrival of estimating accurately arrival bearing is the key that solves indoor positioning problem.

Traditional AOA algorithm for estimating is as MUSIC algorithm, and ESPRIT algorithm mainly adopts special many antennasArray apparatus, the signal subspace of utilization and the orthogonality of noise subspace realize angle estimation. But traditionalAlso there is obvious defect in angle estimation algorithm, must adopt special large scale array antenna just can accurately estimateCount out AOA information, this just estimates to be provided with obstacle in indoor application for AOA.

At present, mostly adopt intelligent antenna technology in GSM, therefore serving BS can provide more accurateTrue angle of arrival information, and the location-based service that provides side Network Based to locate. For example,, at present based on LTEThe AOA navigation system of network, this system utilizes the precoding mechanism of MIMO to realize obtaining of AOA.In indoor application, existing Ubicarse system at present, under WLAN environment, adopts SAR (SyntheticApertureRadar) mentality of designing, realizes angle estimation by the large aerial array of rotation receiving terminal antenna modeling.In addition, current other a system ArrayPhaser system, by cascade WLAN access point, realizes many daysThe design of the system of linear array structure. In addition, DirectionFinding system is utilized WLAN double antenna,Adopt the principle of interferometer direction finding to realize arrival angular measurement. Although said system can arrive and estimate accuratelyMeter, but all there is defect in various degree in said system. The realization of Ubicarse system need to be to receiving equipmentAntenna improves; ArrayPhaser system need to be done larger transformation to WLAN access point apparatus, simultaneouslyIn cascaded design, must fully take into account distinct device exist synchronous error, this also the maintenance of later stage system carryGo out larger challenge. DirectionFinding system is carried out angle estimation by two antennas, can not be effectiveEstimate the multipath signal existing in environment, the angular resolution simultaneously estimating is poor. These of above-mentioned existenceProblem brings limitation all can to the popularization of AOA navigation system in indoor positioning field.

Summary of the invention

The object of this invention is to provide the method for estimation of a kind of WLAN access point AOA based on double antenna,It estimates that the mean error of angle, in 5 degree left and right, can meet the need of the angle estimation of indoor multipath signal far awayAsk.

The method of estimation of WLAN access point AOA based on double antenna of the present invention, comprises following stepRapid:

Step 1, configuration WLAN;

Step 2, adopt multiple subcarrier pair initial data to carry out orthogonal modulation at transmitting terminal;

Step 3, estimate the channel condition information of each subcarrier: the CSI that receives OFDM at receiving terminalChannel information Matrix C SI_matrix：

${\mathrm{CSI}}_{matrix}=\left[\begin{array}{ccccc}{\mathrm{csi}}_{1,1}& {\mathrm{csi}}_{1,2}& ...& {\mathrm{csi}}_{1,19}& {\mathrm{csi}}_{1,20}\\ {\mathrm{csi}}_{2,1}& {\mathrm{csi}}_{2,2}& ...& {\mathrm{csi}}_{2,19}& {\mathrm{csi}}_{2,20}\end{array}\right];$

Wherein, csi_i,jIt is the channel information value of j subcarrier on i root antenna;

Step 4, employing space smoothing algorithm carry out decorrelation processing to the CSI channel matrix receiving;

Step 5, utilize CSI information after decorrelation to signal arrive array antenna flight time TOF andArrive angle algorithm for estimating and combine estimation, obtain the estimated value of direction of arrival.

In described step 3, estimate the channel condition information of each subcarrier, adopt least square (LS) skillArt is carried out channel estimating;

Least square technology is that N-1 subcarrier of hypothesis is orthogonal, there is no ISI (InterSymbolInterference is intersymbol interference), the training signal receiving is Y[k], k=0,1,2 ..., N-1 is to channel HBe estimated asBy cost functionMinimize:

$J\left(\hat{H}\right)=\left|\right|Y-X\hat{H}|{|}^2=Y^{H}Y-Y^{H}X\hat{H}-{\hat{H}}^H{X}^{H}Y+{\hat{H}}^H{X}^{H}A\hat{H};$

Make cost functionAboutPartial derivative equal 0, the solution that obtains LS channel estimating is:

${\hat{H}}_{LS}=\left(X^H\right)X^{-1}{X}^{H}Y=X^{-1}Y;$

OrderRepresentIn element, k=0,1,2 ..., N-1, by can under the assumed condition without ICIKnow that X is diagonal matrix, therefore the LS Signal estimation on each subcarrier can be expressed as:

${\hat{H}}_{LS}\[k\]=\frac{Y\[k\]}{X\[k\]},k=0,1,2,...N-1.$

In described step 5, the CSI information after decorrelation arrives the flight time TOF of array antenna to signalThe two-dimensional directional square that while combining estimation with arrival angle algorithm for estimating, structure comprises AOA and TOF informationBattle array:

$A=\[\stackrel{\→}{a}({\mathrm{\θ}}_1,{\mathrm{\τ}}_1),\stackrel{\→}{a}({\mathrm{\θ}}_2,{\mathrm{\τ}}_2),...,\stackrel{\→}{a}({\mathrm{\θ}}_N,{\mathrm{\τ}}_N)\];$

WhereinThe direction vector of M*1, θ_kThe angle of arrival of k paths, τ_kK articleThe flight time in path, finally set up and utilized OFDM multicarrier information to combine TOF and AOAThe system of estimating;

AOA method of estimation is to utilize MUSIC algorithm to complete the estimation that arrives angle; MUSIC algorithm is rootObtain the estimated value of covariance matrix according to N reception signal phasor;

$R=\frac{1}{N}\underset{n=1}{\overset{N}{\Σ}}x\left(n\right)x^H\left(n\right);$

Wherein, R is covariance matrix, then it is carried out to Eigenvalues Decomposition R=U Σ U^H, according to the size of characteristic valueSequentially, the eigenvalue of maximum characteristic of correspondence vector equating with signal number K is regarded as to signal subspace, surplusUnder (M-K) individual characteristic value character pair vector regard noise subspace as, $R=U_S{\mathrm{\Σ}}_S{U}_S^H+U_N{\mathrm{\Σ}}_N{U}_N^H,$ Make again angle traversal 360 degree, according to $P_{MUSIC}=1/\left(a\right(\mathrm{\θ}\left)U_N{U}_N^{H}a\right(\mathrm{\θ}\left)\right)$ Calculate spectral function P_MUSIC,, wherein a (θ) is for obtaining the estimated values theta of direction of arrival by seeking peak value.

The present invention has the following advantages: carry out the estimation of AOA based on existing WiFi equipment, utilization of the present inventionChannel condition information (ChannelStateInformation, the CSI) square of the OFDM orthogonal sub-carriers receivingBattle array arrives flight time (TimeofFlight, TOF) of array antenna and AOA to multipath signal and combines and estimateMeter, than traditional WLAN angle measurement technique, the present invention is without any transceiver of change, Neng GoudaTo identical angle measurement accuracy, estimate that the mean error of angle, in 5 degree left and right, can meet indoor multipath letter far awayNumber the demand of angle estimation, there is higher promotional value.

Brief description of the drawings

Fig. 1 theory diagram of the present invention.

Detailed description of the invention

Below in conjunction with the drawings and specific embodiments, the present invention is described in detail.

As shown in Figure 1, entire system theory diagram.

(1) on WLAN transmitting network interface card, produce initial data, after OFDM baseband modulation, pass through radio frequencySend.

(2) receive on network interface card and receive the radiofrequency signal through multipath channel at WLAN, then carry out down coversionObtain baseband signal.

(3) obtain after baseband signal, first baseband signal is carried out to demodulation, obtain the CSI information that receives signal.

(4) utilize the CSI information architecture observational equation getting, utilize TOF to combine algorithm for estimating with AOAEstimate AOA information.

The method of estimation of WLAN access point AOA based on double antenna of the present invention, comprises following stepRapid:

Step 1, configuration WLAN:

Center frequency point, bandwidth, array antenna number (2), antenna distance, orthogonal modulation subcarrier number,Sub-carrier frequencies spacing, signal to noise ratio etc.

Step 2, adopt multiple subcarrier pair initial data to carry out orthogonal modulation.

Step 3, estimate the channel condition information of each subcarrier:

Suppose that all subcarriers are orthogonal, there is no ICI, the training symbol of N subcarrier can be representedBecome matrix form:

Wherein, X[k] represent the pilot signal on k subcarrier, meet E{X[k]=0,Var{X[k]}＝σ², k=0,1,2 ... N-1. Because suppose that all subcarriers are all orthogonal, so X isA diagonal matrix. Channel gain [H[0] H[1 of a given k carrier wave] ... H[k-1]]^T, receiveK training signal [Y[0] Y[1] ... Y[k-1]]^TCan be expressed as:

Wherein, H is channel vector, H=[H[0], H[1] ..., H[k-1]]^T; Z is noise vectorZ＝[Z[0],Z[1],…,Z[k-1]]^T, Y is the training signal vector receiving,Y＝[Y[0]Y[1]…Y[k-1]]^T, meet E{Z[k] }=0,k＝0,1,…N-1。

In order to obtain channel estimatingLS channel estimating method need to minimize cost function below

$J\left(\hat{H}\right)=\left|\right|Y-X\hat{H}|{|}^2=Y^{H}Y-Y^{H}X\hat{H}-{\hat{H}}^H{X}^{H}Y+{\hat{H}}^H{X}^{H}A\hat{H};$

Make cost functionAboutPartial derivative equal 0, that is:

$\frac{\∂J\left(\hat{H}\right)}{\∂\hat{H}}=-2{\left(X^{H}Y\right)}^{*}+2\left(X^{H}X\hat{H}\right)=0;$

Then can obtainObtain thus the solution of LS channel estimatingFor:

${\hat{H}}_{LS}=\left(X^H\right)X^{-1}{X}^{H}Y=X^{-1}Y;$

OrderRepresentIn element, k=0,1,2 ..., N-1. By can under the assumed condition without ICIKnow that X is diagonal matrix, therefore the LS Signal estimation on k subcarrier can be expressed as:

${\hat{H}}_{LS}=\frac{Y\[k\]}{X\[k\]},k=0,1,2,...N-1;$

Receive CSI (ChannelStateInformation) the channel information matrix of OFDM at receiving terminalCSI_matrix：

${\mathrm{CSI}}_{matrix}=\left[\begin{array}{ccccc}{\mathrm{csi}}_{1,1}& {\mathrm{csi}}_{1,2}& \mathrm{...}& {\mathrm{csi}}_{1,19}& {\mathrm{csi}}_{1,20}\\ {\mathrm{csi}}_{2,1}& {\mathrm{csi}}_{2,2}& \mathrm{...}& {\mathrm{csi}}_{2,19}& {\mathrm{csi}}_{2,20}\end{array}\right];$

Wherein, csi_i,jIt is the channel information value of j subcarrier on i root antenna.

Step 4, employing space smoothing algorithm carry out decorrelation processing to the CSI channel matrix receiving:

Uniform line-array is divided into several equitant subarrays. If identical (this of the array manifold of each subarraySuppose to be applicable to uniform line-array), subarray covariance matrix can be added in the original meaning of rear average replacementArray covariance matrix R_s. The uniform line-array of M unit is divided into L submatrix with sliding, and each submatrix has NUnit, wherein N=M-L+1. L forward direction submatrix is output as:

$x_l^f\left(t\right)={\[x_l\left(t\right),x_{l+1}\left(t\right),...,x_{l+N-1}\left(t\right)\]}^T=A_M{D}^{1-1}s\left(t\right)+n_l\left(t\right),1\≤l\≤L;$

Wherein s (t) is the complex amplitude of receiving plane ripple, n_l(t) be noise vector, A_MFor the direction of N × M dimensionMatrix, it classifies the steering vector a of N dimension as_M(θ_i)(i＝1,2,…,K)，

$D=diag(e^{j\frac{2\mathrm{\π}d}{\mathrm{\λ}}{\mathrm{sin\θ}}_1},e^{j\frac{2\mathrm{\π}d}{\mathrm{\λ}}{\mathrm{sin\θ}}_2},\mathrm{...},e^{j\frac{2\mathrm{\π}d}{\mathrm{\λ}}{\mathrm{sin\θ}}_K});$

D is the interval between array antenna, and λ is the wavelength that receives signal, θ_iIt is the incoming wave side of i information sourceTo, so the covariance matrix of l forward direction submatrix is:

$R_l^f=E\[x_l^f\left(t\right)x_l^f{\left(t\right)}^H\]=A_M{D}^{l-1}{R}_S{\left(D^{l-1}\right)}^H{A}_M^H;$

Forward direction space smoothing covariance matrix R_fFor:

$R_f=\frac{1}{L}\underset{l=1}{\overset{L}{\Σ}}{R}_l^f;$

On this basis, the inverted order battle array of investigation line array (press M, M-1 ..., 2,1 orders are arranged). In like manner canTo obtain the covariance matrix R of backward space smoothing_bFor:

$R_b=\frac{1}{L}\underset{l=1}{\overset{L}{\Σ}}{R}_l^b;$

Wherein, the covariance matrix of l backward submatrixR in fact_bBe exactly R_fReverse conjugate array, heBetween this relation be exactly the conjugation inverted order consistency of often saying. Therefore, front and back to smoothed covariance matrix are:

$\tilde{R}=\frac{1}{2}(R_f+R_b);$

The advantage of conjugation inverted order consistency is to increase submatrix number, but relatively former array, arrayStill has reduced effective aperture, because subarray is less than former array. Although there is this aperture loss, it changesBecome the limitation of the feature decomposition class AOA algorithm based on aerial array covariance matrix.

Step 5, utilize CSI information after decorrelation arrive the flight time of array antenna and arrive angle signalAlgorithm for estimating is combined estimation:

CSI information after decorrelation to signal arrive array antenna flight time TOF (TimeofFlight) andArrive angle algorithm for estimating and combine the two-dimensional directional square that while estimating location, structure comprises AOA and TOF informationBattle array:

$A=\[\stackrel{\→}{a}({\mathrm{\θ}}_1,{\mathrm{\τ}}_1),\stackrel{\→}{a}({\mathrm{\θ}}_2,{\mathrm{\τ}}_2),...,\stackrel{\→}{a}({\mathrm{\θ}}_N,{\mathrm{\τ}}_N)\];$

WhereinThe direction vector of M*1, θ_kThe angle of arrival of k paths, τ_kK article of roadThe flight time in footpath, finally set up and utilized OFDM multicarrier information that TOF and AOA are combined and estimatedThe system of meter.

AOA method of estimation is to utilize MUSIC algorithm to complete the estimation that arrives angle. Suppose between adjacent two antennasApart from being d, when signal is after different propagation paths, establish the certain angle θ of signal of certain multipath_kIncideWhen receiving antenna array, due to incident angle θ_kExistence, on adjacent two antennas can introduce and θ_kBe correlated withPath length difference dsin θ_k, the phase deviation being caused by this path length difference is 2 π fdsin θ_k/ c. If with first skyLine is reference, and establishing by the wavefront signals of v paths arrival antenna 1 is s_k(t), i antenna receivedThe wavefront signals in k article of footpath be:

s_i,k(t)＝a_ks_k(t)exp(-jw₀(i-1)dsinθ_k/c)；

Wherein s_i,k(t) represent the wavefront signals of the k paths that i antenna receive, a_kRepresent k article of roadThe amplitude fading in footpath, w₀Represent the angular speed of carrier wave, d representative antennas interval, c represents the light velocity. ConsiderThe impact of the noise of i strip antenna, total N bar multipath signal in assumptions' environment, i antenna received soWavefront signals is the summation of many multipath signals, that is:

$s_i\left(t\right)=\underset{k=1}{\overset{N}{\Σ}}{s}_{i,k}\left(t\right)=\underset{k=1}{\overset{N}{\Σ}}{a}_k{s}_k\left(t\right)\exp (-{\mathrm{jw}}_0(i-1)d{\mathrm{sin\θ}}_k/c)+n_i\left(t\right);$

Suppose that array antenna number is M, the noise of each antenna is that average is 0, and variance is σ²Stationary white noiseProcess, and uncorrelated mutually between noise, simultaneously uncorrelated between noise and signal, to receiving on each antennaWavefront signals write as vector form and be:

X(t)＝AS(t)+N(t)；

Wherein, X (t)=[x₁(t),x2(t),...,x_M(t)]^TFor the reception data vector of M dimension,S(t)＝[S₁(t),S₂(t),...,S_N(t)]^TFor N dimensional signal vector,ForM*N ties up direction matrix.For M dimension direction vector,τ_k＝dsinθ_k/c，N(t)＝[n₁(t),n₂(t),...,n_M(t)]^TFor M dimension noise, wherein N is mulitpathNumber.

MUSIC algorithm is the singular value decomposition of utilizing the autocorrelation matrix that receives signal X (t), by special category spaceBe divided into mutually orthogonal signal subspace and noise subspace, utilize subspace orthogonal property obtain spectrum estimateResult. Suppose, the autocorrelation matrix of X (t) is R_xx(t)，

R_xx(t)＝E[X(t)X^H(t)]＝R_s+R_noise；

Wherein R_xx(t) the M*M symmetrical matrix that is M for order, R_s＝E[S(t)S^H(t)] for order is NN*N matrix, R_noiseFor the order matrix that is M-N. By to R_xx(t) eig, andBy size characteristic value is sorted, get M-N minimum characteristic value characteristic of correspondence vector and form noiseSpace, is designated as E_n＝{v_N+1,v_N+2,...,v_M, due to noise subspace and signal subspace orthogonal, soSense θ_kUpper, obviously have:

$E_N^H\stackrel{\→}{a}\left({\mathrm{\θ}}_k\right)=0;$

For the arrival bearing who realizes N multipath signal estimates, by continuously changing θ value, utilize above formula to carry outSpectrum peak search, obtains N spectrum peak corresponding to θ and is the arrival bearing of multipath signal:

$P_{music}\left(\mathrm{\θ}\right)=\frac{1}{{\stackrel{\→}{a}}^H\left(\mathrm{\θ}\right)E_N{E}_N^H\stackrel{\→}{a}\left(\mathrm{\θ}\right)}.$

Further, described step 3 is estimated the channel condition information of each subcarrier, adopts least square (LS)Technology is carried out channel estimating.

Least square technology is that N-1 subcarrier of hypothesis is orthogonal, there is no ICI, the training letter receivingNumber be Y[k], k=0,1,2 ..., N-1 is estimated as channel H'sMinimize cost function below:

$J\left(\hat{H}\right)=\left|\right|Y-A\hat{H}|{|}^2=Y^{H}Y-Y^{H}X\hat{H}-{\hat{H}}^H{X}^{H}Y+{\hat{H}}^H{X}^{H}X\hat{H};$

Cost function above of order aboutPartial derivative equal 0, the solution that obtains LS channel estimating is:

${\hat{H}}_{LS}=\left(X^H\right)X^{-1}{X}^{H}Y=X^{-1}Y;$

OrderRepresentIn element, k=0,1,2 ..., N-1. By can under the assumed condition without ICIKnow that X is diagonal matrix, therefore the LS Signal estimation on each subcarrier can be expressed as:

${\hat{H}}_{LS}=\frac{Y\[k\]}{X\[k\]},k=0,1,2,...N-1.$

Further, the CSI information after described step 5 decorrelation arrives the flight time of array antenna to signalTOF and arrival angle algorithm for estimating are combined structure while estimating location and are comprised two of AOA and TOF informationDimension direction matrix:

$A=\[\stackrel{\→}{a}({\mathrm{\θ}}_1,{\mathrm{\τ}}_1),\stackrel{\→}{a}({\mathrm{\θ}}_2,{\mathrm{\τ}}_2),...,\stackrel{\→}{a}({\mathrm{\θ}}_N,{\mathrm{\τ}}_N)\];$

WhereinThe direction vector of M*1, θ_kThe angle of arrival of k paths, τ_kK article of roadThe flight time in footpath, finally set up and utilized OFDM multicarrier information that TOF and AOA are combined and estimatedThe system of meter.

MUSIC algorithm is the estimated value that obtains covariance matrix according to N reception antenna reception signal,

$R=\frac{1}{N}\underset{n=1}{\overset{N}{\Σ}}x\left(n\right)x^H\left(n\right)$

Wherein, R is covariance matrix, then it is carried out to Eigenvalues Decomposition R=U Σ U^H, according to characteristic valueSize order, regards signal subspace as the eigenvalue of maximum characteristic of correspondence vector equating with signal number K,Remaining (M-K) individual characteristic value character pair vector is regarded as to noise subspace, $R=U_S{\mathrm{\Σ}}_S{U}_S^H+U_N{\mathrm{\Σ}}_N{U}_N^H,$ Make again angle traversal 360 degree, according to $P_{MUSIC}=1/\left(a\right(\mathrm{\θ}\left)U_N{U}_N^{H}a\right(\mathrm{\θ}\left)\right)$ Calculate spectral function P_MUSIC,, wherein a (θ) is for obtaining the estimated values theta of direction of arrival by seeking peak value.

The present invention has taken into full account the impact that the multi-path environment of indoor complexity causes angle estimation, adopts this systemAnd method, estimate that the mean error of angle is in 5 degree left and right, the angle that can meet indoor multipath signal is far away estimatedThe demand of meter.

Claims (3)

1. a method of estimation of the WLAN access point AOA based on double antenna, is characterized in that, comprisesFollowing steps:

Step 1, configuration WLAN;

Step 2, adopt multiple subcarrier pair initial data to carry out orthogonal modulation at transmitting terminal;

Step 3, estimate the channel condition information of each subcarrier:

Receive the CSI channel information Matrix C SI of OFDM at receiving terminal_matrix：

${\mathrm{CSI}}_{matrix}=\left[\begin{array}{ccccc}{\mathrm{csi}}_{1,1}& {\mathrm{csi}}_{1,2}& ...& {\mathrm{csi}}_{1,19}& {\mathrm{csi}}_{1,20}\\ {\mathrm{csi}}_{2,1}& {\mathrm{csi}}_{2,2}& ...& {\mathrm{csi}}_{2,19}& {\mathrm{csi}}_{2,20}\end{array}\right];$

Wherein, csi_i,jIt is the channel information value of j subcarrier on i root antenna;

Step 4, employing space smoothing algorithm carry out decorrelation processing to the CSI channel matrix receiving;

Step 5, utilize CSI information after decorrelation to signal arrive array antenna flight time TOF andArrive angle algorithm for estimating and combine estimation, obtain the estimated value of direction of arrival.

2. the method for estimation of the WLAN access point AOA based on double antenna according to claim 1,It is characterized in that: in described step 3, estimate the channel condition information of each subcarrier, adopt least square (LS)Technology is carried out channel estimating;

Least square technology is that N-1 subcarrier of hypothesis is orthogonal, there is no ICI, the training letter receivingNumber be Y[k], k=0,1,2 ..., N-1 is estimated as channel H'sBy cost functionMinimize:

$J\left(\hat{H}\right)=\left|\right|Y-X\hat{H}|{|}^2=Y^{H}Y-Y^{H}X\hat{H}-{\hat{H}}^H{X}^{H}Y+{\hat{H}}^H{X}^{H}X\hat{H};$

Make cost functionAboutPartial derivative equal 0, the solution that obtains LS channel estimating is:

${\hat{H}}_{LS}=\left(X^H\right)X^{-1}{X}^{H}Y=X^{-1}Y;$

OrderRepresentIn element, k=0,1,2 ..., N-1, by can under the assumed condition without ISIKnow that X is diagonal matrix, therefore the LS Signal estimation on each subcarrier can be expressed as:

${\hat{H}}_{LS}\[k\]=\frac{Y\[k\]}{X\[k\]},k=0,1,2,...N-1.$

3. the estimation side of the WLAN access point AOA based on double antenna according to claim 1 and 2Method, is characterized in that: in described step 5, the CSI information after decorrelation arrives flying of array antenna to signalLine time TOF and arrival angle algorithm for estimating are combined while estimation, and structure comprises AOA and TOF informationTwo-dimensional directional matrix:

$A=\[\stackrel{\→}{a}({\mathrm{\θ}}_1,{\mathrm{\τ}}_1),\stackrel{\→}{a}({\mathrm{\θ}}_2,{\mathrm{\τ}}_2),...,\stackrel{\→}{a}({\mathrm{\θ}}_N,{\mathrm{\τ}}_N)\];$

WhereinThe direction vector of M*1, θ_kThe angle of arrival of k paths, τ_kK articleThe flight time in path, finally set up and utilized OFDM multicarrier information to combine TOF and AOAThe system of estimating;

AOA method of estimation is to utilize MUSIC algorithm to complete the estimation that arrives angle; MUSIC algorithm is rootObtain the estimated value of covariance matrix according to N reception signal phasor x (n);

$R=\frac{1}{N}\underset{n=1}{\overset{N}{\Σ}}x\left(n\right)x^H\left(n\right);$

Wherein, R is covariance matrix, then it is carried out to Eigenvalues Decomposition R=U Σ U^H, large according to characteristic value ULittle order, regards signal subspace as the eigenvalue of maximum characteristic of correspondence vector equating with signal number K,(M-K) the individual characteristic value character pair vector being left is regarded noise subspace as, $R=U_S{\mathrm{\Σ}}_S{U}_S^H+U_N{\mathrm{\Σ}}_N{U}_N^H,$ Make again angle traversal 360 degree, according to $P_{MUSIC}=1/\left(a\right(\mathrm{\θ}\left)U_N{U}_N^{H}a\right(\mathrm{\θ}\left)\right)$ Calculate spectral function P_MUSIC, wherein a (θ) is for obtaining the estimated values theta of direction of arrival by seeking peak value.