CN109274444B - Space-time-frequency three-dimensional spectrum sensing method - Google Patents

Space-time-frequency three-dimensional spectrum sensing method Download PDF

Info

Publication number
CN109274444B
CN109274444B CN201811124363.0A CN201811124363A CN109274444B CN 109274444 B CN109274444 B CN 109274444B CN 201811124363 A CN201811124363 A CN 201811124363A CN 109274444 B CN109274444 B CN 109274444B
Authority
CN
China
Prior art keywords
matrix
dimensional
frequency
time
signal
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
Application number
CN201811124363.0A
Other languages
Chinese (zh)
Other versions
CN109274444A (en
Inventor
杜利平
薛慧
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Science and Technology Beijing USTB
Original Assignee
University of Science and Technology Beijing USTB
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by University of Science and Technology Beijing USTB filed Critical University of Science and Technology Beijing USTB
Priority to CN201811124363.0A priority Critical patent/CN109274444B/en
Publication of CN109274444A publication Critical patent/CN109274444A/en
Application granted granted Critical
Publication of CN109274444B publication Critical patent/CN109274444B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B17/00—Monitoring; Testing
    • H04B17/30—Monitoring; Testing of propagation channels
    • H04B17/382—Monitoring; Testing of propagation channels for resource allocation, admission control or handover
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
    • H04W16/02—Resource partitioning among network components, e.g. reuse partitioning
    • H04W16/10—Dynamic resource partitioning

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Electromagnetism (AREA)
  • Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

The invention provides a space-time-frequency three-dimensional spectrum sensing method which can obtain time, frequency and angle information of a main user signal. The method comprises the following steps: whitening preprocessing is carried out on main user signals received by a plurality of cognitive users; performing space-time frequency distribution transformation on the signal subjected to whitening preprocessing to obtain a four-dimensional matrix, wherein information in the four-dimensional matrix comprises: time, frequency and two-dimensional spatial information; filtering time-frequency two-dimensional information of the four-dimensional matrix, and extracting relevant points of the single source; performing joint diagonalization on the obtained single-source autocorrelation points to obtain a guide vector estimation for angle estimation; and determining the arrival angle of each main user signal according to the obtained guide vector. The invention relates to the technical field of radio frequency spectrum sensing.

Description

Space-time-frequency three-dimensional spectrum sensing method
Technical Field
The invention relates to the technical field of radio frequency spectrum sensing, in particular to a space-time-frequency three-dimensional frequency spectrum sensing method.
Background
In recent years, with the development of global wireless communication technology, new radio systems with different functions have been developed. At present, a static spectrum allocation strategy is adopted by a wireless communication system, and spectrum resources are in increasing shortage. The cognitive radio can sense the change of the external environment, improve the spectrum utilization rate and the system capacity, and solve the problem of spectrum shortage. Currently, commonly used spectrum sensing methods are mainly divided into single-cognitive user spectrum sensing and multi-cognitive user spectrum sensing technologies.
Most of the existing spectrum sensing technologies can only obtain the existence of a main user signal, and cannot obtain detailed time, frequency and angle information.
Disclosure of Invention
The invention aims to provide a space-time-frequency three-dimensional spectrum sensing method to solve the problem that the spectrum sensing technology in the prior art cannot obtain detailed time, frequency and angle information.
In order to solve the above technical problem, an embodiment of the present invention provides a space-time-frequency three-dimensional spectrum sensing method, including:
carrying out whitening pretreatment on main user signals received by a plurality of cognitive users, wherein the main user signals are not related to each other;
performing space-time frequency distribution transformation on the signal subjected to whitening preprocessing to obtain a four-dimensional matrix, wherein information in the four-dimensional matrix comprises: time, frequency and two-dimensional spatial information;
filtering time-frequency two-dimensional information of the four-dimensional matrix, and extracting a single source from a relevant point, wherein the single source refers to each independent main user signal source;
performing joint diagonalization on the obtained single-source autocorrelation points to obtain a guide vector estimation for angle estimation;
and determining the arrival angle of each main user signal according to the obtained guide vector.
Further, the whitening preprocessing the main user signals received by the multiple cognitive users comprises:
forming a receiving signal matrix by using main user signals received by a plurality of cognitive users;
and carrying out whitening preprocessing on the received signal matrix.
Further, the forming a receiving signal matrix by the primary user signals received by the multiple cognitive users includes:
the method comprises the following steps that N master users transmit signals from different directions to reach M cognitive user sides, wherein M is larger than N;
forming a received signal matrix Y (Y) by the signals received by M cognitive users1,Y2,...,YM]TWhere superscript T represents the transpose of the matrix.
Further, the whitening preprocessing the received signal matrix includes:
carrying out eigenvalue decomposition on a covariance matrix of a received signal matrix Y;
constructing a whitening matrix W according to the eigenvalue decomposition result;
the whitening matrix W is point-multiplied with the received signal matrix Y to obtain a whitening signal Z ═ W.Y, wherein Z ═ Z1,Z2,...,ZN]T。
Further, the constructing the whitening matrix W according to the eigenvalue decomposition result includes:
the obtained characteristic values are subjected to descending order arrangement to obtain characteristic values [ lambda ] after descending order arrangement1,...,λM]And its corresponding feature vector [ h ]1,...,hM]；
Taking the first 1-N eigenvalues and eigenvectors thereof from the M eigenvalues after descending order arrangement to construct a whitening matrix W, wherein the whitening matrix W is expressed as:
Figure BDA0001812064960000021
wherein σ2The superscript H represents the conjugate transpose of the matrix for the average of the next M-N smaller eigenvalues after descending order.
Further, the four-dimensional matrix obtained by the space-time-frequency distribution transform is represented as:
Figure BDA0001812064960000031
wherein D isZZ(t, f) is a two-dimensional covariance matrix representing spatial information; each element
Figure BDA0001812064960000032
Is a two-dimensional time-frequency matrix,
Figure BDA0001812064960000033
representing the ith signal ZiAnd the jth signal ZjThe mutual time-frequency distribution of (1); t represents time; f represents frequency.
Further, the filtering the time-frequency two-dimensional information of the four-dimensional matrix and extracting the single-source autocorrelation points includes:
by means of a relational expression
Figure BDA0001812064960000034
Extracting autocorrelation points containing all primary user signals, wherein epsilon2To extract the threshold, trace (·) represents the trace of the matrix;
dividing the extracted autocorrelation points of all the main user signals into N classes through clustering, wherein one class represents the autocorrelation point of one signal to obtain K of N classes of single-source signalsiAn autocorrelation point
Figure BDA0001812064960000035
Wherein, i is 1,2i＝1,2,...,Ki。
Further, the jointly diagonalizing the obtained single-source autocorrelation points to obtain a steering vector estimation for angle estimation comprises:
for the obtained N-type autocorrelation points
Figure BDA0001812064960000036
Extracting two-dimensional space information of the four-dimensional matrix to form a correlation matrix of the signal on corresponding time frequency points
Figure BDA0001812064960000037
For the obtained correlation matrix
Figure BDA0001812064960000038
Performing joint diagonalization to obtain
Figure BDA0001812064960000039
A minimum NxN dimensional unitary matrix U, where I is an identity matrix and off (-) is the sum of the squares of the non-diagonal elements of the matrix;
and (3) carrying out guide vector estimation by using a unitary matrix U obtained by joint diagonalization:
Figure BDA00018120649600000310
wherein, W#In order to whiten the pseudo-inverse of the matrix,
Figure BDA00018120649600000311
is a steering vector.
Further, the determining the arrival angle of each primary user signal according to the obtained steering vector includes:
and determining the arrival angle of each main user signal by utilizing multiple signal classification according to the obtained guide vector.
Further, the determining the arrival angle of each primary user signal by using multiple signal classification according to the obtained steering vector comprises:
to pair
Figure BDA0001812064960000041
Coordination of the best of the Chinese traditional medicineVariance matrix:
Figure BDA0001812064960000042
wherein R isAAA covariance matrix of dimension M × M;
r is to beAAThe feature vector of (2) is divided into two parts according to the size of the feature value: u shapesAnd UnWherein, Us·Us H+Un·Un H＝I，Us＝[U1,U2,...,UN]Stretching a unitary space formed by eigenvectors corresponding to the first N maximum eigenvalues into a signal subspace; u shapen＝[UN+1,...,UM]Expanding a unitary matrix formed by eigenvectors corresponding to the last M-N minimum eigenvalues into a noise subspace;
by passing
Figure BDA0001812064960000043
And estimating the angle of the main user signal to obtain the arrival angle of each main user signal.
The technical scheme of the invention has the following beneficial effects:
in the scheme, whitening preprocessing is carried out on main user signals received by a plurality of cognitive users; performing space-time frequency distribution transformation on the signal subjected to whitening preprocessing to obtain a four-dimensional matrix, wherein information in the four-dimensional matrix comprises: time, frequency and two-dimensional spatial information; filtering the time-frequency two-dimensional information of the four-dimensional matrix, and extracting the relevant points of the single sources so as to obtain the accurate time-frequency distribution of each main user signal; performing joint diagonalization on the obtained single-source autocorrelation points to obtain a guide vector estimation for angle estimation; and determining the arrival angle of each main user signal according to the obtained guide vector. Therefore, the accurate time-frequency distribution map of the signal is obtained by extracting the autocorrelation point of the signal based on the space-time frequency distribution transformation, and the estimation accuracy and the resolution of the main user angle are improved.
Drawings
Fig. 1 is a schematic flowchart of a space-time-frequency three-dimensional spectrum sensing method according to an embodiment of the present invention;
fig. 2 is a schematic diagram of multiple primary users and cognitive users in a cognitive radio network according to an embodiment of the present invention;
fig. 3 is a schematic time-frequency energy diagram of a space-time-frequency three-dimensional spectrum sensing method for a signal 1 according to an embodiment of the present invention;
fig. 4 is a schematic time-frequency energy diagram of a space-time-frequency three-dimensional spectrum sensing method for a signal 2 according to an embodiment of the present invention;
fig. 5 is a schematic diagram of angle estimation of a space-time-frequency three-dimensional spectrum sensing method for a signal 1 according to an embodiment of the present invention;
fig. 6 is a schematic diagram of angle estimation of a space-time-frequency three-dimensional spectrum sensing method for a signal 2 according to an embodiment of the present invention.
Detailed Description
In order to make the technical problems, technical solutions and advantages of the present invention more apparent, the following detailed description is given with reference to the accompanying drawings and specific embodiments.
The invention provides a space-time-frequency three-dimensional spectrum sensing method aiming at the problem that the existing spectrum sensing technology cannot obtain detailed time, frequency and angle information.
As shown in fig. 1, a space-time-frequency three-dimensional spectrum sensing method provided in an embodiment of the present invention includes:
s101, whitening preprocessing is carried out on main user signals received by a plurality of cognitive users, wherein the main user signals are not related to each other;
s102, performing space-Time-frequency distribution (STFD) transformation on the whitened and preprocessed signal to obtain a four-dimensional matrix, where information in the four-dimensional matrix includes: time, frequency and two-dimensional spatial information;
s103, filtering time-frequency two-dimensional information of the four-dimensional matrix, and extracting a single source from a relevant point, wherein the single source refers to each independent main user signal source;
s104, performing joint diagonalization on the obtained single-source autocorrelation points to obtain a guide vector estimation for angle estimation;
and S105, determining the arrival angle of each main user signal according to the obtained guide vector.
The space-time-frequency three-dimensional spectrum sensing method provided by the embodiment of the invention is used for carrying out whitening pretreatment on main user signals received by a plurality of cognitive users; performing space-time frequency distribution transformation on the signal subjected to whitening preprocessing to obtain a four-dimensional matrix, wherein information in the four-dimensional matrix comprises: time, frequency and two-dimensional spatial information; filtering the time-frequency two-dimensional information of the four-dimensional matrix, and extracting the relevant points of the single sources so as to obtain the accurate time-frequency distribution of each main user signal; performing joint diagonalization on the obtained single-source autocorrelation points to obtain a guide vector estimation for angle estimation; and determining the arrival angle of each main user signal according to the obtained guide vector. Therefore, the accurate time-frequency distribution map of the signal is obtained by extracting the autocorrelation point of the signal based on the space-time frequency distribution transformation, and the estimation accuracy and the resolution of the main user angle are improved.
In a specific implementation manner of the foregoing space-time-frequency three-dimensional spectrum sensing method, further, the whitening preprocessing performed on the primary user signals received by the multiple cognitive users includes:
forming a receiving signal matrix by using main user signals received by a plurality of cognitive users;
and carrying out whitening preprocessing on the received signal matrix.
In a specific implementation manner of the foregoing space-time-frequency three-dimensional spectrum sensing method, further, the forming a received signal matrix by using primary user signals received by a plurality of cognitive users includes:
the method comprises the following steps that N master users transmit signals from different directions to reach M cognitive user sides, wherein M is larger than N;
forming a received signal matrix Y (Y) by the signals received by M cognitive users1,Y2,...,YM]TWhere superscript T represents the transpose of the matrix.
In this embodiment, it is assumed that N primary users transmit signals from different directions and reach M (M > N) cognitive user terminals at angles θ1,θ2,...,θN. The main user signal sources are mutually independent, the main user signals are mutually uncorrelated, and the main user signals and noise are mutually uncorrelated. Each cognitive user sends the received signal to a decision center, and the decision center forms a received signal matrix Y ═ Y1,Y2,...,YM]T。
In an embodiment of the foregoing space-time-frequency three-dimensional spectrum sensing method, further performing whitening preprocessing on the received signal matrix includes:
carrying out eigenvalue decomposition on a covariance matrix of a received signal matrix Y;
constructing a whitening matrix W according to the eigenvalue decomposition result;
the whitening matrix W is point-multiplied with the received signal matrix Y to obtain a whitening signal Z ═ W.Y, wherein Z ═ Z1,Z2,...,ZN]T。
In this embodiment, the specific step of constructing the whitening matrix W according to the eigenvalue decomposition result may include:
carrying out eigenvalue decomposition on a covariance matrix of a received signal matrix Y;
the obtained characteristic values are subjected to descending order arrangement to obtain characteristic values [ lambda ] after descending order arrangement1,...,λM]And its corresponding feature vector [ h ]1,...,hM]And taking the first 1-N eigenvalues and eigenvectors thereof from the M eigenvalues after descending order arrangement to construct the whitening matrix W, namely, the formula of the whitening matrix W only uses 1-N eigenvalues and eigenvectors thereof.
In this embodiment, the whitening matrix W is represented as:
Figure BDA0001812064960000071
wherein σ2The superscript H represents the conjugate transpose of the matrix for the average of the next M-N smaller eigenvalues after descending order.
In a specific embodiment of the foregoing space-time-frequency three-dimensional spectrum sensing method, further, a four-dimensional matrix obtained by space-time-frequency distribution transformation is represented as:
Figure BDA0001812064960000072
wherein D isZZ(t, f) is a two-dimensional covariance matrix representing spatial information; each element
Figure BDA0001812064960000073
Is a two-dimensional time-frequency matrix,
Figure BDA0001812064960000074
representing the ith signal ZiAnd the jth signal ZjThe mutual time-frequency distribution of (1); t represents time; f represents frequency.
In this embodiment, the signal after whitening preprocessing is subjected to space-time-frequency distribution transform to obtain a four-dimensional matrix DZZ(t, f) in the specific form:
Figure BDA0001812064960000075
in this example, DZZ(t, f) is a two-dimensional covariance matrix in which each element
Figure BDA0001812064960000076
And is a two-dimensional time-frequency matrix which is finally expressed in a four-dimensional form.
In this embodiment, the four-dimensional matrix D is obtainedZZAnd (t, f) extracting autocorrelation points, performing characteristic value decomposition on the extracted autocorrelation points, and realizing signal parameter estimation by using a subspace method.
In a specific implementation manner of the aforementioned space-time-frequency three-dimensional spectrum sensing method, further, the filtering the time-frequency two-dimensional information of the four-dimensional matrix, and extracting the single-source autocorrelation point includes:
by means of a relational expression
Figure BDA0001812064960000081
Extracting signals containing all primary usersWhere e is2To extract the threshold, trace (·) represents the trace of the matrix;
dividing the extracted autocorrelation points of all the main user signals into N classes through clustering, wherein one class represents the autocorrelation point of one signal to obtain K of N classes of single-source signalsiAn autocorrelation point
Figure BDA0001812064960000082
Wherein, i is 1,2i＝1,2,...,Ki。
In this example,. epsilon2Is a positive real number close to but less than 1, is noise dependent, and thus the autocorrelation points can be extracted.
In this example, KiThe value of (1) is the number of the time-frequency points of each type in the N types extracted by all the extraction steps, is not a determined value, and different results can be generated in each operation of the method.
In an embodiment of the foregoing space-time-frequency three-dimensional spectrum sensing method, further, the jointly diagonalizing the obtained single-source autocorrelation points to obtain a steering vector estimate for angle estimation includes:
for the obtained N-type autocorrelation points
Figure BDA0001812064960000083
Extracting two-dimensional space information of the four-dimensional matrix to form a correlation matrix of the signal on corresponding time frequency points
Figure BDA0001812064960000084
For the obtained correlation matrix
Figure BDA0001812064960000085
Performing joint diagonalization to obtain
Figure BDA0001812064960000086
A minimum NxN dimensional unitary matrix U, where I is an identity matrix and off (-) is the sum of the squares of the non-diagonal elements of the matrix;
and (3) carrying out guide vector estimation by using a unitary matrix U obtained by joint diagonalization:
Figure BDA0001812064960000087
wherein, W#In order to whiten the pseudo-inverse of the matrix,
Figure BDA0001812064960000088
is a steering vector.
In an embodiment of the foregoing space-time-frequency three-dimensional spectrum sensing method, further, the determining, according to the obtained steering vector, an angle of arrival of each primary user signal includes:
and determining the arrival angle of each main user Signal by using Multiple Signal Classification (MUSIC) according to the obtained guide vector.
In an embodiment of the foregoing space-time-frequency three-dimensional spectrum sensing method, further, the determining, according to the obtained steering vector, an angle of arrival of each primary user signal by using multiple signal classification includes:
to pair
Figure BDA0001812064960000089
Solving a covariance matrix:
Figure BDA0001812064960000091
wherein R isAAA covariance matrix of dimension M × M;
r is to beAAThe feature vector of (2) is divided into two parts according to the size of the feature value: u shapesAnd UnWherein, Us·Us H+Un·Un H＝I，Us＝[U1,U2,...,UN]Stretching a unitary space formed by eigenvectors corresponding to the first N maximum eigenvalues into a signal subspace; u shapen＝[UN+1,...,UM]Expanding noise for unitary matrix formed by eigenvector corresponding to last M-N minimum eigenvaluesA subspace;
by passing
Figure BDA0001812064960000092
And estimating the angle of the main user signal to obtain the arrival angle of each main user signal.
In order to better understand the space-time-frequency three-dimensional spectrum sensing method according to the embodiment of the present invention, a specific embodiment is described, as shown in fig. 2, where N is 2, M is 6:
in the first step, N main users transmit signals from different directions (supposing that the main user 1 transmits a signal 1 and the main user 2 transmits a signal 2) and reach M cognitive user terminals, the arrival angles are-2 degrees and 2 degrees respectively, the instantaneous frequencies of the two signals are 1e9 and 2e9 respectively, and the signal-to-noise ratio is-5 dB.
In this example, 6 mixed signals containing white noise were observed through 2 "statistically" independent signals.
In this embodiment, N primary users transmit signals from different directions and reach M (M > N) cognitive user terminals at angles θ respectively1,θ2,...,θN. The main user signals are not related to each other, and the main user signals and noise are not related to each other. Each cognitive user sends the received signal to a decision center, and the decision center forms a received signal matrix Y ═ Y1,Y2,...,YM]T。
Step two, whitening preprocessing is carried out on the received signal matrix, and the specific steps are as follows:
a21, decomposing the characteristic value of the covariance matrix of the received signal, and setting [ lambda ]1,...,λM]For the descending order of the eigenvalues, [ h ]1,...,hM]For corresponding feature vector, σ2Is the average of the last M-N smaller eigenvalues of the covariance matrix. The whitening matrix W is defined as:
Figure BDA0001812064960000093
wherein, the superscript H is the conjugate transpose of the matrix.
A22, the whitened signal is:
Z＝W·Y
wherein Z ═ Z1,Z2,...,ZN]T。
Step three, carrying out space-time frequency distribution transformation on the signals after whitening pretreatment to obtain a four-dimensional matrix DZZ(t, f) in the specific form:
Figure BDA0001812064960000101
wherein, DZZ(t, f) is a two-dimensional covariance matrix representing spatial information; each element
Figure BDA0001812064960000102
Is a two-dimensional time-frequency matrix,
Figure BDA0001812064960000103
representing the ith signal ZiAnd the jth signal ZjThe mutual time-frequency distribution of (1); t represents time; f represents frequency.
Step four, filtering the time-frequency two-dimensional information of the four-dimensional matrix, extracting the relevant points of the single sources, and obtaining the accurate time-frequency distribution of each main user signal, wherein the specific steps are as follows:
a41, the process of extracting the autocorrelation points including all the main user signals may represent extracting time-frequency points satisfying the following formula:
Figure BDA0001812064960000104
wherein trace (. epsilon.) represents the trace of the matrix2Is a positive real number close to but less than 1, is noise dependent, and thus the autocorrelation points can be extracted.
A42, dividing the autocorrelation points of all main user signals into N classes by a clustering method, wherein one class represents the autocorrelation point of one signal to obtain K of N classes of single-source signalsiAn autocorrelation point
Figure BDA0001812064960000105
Wherein i is 1,2i＝1,2,...,Ki。
And fifthly, performing joint diagonalization on the four-dimensional matrix subjected to filtering processing to obtain a guide vector estimation for angle estimation, wherein the specific steps are as follows:
a51, obtaining the autocorrelation points of N types
Figure BDA0001812064960000106
Extracting a four-dimensional matrix DZZThe two-dimensional space information forms a correlation matrix on corresponding time frequency points
Figure BDA0001812064960000107
A52 to the obtained
Figure BDA0001812064960000108
Joint diagonalization is performed to obtain an N × N dimensional unitary matrix U that minimizes the following equation:
Figure BDA0001812064960000111
where I is the identity matrix and off (-) is the sum of the squares of the off-diagonal elements of the matrix.
A53, performing steering vector estimation by using unitary matrix U obtained by joint diagonalization:
Figure BDA0001812064960000112
wherein, W#In order to whiten the pseudo-inverse of the matrix,
Figure BDA0001812064960000113
is a steering vector.
And step six, solving the angle estimation of the signal by the obtained guide vector estimation by using an MUSIC method, wherein the specific steps are as follows:
a61, to
Figure BDA0001812064960000114
Solving a covariance matrix:
Figure BDA0001812064960000115
wherein R isAAA covariance matrix of dimension M × M;
a62, adding RAAThe feature vector of (2) is divided into two parts according to the size of the feature value: u shapesAnd UnWherein, Us·Us H+Un·Un H＝I，Us＝[U1,U2,...,UN]Stretching a unitary space formed by eigenvectors corresponding to the first N maximum eigenvalues into a signal subspace; u shapen＝[UN+1,...,UM]Expanding a unitary matrix formed by eigenvectors corresponding to the last M-N minimum eigenvalues into a noise subspace;
a63, estimating the angle of the primary user signal using the following formula:
Figure BDA0001812064960000116
in the embodiment of the present invention, fig. 3 and 4 describe that filtering processing such as signal autocorrelation points and clustering is performed on time-frequency two-dimensional information of a four-dimensional matrix, and single-source autocorrelation points are extracted to obtain accurate time-frequency estimation of two main user signals. The abscissa in fig. 3 and 4 is the time operator(s), the ordinate is the instantaneous frequency of the signal, and the result is obtained by repeating the experiment under the conditions that the number of signals is 2, the cognitive user terminal is 6, the number of snapshots is 1000, the time interval is 1s, and the frequencies of the two signals are 1e9HZ and 2e9HZ, respectively. As can be seen from fig. 3 and 4, the time-frequency diagram obtained in the embodiment of the present invention not only reduces the problems of noise, etc. in the system, but also obtains better performance. This demonstrates that the proposed method is comprehensive and efficient. Meanwhile, the result also shows that the method for researching the space-time-frequency three-dimensional spectrum sensing provided by the embodiment of the invention has excellent performance.
Fig. 5 and 6 illustrate the angle of each signal obtained by joint diagonalization, MUSIC, etc. using STFD matrix composed of single-derived correlation points. The abscissa in fig. 5 and fig. 6 is the azimuth, the ordinate is the peak value of the signal, and the result is obtained by repeating the experiment under the conditions that the number of the signals is 2, the number of the cognitive user terminals is 6, the number of the snapshots is 1000, the time interval is 1s, the angles of the two signals are-2 ° and 2 ° respectively. As can be seen from fig. 5 and 6, the angle diagram obtained by the present invention not only has a sharper peak, but also has a greatly improved resolution. This demonstrates that the proposed method is comprehensive and efficient. Meanwhile, the result also shows that the method for researching the space-time-frequency three-dimensional spectrum sensing provided by the embodiment of the invention has excellent performance.
In summary, the method of the design can obtain more concentrated autocorrelation points in the spectrum sensing process, reduce the influence of noise, and improve the resolution of signals, and has the following advantages:
1) the STFD matrix can simultaneously contain the spectrum information of three dimensions of space, time and frequency;
2) filtering processing such as signal autocorrelation points and clustering is carried out on the time-frequency two-dimensional information of the four-dimensional matrix, so that autocorrelation points of each signal can be extracted, and the calculation amount of angle calculation is reduced; meanwhile, the interference and noise of cross terms can be effectively inhibited under the condition of Gaussian white noise, and the time-frequency convergence is higher;
3) the space time-frequency matrix is processed by adopting the joint diagonalization and the classical MUSIC method, so that the robustness Of the algorithm is enhanced, the DOA (Direction Of Arrival) estimation capability Of the algorithm under the condition Of low signal-to-noise ratio is improved, and the angle resolution and accuracy are improved.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions.
While the foregoing is directed to the preferred embodiment of the present invention, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention as defined in the appended claims.

Claims (6)

1. A space-time-frequency three-dimensional spectrum sensing method is characterized by comprising the following steps:
carrying out whitening pretreatment on main user signals received by a plurality of cognitive users, wherein the main user signals are not related to each other;
performing space-time frequency distribution transformation on the signal subjected to whitening preprocessing to obtain a four-dimensional matrix, wherein information in the four-dimensional matrix comprises: time, frequency and two-dimensional spatial information;
filtering time-frequency two-dimensional information of the four-dimensional matrix, and extracting a single source from a relevant point, wherein the single source refers to each independent main user signal source;
performing joint diagonalization on the obtained single-source autocorrelation points to obtain a guide vector estimation for angle estimation;
determining the arrival angle of each main user signal according to the obtained guide vector;
the whitening preprocessing of the main user signals received by the multiple cognitive users comprises the following steps:
forming a receiving signal matrix by using main user signals received by a plurality of cognitive users;
carrying out whitening pretreatment on a received signal matrix;
wherein, the forming a receiving signal matrix by the main user signals received by a plurality of cognitive users comprises:
the method comprises the following steps that N master users transmit signals from different directions to reach M cognitive user sides, wherein M is larger than N;
forming a received signal matrix Y (Y) by the signals received by M cognitive users1,Y2,...,YM]TWherein, superscript T represents the transpose of the matrix;
wherein the whitening preprocessing of the received signal matrix comprises:
carrying out eigenvalue decomposition on a covariance matrix of a received signal matrix Y;
constructing a whitening matrix W according to the eigenvalue decomposition result;
the whitening matrix W is point-multiplied with the received signal matrix Y to obtain a whitening signal Z ═ W.Y, wherein Z ═ Z1,Z2,...,ZN]T；
Wherein the constructing the whitening matrix W according to the eigenvalue decomposition result comprises:
the obtained characteristic values are subjected to descending order arrangement to obtain characteristic values [ lambda ] after descending order arrangement1,...,λM]And its corresponding feature vector [ h ]1,...,hM]；
Taking the first 1-N eigenvalues and eigenvectors thereof from the M eigenvalues after descending order arrangement to construct a whitening matrix W, wherein the whitening matrix W is expressed as:
Figure FDA0002310958770000021
wherein σ2The superscript H represents the conjugate transpose of the matrix for the average of the next M-N smaller eigenvalues after descending order.
2. The space-time-frequency three-dimensional spectrum sensing method according to claim 1, wherein the four-dimensional matrix obtained by space-time frequency distribution transform is represented as:
Figure FDA0002310958770000022
wherein D isZZ(t, f) is a two-dimensional covariance matrix representing spatial information; each element
Figure FDA0002310958770000023
Is a two-dimensional time-frequency matrix,
Figure FDA0002310958770000024
representing the ith signal ZiAnd the jth signal ZjMutual time-frequency distribution of(ii) a t represents time; f represents frequency.
3. The space-time-frequency three-dimensional spectrum sensing method according to claim 2, wherein the filtering the time-frequency two-dimensional information of the four-dimensional matrix and the extracting the single-source autocorrelation points comprises:
by means of a relational expression
Figure FDA0002310958770000025
Extracting autocorrelation points containing all primary user signals, wherein epsilon2To extract the threshold, trace (·) represents the trace of the matrix;
dividing the extracted autocorrelation points of all the main user signals into N classes through clustering, wherein one class represents the autocorrelation point of one signal to obtain K of N classes of single-source signalsiAn autocorrelation point
Figure FDA0002310958770000026
Wherein, i is 1,2i＝1,2,...,Ki。
4. The space-time-frequency three-dimensional spectrum sensing method according to claim 3, wherein the jointly diagonalizing the obtained single-source autocorrelation points to obtain a steering vector estimation for angle estimation comprises:
for the obtained N-type autocorrelation points
Figure FDA0002310958770000027
Extracting two-dimensional space information of the four-dimensional matrix to form a correlation matrix of the signal on corresponding time frequency points
Figure FDA0002310958770000028
For the obtained correlation matrix
Figure FDA0002310958770000029
Performing joint diagonalization to obtain
Figure FDA0002310958770000031
A minimum NxN dimensional unitary matrix U, where I is an identity matrix and off (-) is the sum of the squares of the non-diagonal elements of the matrix;
and (3) carrying out guide vector estimation by using a unitary matrix U obtained by joint diagonalization:
Figure FDA0002310958770000032
wherein, W#In order to whiten the pseudo-inverse of the matrix,
Figure FDA0002310958770000033
is a steering vector.
5. The space-time-frequency three-dimensional spectrum sensing method according to claim 4, wherein the determining the angle of arrival of each primary user signal according to the obtained steering vector comprises:
and determining the arrival angle of each main user signal by utilizing multiple signal classification according to the obtained guide vector.
6. The space-time-frequency three-dimensional spectrum sensing method according to claim 5, wherein the determining the angle of arrival of each primary user signal by using multiple signal classifications according to the obtained steering vector comprises:
to pair
Figure FDA0002310958770000034
Solving a covariance matrix:
Figure FDA0002310958770000035
wherein R isAAA covariance matrix of dimension M × M;
r is to beAAThe feature vector of (2) is divided into two parts according to the size of the feature value: u shapesAnd UnWherein, Us·Us H+Un·Un H＝I，Us＝[U1,U2,...,UN]Stretching a unitary space formed by eigenvectors corresponding to the first N maximum eigenvalues into a signal subspace; u shapen＝[UN+1,...,UM]Expanding a unitary matrix formed by eigenvectors corresponding to the last M-N minimum eigenvalues into a noise subspace;
by passing
Figure FDA0002310958770000036
And estimating the angle of the main user signal to obtain the arrival angle of each main user signal.
CN201811124363.0A 2018-09-26 2018-09-26 Space-time-frequency three-dimensional spectrum sensing method Active CN109274444B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811124363.0A CN109274444B (en) 2018-09-26 2018-09-26 Space-time-frequency three-dimensional spectrum sensing method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811124363.0A CN109274444B (en) 2018-09-26 2018-09-26 Space-time-frequency three-dimensional spectrum sensing method

Publications (2)

Publication Number Publication Date
CN109274444A CN109274444A (en) 2019-01-25
CN109274444B true CN109274444B (en) 2020-03-10

Family

ID=65198232

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811124363.0A Active CN109274444B (en) 2018-09-26 2018-09-26 Space-time-frequency three-dimensional spectrum sensing method

Country Status (1)

Country Link
CN (1) CN109274444B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109697758A (en) * 2019-02-18 2019-04-30 中电科仪器仪表有限公司 A kind of multi-view angle three-dimensional curved surface spectrogram display methods
CN111262617B (en) * 2020-01-16 2021-09-21 南京邮电大学 Spectrum sensing method based on cooperation of multiple low-orbit satellites

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106291453A (en) * 2016-09-18 2017-01-04 电子科技大学 A kind of two-dimentional ripple based on beam space conversion reaches angle assay method

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101592721B (en) * 2009-06-19 2011-09-14 中国人民解放军理工大学 Eigenvalue reconstruction based method for estimating angle of arrival of coherent signal

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106291453A (en) * 2016-09-18 2017-01-04 电子科技大学 A kind of two-dimentional ripple based on beam space conversion reaches angle assay method

Also Published As

Publication number Publication date
CN109274444A (en) 2019-01-25

Similar Documents

Publication Publication Date Title
CN109932680B (en) Non-circular signal direction of arrival estimation method based on translational co-prime array
CN101420758B (en) Method for resisting simulated main customer attack in cognitive radio
CN110197209B (en) Radiation source identification method based on multi-feature fusion
CN112910578B (en) Path parameter extraction method for millimeter wave 3D MIMO channel
Zhang et al. A spectrum sensing method based on signal feature and clustering algorithm in cognitive wireless multimedia sensor networks
CN102694588B (en) Arrival direction estimation method based on conjugation expansion
CN106019215A (en) Nested array direction-of-arrival angle estimation method based on fourth-order cumulants
CN109274444B (en) Space-time-frequency three-dimensional spectrum sensing method
CN106226729B (en) Relatively prime array direction of arrival angle method of estimation based on fourth-order cumulant
Zhuang et al. A multi-antenna spectrum sensing scheme based on main information extraction and genetic algorithm clustering
Shao et al. Machine Learning-Assisted Sensing Techniques for Integrated Communications and Sensing in WLANs: Current Status and Future Directions.
CN103714340A (en) Self-adaptation feature extracting method based on image partitioning
CN109962745B (en) Spectrum sensing method, system and device
Cheng et al. Further study of the pencil-MUSIC algorithm
Cheng et al. Performance of a novel automatic identification algorithm for the clustering of radio channel parameters
Pu et al. Novel three-hierarchy multiple-tag-recognition technique for next generation RFID systems
CN110365434B (en) Multi-antenna cooperative spectrum sensing method based on information geometry and differential evolution clustering algorithm
Dang et al. Air gesture recognition using WLAN physical layer information
Feng et al. Bi-iterative algorithm for extracting independent components from array signals
CN104363078B (en) The real orthogonal space time packet blind-identification method of under determined system based on robust Competition Clustering
Ding et al. Multimodal Fusion-GMM based Gesture Recognition for Smart Home by WiFi Sensing
CN108594165A (en) A kind of narrow band signal Wave arrival direction estimating method based on expectation-maximization algorithm
Shi et al. CSI-based fingerprinting for indoor localization with multi-scale convolutional neural network
CN109039490B (en) Frequency-space two-dimensional spectrum hole detection method for MIMO-OFDM system
Weigang et al. Study of improved multiple signal classification algorithm based on coherent signal sources

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