CN110912630A - Airspace spectrum sensing method based on multiple antennas - Google Patents

Abstract

The invention belongs to the technical field of cognitive radio, and in particular relates to a multi-antenna-based spatial spectrum sensing method. The present invention uses the spatial angle dimension information as a new spectrum opportunity space domain spectrum sensing method, estimates the arrival angle of the signal space angle dimension, and can avoid the communication direction of the main user or the communication direction of the main user through the beam forming technology Carry out null-notch antenna beam design, so that cognitive users can access the spectrum through different spatial angles at the same frequency, at the same time or even at the same place, avoiding the communication direction of the main user, thereby increasing system capacity and improving spectrum utilization Rate. Compared with the spectrum sensing algorithm of the traditional dimension, the present invention takes the spatial angle dimension information as a new spectrum opportunity, and detects the spectrum holes in the spatial angle dimension. utilization.

Description

A Multi-Antenna-Based Spatial Spectrum Sensing Method

technical field

The invention belongs to the technical field of cognitive radio, and in particular relates to a multi-antenna-based spatial spectrum sensing method.

Background technique

With the continuous development of wireless communication, spectrum resources are becoming more and more scarce, which seriously restricts the development of communication technology. To promote the development of wireless communication, it is necessary to improve the utilization rate of spectrum resources. Among them, cognitive radio (CR) technology is an effective method to solve the shortage of spectrum and improve the utilization rate of spectrum resources. The basic idea of CR is spectrum sharing or spectrum reuse, and one of its features is to allow unlicensed cognitive users (CUs) opportunities to access licensed frequency bands without interfering with licensed primary user (PU) communications. To achieve this, a Cognitive User (CU) system must constantly detect whether a Licensed Primary User (PU) is occupying a licensed frequency band, a process known as spectrum sensing.

Although the spectrum sensing algorithm of the traditional dimension improves the detection performance to a certain extent, it mainly detects in the frequency dimension, time dimension and geographic latitude, and the spectrum development capability is limited. On the other hand, the rapid development of multi-antenna technology and the application of 5G large-scale antenna arrays enable mobile terminals and base stations to have the ability to identify angles, which promotes the development of angle-dimensional spectrum resources. If the angle of arrival of the signal space angle dimension is estimated, through beamforming technology, it is possible to avoid the primary user (PU) communication direction or design a null antenna beam for the primary user communication direction, so that the cognitive user can operate at the same frequency. , At the same time or even the same place, avoiding the communication direction of the main user, and performing spectrum access through different spatial angles, thereby increasing system capacity and improving spectrum utilization.

SUMMARY OF THE INVENTION

The present invention proposes a multi-antenna-based spatial spectrum sensing method, aiming at increasing system capacity and improving spectrum utilization.

The technical scheme of the present invention is:

For a cognitive user (CU) carrying multiple antennas, the antennas are isotropic M-element uniform circular arrays (Uniform Circular Array, UCA). Assuming that there are D (D≤M) far-field main user signals incident on the M-element uniform circular array from different directions in the space, and the center of the uniform circular array is taken as the reference point, then the i-th main user of array element j is reached. The signal is:

为空间信号的载波。由于信号满足窄带假设条件，则z_i(t-τ)≈z_i(t)，经过传播延迟τ后的信号可以表示为：where h _ij represents the channel gain between the i-th primary user signal _si (t) and the j-th receiving antenna, _zi (t) is the complex envelope of the i-th primary user signal, including signal information,

is the carrier of the space signal. Since the signal satisfies the narrowband assumption, then _zi (t-τ) _≈zi (t), the signal after the propagation delay τ can be expressed as:

Ideally, the signal received by the jth array element can be expressed as:

Among them, τ _ij is the time delay relative to the reference point when the i-th primary user signal arrives at the array element j, and w _j (t) is the additive white Gaussian noise with a variance of σ ² on the array element j.

The spatial spectrum sensing method of the present invention comprises the following steps:

S1. The array antenna samples the received signal N times, then the signal received by each array antenna of the cognitive user is expressed as:

为信号传播时延造成的相位差，

i＝1,2,...,D表示第i个主用户信号，j＝1,2,…,M表示第j个接收天线，θ_i和

分别表示第i个主用户信号的方位角和仰角，n＝0,1,…,N-1表示第n个采样序号，λ表示波长、

表示载波角频率；令M个阵列天线的接收数据构成一个M×N维矩阵：in,

is the phase difference caused by the signal propagation delay,

i=1,2,...,D represents the i-th primary user signal, j=1,2,...,M represents the j-th receiving antenna, θ _i and

respectively represent the azimuth and elevation of the i-th primary user signal, n=0,1,...,N-1 represents the n-th sampling number, λ represents the wavelength,

represents the carrier angular frequency; let the received data of M array antennas form an M×N-dimensional matrix:

表示主用户和认知用户接收天线的信道增益矩阵，‘.*’表示矩阵点乘，

为信号矩阵，

为阵列流行，

为加性噪声矩阵；in,

Represents the channel gain matrix of the receiving antennas of the primary user and the cognitive user, '.*' represents the matrix dot product,

is the signal matrix,

Pop for arrays,

is the additive noise matrix;

S2. Calculate the sample covariance matrix

然后对

进行特征值分解得到M个特征值及其对应的特征向量，从而获得

的最大特征值

迹

以及特征值几何平均

Obtain the estimated autocorrelation function from the sampling sequence

then right

Perform eigenvalue decomposition to obtain M eigenvalues and their corresponding eigenvectors, thereby obtaining

The largest eigenvalue of

trace

and the geometric mean of the eigenvalues

S3. Take α∈[0,1], and calculate the test statistic T of the fusion detection algorithm:

According to the random matrix theory, the false alarm probability P _fa is obtained:

σ²为高斯白噪声w(n)的方差、

F_TW(·)为一阶Tracy-Widom分布；根据虚警概率P_fa，确定判决门限γ：in,

σ ² is the variance of white Gaussian noise w(n),

F _TW (·) is the first-order Tracy-Widom distribution; according to the false alarm probability P _fa , the decision threshold γ is determined:

为一阶Tracy-Widom分布的逆；in

is the inverse of the first-order Tracy-Widom distribution;

S4. Compare the statistic T with the decision threshold γ:

If the test statistic T is greater than the decision threshold γ, then the subband is occupied, the primary user exists, and the process goes to step S5;

If the test statistic T is less than the decision threshold γ, the subband is not occupied, the primary user does not exist, and the cognitive user directly performs spectrum access;

将步骤b中得到的样本协方差矩阵的特征值从小到大排列，即λ₁≥…≥λ_D>λ_D+1≥…≥λ_M，V＝[q₁,q₂,...,q_M]是对应的特征值，计算γ_k＝λ_k/λ_k+1，k＝1,2,…,M-1，取主信号数的估计值

为使得γ_k＝max(γ₁,γ₂,…,γ_M-1),k＝1,2,…,M-1时的k值；S5. Estimate the number of main signals

Arrange the eigenvalues of the sample covariance matrix obtained in step b from small to large, that is, λ ₁ ≥...≥λ _D >λ _D+1 ≥...≥λ _M , V=[q ₁ ,q ₂ ,..., q _M ] is the corresponding eigenvalue, calculate γ _k =λ _k /λ _k+1 , k=1,2,...,M-1, take the estimated value of the number of main signals

In order to make γ _k =max(γ ₁ ,γ ₂ ,...,γ _M-1 ), the k value when k=1,2,...,M-1;

构造

维的噪声子空间

按照

计算Music空间谱，并搜索Music空间，找出

个峰值，从而得到主信号DOA估计值，认知用户通过波束成形技术对避开主用户通信方向进行频谱接入。S6. Perform DOA estimation on the main signal: estimate the value according to the number of main signals

structure

dimensional noise subspace

according to

Calculate the Music space spectrum, and search the Music space to find out

A peak value is obtained to obtain an estimated DOA value of the main signal, and the cognitive user uses the beamforming technology to perform spectrum access to the communication direction that avoids the main user.

The beneficial effect of the present invention is that the spatial angle dimension information is used as a new spectrum opportunity to detect the spectrum holes in the spatial angle dimension. Compared with the spectrum sensing algorithm of the traditional dimension, although the implementation complexity is increased, the system capacity and improve spectrum utilization.

Description of drawings

Fig. 1 is the system diagram of the spatial spectrum sensing scheme of the present invention;

Fig. 2 is a uniform circular array (UCA) model diagram;

Figure 3 and Figure 5 are schematic diagrams of detection probability VS signal-to-noise ratio when α∈[0.1,1] under Gaussian channel and Rayleigh fading channel respectively;

FIG. 4 and FIG. 6 are schematic diagrams of DOA estimation root mean square error (RMSE) VS signal-to-noise ratio under Gaussian channel and Rayleigh fading channel, respectively.

Detailed ways

The technical solution of the present invention has been described in detail in the section of the content of the invention, and the practicability of the present invention is described below in conjunction with a simulation example.

Assuming that there is only one primary user (D=1) with frequency f, the transmit signal is a QPSK signal, the number of uniform circular array antennas is M=16, and the number of sampling points is N=10000.

First, the relationship between the signal-to-noise ratio and detection probability of the detection scheme with different α values is compared. The simulation results are shown in Fig. In this simulation, set the false alarm probability P _fa =0.01, SNR = -24:2:4, and the number of Monte Carlo simulations under different signal-to-noise ratios (SNR) is 2000 times. It can be seen from Figure 3 and Figure 5 that when α∈[0.1,1], the smaller the value of α, the better the detection performance of the detection scheme used in this scheme; when α=0.5 and α=1, the detection performance of this scheme used in this scheme is better. The detection scheme is equivalent to the ME-GM (maximum-eigenvalue-geometric-mean) algorithm and the MET (maximum-eigenvalue-trace) algorithm respectively, and it can be seen from Figure 3 and Figure 5 that when α≤0.4, this scheme uses The detection performance of the detection scheme is better than that of the ME-GM algorithm and the MET algorithm.

Compare the root mean square error (RMSE) of DOA estimation under different signal-to-noise ratios, set the main signal direction as (θ, φ)=(125°, 80.1°), SNR=-22:2:4, different signal-to-noise ratios The number of Monte Carlo simulations under (SNR) is 200. It can be seen from Figure 4 and Figure 6 that when SNR≥-15dB, the root mean square error RMSE of DOA estimation is less than 1°, the DOA estimation scheme used can more accurately estimate the direction of arrival of the main user signal, and the uniform circular array 360° omnidirectional estimation can be achieved. After estimating the DOA of the primary user signal, using beamforming technology, cognitive users can avoid the access direction of the primary user for spectrum access and improve spectrum utilization, which also proves that this solution can improve spectrum utilization and increase system size. capacity.

Claims (1)

1. A space domain spectrum sensing method based on multiple antennas is characterized in that for cognitive users carrying multiple antennas, the antennas are isotropic M-element uniform circular arrays, and D far-field main user signals enter the M-element uniform circular arrays from different directions in space, and the method comprises the following steps:

s1, the array antennas perform N times of sampling on the received signal, and the signal received by each array antenna of the cognitive user is represented as:

wherein,

for the phase difference caused by the propagation delay of the signal,

1,2, D denotes the ith primary user signal, 1,2, …, M denotes the jth receiving antenna, θ_iAnd

respectively indicates the azimuth angle and the elevation angle of the ith primary user signal, N is 0,1, …, N-1 indicates the nth sampling number, and λ indicates the wavelength,

Represents the carrier angular frequency; let the received data of M array antennas form an M × N dimensional matrix:

wherein,

the channel gain matrix of the receiving antenna of the main user and the cognitive user is shown,' indicates the dot product of the matrix,

in the form of a matrix of signals,

in order for the array to be popular,

is an additive noise matrix;

s2, calculating a sample covariance matrix

Obtaining an estimated autocorrelation function by sampling a sequence

Then to

Decomposing the eigenvalues to obtain M eigenvalues and corresponding eigenvectors thereof, thereby obtaining

Maximum eigenvalue of

Trace

And geometric mean of eigenvalues

S3, taking α E [0,1], calculating a test statistic T of the fusion detection algorithm:

obtaining false alarm probability P according to random matrix theory_fa：

Wherein,

σ²is the variance of white Gaussian noise w (n),

F_TW(.) is a first order Tracy-Widom distribution; according to false alarm probability P_faDetermining a decision threshold gamma:

wherein

Is the inverse of the first-order Tracy-Widom distribution;

s4, comparing the statistic T with a decision threshold gamma:

if the test statistic T is larger than the judgment threshold gamma, the sub-band is occupied, a master user exists, and the step S5 is carried out;

if the test statistic T is smaller than the judgment threshold gamma, the sub-band is not occupied, a master user does not exist, and the cognitive user directly performs spectrum access;

s5, estimating the number of main signals

C, arranging the eigenvalues of the sample covariance matrix obtained in the step b from small to large, namely lambda₁≥…≥λ_D>λ_D+1≥…≥λ_M，V＝[q₁,q₂,...,q_M]Is a corresponding characteristic value, calculating gamma_k＝λ_k/λ_k+1K is 1,2, …, M-1, and an estimate of the number of primary signals is taken

To make gamma_k＝max(γ₁,γ₂,…,γ_M-1) K is 1,2, …, value of k at M-1;

s6, DOA estimation is carried out on the main signal: estimating the value according to the number of main signals

Structure of the device

Noise subspace of dimension

According to

Calculating Music space spectrum, searching Music space and finding out

Peak value, thereby obtaining the DOA estimated value of the main signal and cognizing the userAnd performing spectrum access on the direction avoiding the main user communication through a beam forming technology.

Images