KR102127751B1 - Multi-target control system for separating interference signals - Google Patents

Abstract

The present invention relates to a multi-target control system for separating interference signals, and more specifically, to a multi-target control system for separating interference signals, capable of separating and extracting a plurality of interference signals received to an array antenna. A multi-target control system for separating interference signals according to an embodiment of the present invention comprises: a signal receiving module for receiving a plurality of signals through an array antenna; a plurality of cyclic frequency setting modules for setting a cyclic frequency for a signal of interest desired to be extracted, and connected in parallel to the signal receiving module; and a plurality of signal of interest extraction modules for extracting a corresponding signal of interest from among interference signals input through the signal receiving module based on each cyclic frequency set in the cyclic frequency setting modules. Also, the present invention further comprises an initial beamforming vector estimating module for calculating a unique vector orthogonal to a signal subspace included in a covariance matrix of a signal, when a new signal is input through the signal receiving module, and allocating the value as an initial beamforming vector value of a corresponding extraction module among the signal of interest extraction modules.

Description

Multi-target control system for separating interference signals

The present invention relates to a multi-target control system for separating the interference signals, and more particularly, to a multi-target control system for separating the interference signals that can be extracted by separating a plurality of interference signals received by the array antenna.

The contents described in this section merely provide background information for this embodiment, and do not constitute a prior art.

As the use of frequency resources has recently increased, and overcrowding of a specific frequency band has occurred, there is a need to solve the problem of detecting and extracting interference signals received in the same frequency band. In particular, when performing various military operations using a communication signal as in the military operation environment, and when it is used to ensure the safety of allies, extraction of the interference signal is a very important part.

Korean Registered Patent No. 10-0351125 (Invention name: smart antenna system having a pre-beam former for each multipath and an adaptive equalization combiner structure for each frequency component of a multipath signal) is provided for each multiple received through multiple antenna arrays. Pre-beam formation is performed on the path signals and the delay time is compensated to efficiently extract a desired signal through an adaptive equalizer in the frequency domain.

The patent has the characteristics of the pilot signal during the training or learning period, and the synchronization unit for performing frame synchronization, timing synchronization, and frequency synchronization of signals using signals received from multiple antennas, respectively. Is used to estimate the delay time for multipath. In addition, the estimation unit estimates a pre-beam forming weight vector through the estimated delay time information, and applies the pre-beam forming weight vector output from the estimator to a signal output from the synchronization unit to combine signal components for each path. In order to compensate for the delayed time of each multipath signal, the pre-beam forming unit that forms an independent beam for each multi-path signal and the multi-path signal output through the pre-beam forming unit by the estimated delay time. It comprises a matching compensation unit and an adaptive equalization unit that equalizes the signal output from the delay time compensation unit in the frequency domain.

The patent requires prior information on the incident signal to form a beam for removing the interference signal. However, since the pre-information on the incident signal has a problem that changes with time, a blind-based interference signal extraction technology capable of adaptively responding to the characteristics of the interest signal and the interference signal in a time-varying environment to separate and extract multiple interest signals. need.

Accordingly, it is an object of the present invention to provide a multi-target control system for separating interference signals capable of separating and extracting a plurality of signals existing in a specific frequency band.

Another object of the present invention is to provide a multi-target control system for separating the whole signal capable of improving convergence speed and ensuring convergence of a beam forming vector when multiple signals existing in a narrow band frequency band are separated and extracted. .

Another object of the present invention is to provide a multi-target control system for interference signal separation capable of separating and extracting a plurality of signals existing in a specific frequency band through blind-based adaptive beamforming control.

In order to solve the above technical problem, a multi-target control system for separating a mixed signal according to an embodiment of the present invention includes a signal receiving module for receiving a plurality of signals through an array antenna; A plurality of cyclic frequency setting modules connected in parallel to the signal receiving module to set a circulating frequency for a signal of interest to be extracted; And it characterized in that it comprises a plurality of interest signal extraction module for extracting the interest signal from the interference signal input through the signal receiving module based on each of the cyclic frequency set in the plurality of cyclic frequency setting module.

Preferably, when a new signal is input through the signal receiving module, an eigenvector orthogonal to the signal subspace included in the signal covariance matrix is calculated, and this value is extracted from the plurality of interest signal extraction modules. It characterized in that it further comprises an initial beamforming vector estimation module for assigning the initial beamforming vector value of.

Preferably, the initial beamforming vector estimation module is characterized in that the eigenvector value calculated by Equation 16 is calculated as the initial beamforming vector value.

Equation 16

은 샘플마다 갱신되는 공분산 행렬이다.

는 임의로 설정한 주기 P로 업데이트되는 공분산 행렬이며, n은 신호의 샘플 단위이고, j는 n의 블록 단위로서,

의 관계를 가진다.][From here

Is a covariance matrix that is updated for each sample.

Is a covariance matrix updated with a randomly set period P, n is the sample unit of the signal, and j is the block unit of n,

Has a relationship of.]

Preferably, the plurality of interest signal extracting modules generate a reference signal based on the cyclic frequency, and update the beamforming vector and the control vector so that the correlation between the reference signal and the beamforming output is maximized. .

Preferably, the initial beamforming vector output from the initial beamforming vector estimation module is updated in units of blocks, and the signal of interest extraction module updates the covariance matrix in units of samples, and is initially output from the initial beamforming vector estimation module. When the beamforming vector is applied to the signal extraction module of interest, the orthogonality of the covariance matrix updated in units of samples is determined, and a vector value when an eigenvalue greater than a set threshold is assigned is assigned as an initial beamforming vector. Should be

Preferably, the plurality of interest signal extraction modules are configured in parallel by using a cross-SCORE algorithm utilizing a cyclic frequency of each interest signal.

Preferably, the initial beamforming vector estimation module is characterized by using a DMP algorithm.

The multi-target control system for separating the whole signal according to an embodiment of the present invention is capable of separating and extracting a plurality of signals existing in a specific frequency band. At this time, by removing the interference signal present in the channel, only the desired signal of interest can be extracted and obtained, thereby improving the reception performance of each signal of interest.

In addition, the present invention can improve the convergence rate and the convergence speed of the beam forming vector when separating and extracting a plurality of signals existing in a narrow band frequency band. Such a point has an effect of improving reception performance by minimizing the time delay required to obtain a signal of interest extracted separately.

In addition, according to the present invention, a plurality of signals existing in a specific frequency band may be separated and extracted through blind-based adaptive beamforming control. That is, the present invention can obtain an effect capable of extracting a signal of interest by adaptively responding to a characteristic of the signal of interest even without prior information on the received signal.

1 is a diagram illustrating the configuration of a multi-target control system for segregation of interference signals according to an embodiment of the present invention.
2 is a configuration diagram of a multi-target control system for segregating the interference signal according to another embodiment of the present invention.
3 is an exemplary view of an experimental environment in which an embodiment of the present invention is tested.
4 is an exemplary diagram of a maximum eigenvalue detected when a new signal appears in an initial beamforming vector estimation module according to an embodiment of the present invention.
5 is an exemplary diagram comparing an initial beamforming vector pattern estimated by an initial beamforming vector estimation module with an existing vector pattern according to an embodiment of the present invention.
6 is an exemplary view of a beamforming vector pattern formed in a signal extraction module of interest according to an embodiment of the present invention.
7 is a result of the performance comparison according to repeated implementation according to an embodiment of the present invention.

Hereinafter, exemplary embodiments disclosed herein will be described in detail with reference to the accompanying drawings, but the same or similar elements are assigned the same reference numbers regardless of the reference numerals, and overlapping descriptions thereof will be omitted. The suffixes "parts" and "groups", "modules" and "parts", "units" and "parts", "devices" and "systems" for the components used in the following description are considered for ease of specification only. It is given or used interchangeably, and does not have a meaning or a role that is distinguished from itself.

In addition, in describing the embodiments disclosed in this specification, detailed descriptions of related well-known technologies are omitted when it is determined that they may obscure the gist of the embodiments disclosed herein. In addition, the accompanying drawings are only for easy understanding of the embodiments disclosed herein, and the technical spirit disclosed in the specification is not limited by the accompanying drawings, and all modifications included in the spirit and technical scope of the present invention , It should be understood to include equivalents or substitutes.

Terms including ordinal numbers such as first and second may be used to describe various components, but the components are not limited by the terms. The terms are used only for the purpose of distinguishing one component from other components.

When an element is said to be "connected" or "connected" to another component, it is understood that other components may be directly connected or connected to the other component, but other components may exist in the middle. It should be. On the other hand, when a component is said to be "directly connected" or "directly connected" to another component, it should be understood that no other component exists in the middle.

Singular expressions include plural expressions unless the context clearly indicates otherwise.

In this application, the terms "comprises" or "have" are intended to indicate the presence of features, numbers, steps, operations, components, parts or combinations thereof described in the specification, one or more other features. It should be understood that the existence or addition possibilities of fields or numbers, steps, operations, components, parts or combinations thereof are not excluded in advance.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. It will be apparent to those skilled in the art that the present invention may be embodied in other specific forms without departing from the spirit and essential features of the present invention.

1 is a diagram illustrating the configuration of a multi-target control system for segregation of interference signals according to an embodiment of the present invention.

1, a multi-target control system for segregating a signal in accordance with an embodiment of the present invention is a configuration diagram capable of separating all received signals by arranging a plurality of signal extraction modules of interest in parallel. In the present invention, different cyclic frequencies are applied to each of the signal extraction modules of interest, and when the interference signal is received, a signal having a set cyclic frequency is obtained as the output of the signal extraction module of interest.

As illustrated, the signal receiving module 110 receiving a plurality of signals, such as an array antenna, is connected in parallel to a plurality of signal extraction modules 130 to 130n through a plurality of cyclic frequency setting modules 120 to 120n. Is becoming. And a plurality of interest signal extraction modules (103 ~ 130n) for extracting a signal of interest of different characteristics from a mixed signal containing a plurality of signals is connected to the signal processing unit 140 for classifying and using the extracted signal have.

The signal receiving module 110 is configured to receive a signal through an array antenna (not shown) composed of M sensors. In particular, in the case of a narrowband signal, an environment composed of a signal of interest, an interference signal, and background noise may be considered, and the signal vector x(n) received by the array antenna composed of the M sensors may be defined as in Equation 1. .

Equation 1

Here, s _SOI (n) is an interest signal incident at an incident angle θ _SOI , and s _int (n) is an interference signal incident at an incident angle θ _int . a(θ) represents the steering vector at θ, and v(n) represents the noise signal vector.

Assuming that the signal of interest s _SOI (n) is spectrally self-consistent with respect to the cyclic frequency α and the interference signal s _int (n) has no self-consistency, there is no correlation between the two signals. Using this feature, the beamforming vector W can be obtained. Therefore, the beamforming output for obtaining the beamforming vector W for obtaining the signal of interest s _SOI (n) can be defined as in Equation 2.

Equation 2

Here, H is conjugate transpose, and when an appropriate beamforming vector W is obtained, the signal of interest s _SOI (n) can be extracted.

On the other hand, since the digital modulated signal has a carrier frequency and a bandwidth, the average and autocorrelation have cyclostationary characteristics. The average and autocorrelation of the signal x(n) having periodic characteristics has a period of n _T as shown in Equations 3 and 4.

Equation 3

Equation 4

The autocorrelation of x(n) is considered as the main statistical characteristic for extracting the signal of interest, and the autocorrelation with periodicity is expressed by the cyclic autocorrelation function and the conjugate cyclic autocorrelation function. It can be represented by 5 and equation (6).

Equation 5

Equation 6

Here, [X ^H ] ^(*) = X ^T , and <·>∞ means the time average over infinite time. The cyclic autocorrelation function has a large value in time transition n ₀ and frequency shift α. T is transpose, and f _s represents the sampling frequency. At this time, it can be said that a signal whose autocorrelation shows periodicity in the time axis and the frequency axis has a cyclostationary, and in this case, the frequency shift α is called a cyclic frequency.

It is the plurality of interest signal extraction modules 130 to 130n that remove the interference signal and acquire only the interest signal using the cyclic frequency feature. Therefore, the cyclic frequency setting module (120 ~ 120n) sets each different cyclic frequency having a periodic characteristic. The cyclic frequency setting values of the cyclic frequency setting modules 120 to 120n are preset values according to an interest signal to be obtained, and are calculated by a symbol rate or a carrier frequency of the interest signal.

The plurality of interest signal extraction modules 130 to 130n are blind beam forming algorithm control modules (self-coherence restoral) using cyclostationary, and among them, a representative cross-SCORE algorithm uses the cyclic frequency of the interest signal. By obtaining the reference signal, the beamforming vector is updated in a direction to maximize the correlation between the beamforming output and the reference signal. Accordingly, the plurality of interest signal extraction modules 130 to 130n update the beamforming vector in a direction to minimize the difference between the beamforming output and the reference signal, and extract the interest signal.

The blind-based adaptive beamforming algorithm can extract an interest signal by forming an antenna beam by adaptively responding to characteristics of an interest signal and an interference signal. The blind-based adaptive beamforming algorithm utilizes characteristics such as constant signal amplitude, periodic cyclostationary characteristics, and high order cumulant, instead of information such as the number of interference signals and the direction of incidence. To extract the signal of interest.

Therefore, the plurality of interest signal extraction modules of the present invention extract the interest signal n through the correlation between the interest signal and the interference signal using periodic characteristics. If the interest signal and the interference signal have different periodic characteristics with respect to the cyclic frequency α, the plurality of interest signal extraction modules 130 to 130n do not have a correlation. Accordingly, a correlation vector for a cyclic frequency is obtained through a cyclic autocorrelation function, and a signal of interest is extracted, and a beamforming vector is formed to remove the interference signal.

The plurality of interest signal extraction modules 130 to 130n obtain a reference signal r(n) using the cyclic frequency of the signal of interest to beam in a direction to maximize the correlation between beamforming output y(n) and r(n) Update the formation vector. The cyclic signal u(n) is a signal in which a time delay and frequency of the input signal are shifted, and the reference signal can be generated using a self-consistency characteristic of the interest signal.

The cyclic signal u(n) and the reference signal r(n) are defined as in Equation 7 and Equation 8 below.

Equation 7

Equation 8

Here, c denotes a control vector and is used as a beamforming vector of the cyclic signal u(n).

The plurality of interest signal extraction modules 130 to 130n extract blind interest signals having self-consistency with respect to the cyclic frequency α and remove interference-based interference signals, thereby removing blind-based interference signals. At this time, the value of the control vector affects performance. The plurality of interest signal extraction modules 130 to 130n update the beamforming vector and the control vector together.

The calculation formulas of the beam forming vector and the control vector of the plurality of interest signal extraction modules 130 to 130n are defined by Equation 9, Equation 10, and Equation 11.

Equation 9

Equation 10

Equation 11

과

은 각각 수신신호와 순환신호의 자기상관 행렬을 의미하며,

과

는 수신신호와 순환신호의 상호상관 (cross-correlation) 행렬을 나타낸다. 수학식 10은 W_CS(n)와 c(n)의 크기를 제한하기 위해 필수적인 정규화 (normalization) 과정이며, 각 상관 행렬은 ,

인 관계를 가지고 있다.here,

and

Denotes an autocorrelation matrix of a received signal and a cyclic signal, respectively,

and

Denotes a cross-correlation matrix of a received signal and a cyclic signal. Equation 10 is a normalization process necessary to limit the size of W _CS (n) and c(n), and each correlation matrix is

Have a relationship.

과

를 구하는 식은 수학식 12, 수학식 13와 같고, 역행렬인 R^-1 _xx(n)은 matrix inverse lemma를 사용하여 계산되며 수학식 14와 같이 표현된다. 수학식 12와 수학식 13에 이용된 N은 이동 사각 윈도우의 크기를 나타낸다.remind

and

The equation for obtaining is equal to Equation 12 and Equation 13, and the inverse matrix R ^-1 _xx (n) is calculated using a matrix inverse lemma and is expressed as Equation 14. N used in Equations (12) and (13) represents the size of the moving square window.

Equation 12

Equation 13

Equation 14

The process of extracting and acquiring a plurality of signals existing in a specific frequency band using a plurality of interest signal extraction modules according to an embodiment of the present invention configured as described above is performed as follows.

When a mixed signal including a plurality of signals existing in a specific frequency band is input through the signal receiving module 110, a plurality of signals according to the cyclic frequency information set in the cyclic frequency setting modules 120 to 120n applied to each port Interest signal extraction module (130 ~ 130n) is operated.

Each interest signal extraction module (130~130n) extracts one interest signal based on the set and assigned cyclic frequency, and simultaneously generates a beamforming vector in a direction to remove the remaining interference signals, and simultaneously generates a desired interest signal or It is extracted sequentially.

Therefore, when a plurality of interest signal extraction modules 130 to 130n performed by various parallel ports are passed, a plurality of signals included in the incident interfering signal are separated and extracted and input to the signal processing unit 140. The separated and extracted signals input to the signal processing unit 140 then perform necessary signal processing to obtain desired information.

Next, FIG. 2 is a configuration diagram of a multi-target control system for segregation of interference signals according to another embodiment of the present invention.

In a multi-target control system for separating and receiving all of the interference signals, the output of each signal of interest extraction module must converge to different signals. In addition, a plurality of interest signal extraction modules must operate stably even when a new signal appears. In particular, a convergence delay may occur in a situation in which a new signal appears as the signal changes, and to improve this, it is necessary to estimate an initial beamforming vector.

Therefore, as illustrated in FIG. 2, when a new signal appears, the initial beamforming vector is estimated through an initial beamforming vector estimation module (Dminant Mode Prediction: DMP algorithm). The initial beamforming vector obtained from the initial beamforming vector estimation module has a high beamforming gain at an incident angle where a new signal appears, and a very low gain at an incident angle of an existing interference signal.

In the illustrated multi-target control system for separation of the mixed signal, the signal receiving module 210 receiving a plurality of signals, such as an array antenna, extracts a plurality of interest signals through a plurality of cyclic frequency setting modules 220 to 220n. (230~230n). In addition, a plurality of interest signal extraction modules 230 to 230n for extracting one interest signal of different characteristics from a mixed signal including a plurality of signals are connected to a signal processing unit 240 for classifying and using the extracted signals. have.

In addition, the multi-target control system of the present invention further includes an initial beamforming vector estimation module 250 for estimating an initial beamforming vector when a new signal appears. The plurality of interest signal extraction modules 230 to 230n are guaranteed to converge to different signals based on the initial beamforming vector allocated by the initial beamforming vector estimation module 250, and the convergence speed can also be improved. .

The initial beamforming vector estimation module 250, when a new signal appears, forms a deep null in the incident angle of the existing signal by using orthogonal characteristics of the sub-space of the signal, and is high as the incident angle of the new signal. The beamforming vector is estimated to have a gain, and is allocated to each of the plurality of interest signal extraction modules 230 to 230n.

At this time, the initial beamforming vector estimation module allocates the beamforming vector estimated using the characteristic that the sub-spaces of different signals are orthogonal to the signal extraction module of interest as the initial beamforming vector. At this time, the initial beamforming vector estimation module updates the covariance matrix in block units, but the interest signal extraction module updates in sample units of signals. Therefore, while the interest signal extraction module updates the autocorrelation matrix for each sample, the initial beamforming vector estimation module should be controlled to update the covariance matrix in block units. Therefore, an orthogonal signal is estimated through an eigenvalue decomposition process between a block-based reference autocorrelation matrix and a sample-updated matrix. Then, the eigenvector when the eigenvalue is greater than the set threshold is assigned as the initial beamforming vector.

In the illustrated figure, the initial beamforming vector estimation module 250 is not connected to the output terminal of the signal receiving module 110, but the initial beamforming vector estimation module 250 includes a new signal, particularly a specific cyclic frequency. You should be able to confirm that the signal of interest. Therefore, the initial beamforming vector estimation module 250 is configured to monitor the input of interest signals based on the circulating frequency to the plurality of interest signal extraction modules 230 to 230n.

The initial beamforming vector estimated by the initial beamforming vector estimation module 150 may be expressed as Equation 16 as an eigenvector corresponding to the maximum eigenvalue of the eigenequation using the covariance matrix shown in Equation (15). .

Equation 15

Equation 16

은 샘플마다 갱신되는 공분산 행렬이고, n은 신호의 샘플 단위이고, 샘플이 증가함에 따라 누적된 샘플을 통해 구해지는 공분산 행렬의 값으로

이 갱신된다.

는 실험에 의해 결정된 기설정한 주기 P로 업데이트되는 공분산 행렬이며, j는 n의 블록 단위로서,

의 관계를 가진다. 이 때,

은 ㆍ보다 크지 않은 정수를 의미한다. 매 샘플마다

와

의 고유값 분해 (eigen-decomposition) 과정을 통해서 고유값과 고유벡터를 구한다. 이때의 고유벡터를 W_init 인 초기 빔형성벡터로 할당한다.From here

Is a covariance matrix that is updated for each sample, n is the sample unit of the signal, and is the value of the covariance matrix obtained through the accumulated samples as the sample increases.

Is updated.

Is a covariance matrix updated with a predetermined period P determined by experiment, j is a block unit of n,

Have a relationship of At this time,

Means an integer not greater than ㆍ. Every sample

Wow

The eigenvalues and eigenvectors are obtained through the eigen-decomposition process. The eigenvector at this time is assigned as the initial beamforming vector, which is W _init .

에서 주기 P의 값이 증가할수록 업데이트되는 빈도가 낮아지므로 새로운 신호가 등장했을 때, 큰 고유값을 얻는데 지연시간이 발생한다. 또한

을 짧은 주기로 업데이트할 경우 알고리즘의 계산 복잡도가 증가하는 단점이 발생하므로 적절한 주기 P의 값을 설정할 필요가 있다. 상기 주기 P는 후술되는 다양한 실험 결과를 통해서 최적의 결과를 얻을 수 있도록 설정되어진다. 즉, 본 발명에서는,

의 관계를 갖도록 설정하고, 실험치로 얻어진 주기를 적용하여 최적의 초기 빔 형성 벡터를 얻고 있다.

Since the frequency of the update increases as the value of the period P increases, a delay occurs in obtaining a large eigenvalue when a new signal appears. Also

If is updated in a short period, it is necessary to set the value of the proper period P because the computational complexity of the algorithm increases. The period P is set to obtain an optimal result through various experimental results described below. That is, in the present invention,

It is set to have a relationship of, and the optimal initial beam forming vector is obtained by applying the period obtained by the experimental value.

의 신호 부공간에 포함되지 않으면서

의 신호 부공간에는 포함되는 고유벡터가 되야 한다. 이와 같이 해서 얻어진 고유벡터는 기존 신호 부공간에 대해서 상관성이 낮기 때문에 기존의 신호 방향에 깊은 널(NULL)을 형성하게 되고, 새로운 신호 방향으로 높은 빔형성 이득을 가지게 된다.When a new signal appears, the eigenvector corresponding to the maximum eigenvalue in Equation 16 is the covariance matrix of the existing signal.

Is not included in the signal subspace of

It must be an eigenvector included in the signal subspace of. Since the eigenvector obtained in this way has a low correlation to the existing signal subspace, deep nulls are formed in the existing signal direction, and a high beamforming gain is obtained in the new signal direction.

The multi-target control system for separation of the interference signal according to the present invention made of the above configuration is operated as follows.

When a mixed signal including a plurality of signals existing in a specific frequency band is input through the signal receiving module 210, a plurality of signals according to the cyclic frequency information set in the cyclic frequency setting modules 220 to 220n applied to each port Interest signal extraction module (230 ~ 230n) is operated.

Each of the interest signal extraction modules 230 to 230n extracts one interest signal based on a set and assigned cyclic frequency, and extracts a desired interest signal while generating a beamforming vector in a direction to remove the remaining interference signals. do.

On the other hand, when multiple signals newly appear in a specific frequency band according to the passage of time, the beamforming vectors must converge to extract the signals of interest from each signal extraction module of interest 230 to 230n. At this time, an appropriate initial beamforming vector must be allocated for fast convergence of the signal.

The initial beamforming vector estimation module 250 estimates an initial beamforming vector at a time when a new signal appears and assigns it to the plurality of interest signal extraction modules 230-230n. The initial beamforming vector estimation module 250 has a large eigenvalue when a new signal appears.

At this time, when the detected eigenvalue is greater than a threshold set as a reference value, the eigenvector corresponding to the eigenvalue is determined as an initial beamforming vector and is allocated to the interest signal extraction modules 230 to 230n. The initial beamforming vector estimation module detects the appearance of a new signal only when an eigenvalue greater than an existing eigenvalue (a value close to 1) occurs in order to detect the appearance of a new signal. Therefore, if the power of the newly emerged signal is less than the noise power, a value greater than 1 does not occur, so the signal to noise ratio (SNR) of the newly emerged signal should be a sufficiently large value.

The interest signal extraction modules 230 to 230n set the initial beamforming vector allocated by the initial beamforming vector estimation module 250 as a beamforming vector and converge to a new signal. In this way, a plurality of interest signals extracted from the interest signal extraction modules 230 to 230n are output to the signal processing unit 240.

That is, the interest signal extraction modules 230 to 230n extract each interest signal before a new signal appears. In addition, an arbitrary signal extraction module to which the cyclic frequency information for a new signal is applied generates a beamforming vector having a low gain in the direction of the existing signal. In addition, when a new signal is incident, an initial beamforming vector is assigned to the corresponding signal extraction module, and extraction of the signal of interest is performed.

Accordingly, when a plurality of interest signal extraction modules 230 to 230n performed by various parallel ports are passed, a plurality of signals included in the incident intermix signal are separated and extracted and input to the signal processing unit 240. The separated and extracted signals input to the signal processing unit 240 then perform necessary signal processing to obtain desired information.

As described above, when the signal of interest is extracted by applying the initial beamforming vector estimation module, it is possible to obtain an effect that the convergence of the signal of interest is better guaranteed and the convergence speed is faster.

Next, the initial values were set based on the algorithm implementation process, and an experimental environment was constructed as shown in FIG. 3 to confirm the effect.

As shown in Fig. 4, when the second signal newly appears at 2 ms, it can be confirmed that the eigenvalue suddenly increases. The initial beamforming vector estimation module 250 assigns the eigenvector corresponding to the maximum eigenvalue to the second interest signal extraction module 231 as the initial beamforming vector.

At this time, the pattern of the existing beamforming vector and the beamforming vector pattern according to the initial beamforming vector are compared and illustrated in FIG. 5.

As shown in the figure, when comparing the gain difference of the beamforming vector at the incident angle of the signal of interest and the interference signal, the performance of removing the interference of the pattern of the initial beamforming vector estimated by the present invention than the conventional beamforming vector pattern has You can see that it is better.

That is, the existing beamforming vector has a gain difference of about 14dB from the incident angle of the interest signal and the interference signal. The beamforming vector of the present invention shows a gain difference of about 28 dB in the incident angle between the interest signal and the interference signal.

Also, FIG. 6 shows the beamforming vector gain for the second signal of interest extraction module 231 with respect to the incident direction of the signal of interest and interference. The second interest signal extraction module 231 for extracting the interest signal can confirm that the beamforming vector gain for the interference signal direction is low. However, from the viewpoint of the beamforming vector gain in the direction of the signal of interest, it can be confirmed that the signal of interest has a fast convergence rate after appearing at 2 ms.

This experiment was repeatedly performed several times, and as a result, the performance is shown in FIG. 7. The gain difference of the initial beamforming vector to extract a new signal is

It was found to have a value that is about 2 times larger than the previous one, and the cumulative number of samples for convergence was confirmed to be reduced to about 13% compared to the previous one.

On the other hand, the present invention is partially affected by the noise effect and the incident angle difference. In order to judge this, the experiment was performed under the same conditions as in FIG. 3 by applying only a low SNR value (0 dB).

The initial beamforming vector estimation module 250 has a characteristic that a signal has a eigenvalue that is sufficiently large when the power of the signal is greater than the power of the noise. Therefore, when the signal-to-noise ratio has a low SNR value, the magnitude of the maximum eigenvalue appears very low, with a value around 1. When the initial beamforming vector estimated according to the eigenvalue obtained with a low threshold value is compared with the existing vector, the difference in beamforming vector gain in the direction of the signal of interest and the interference signal decreases as the value of SNR decreases. Similarly, when the signal-to-noise ratio has a low SNR value, the difference in the beamforming vector gains for the incident direction of the interest signal and the interference signal input to the second port is also small. However, it can be seen that the convergence speed is faster than the previous one. Therefore, although the performance of the embodiment of the present invention is deteriorated at a low noise ratio, the convergence guarantee and the convergence speed are superior to those of the related art.

In addition, the present invention is partially affected by the difference in incident angle. In order to judge this, under the same conditions as in FIG. 3, only the angles of incidence of signals 1 and 2 were set at 20 degrees and 26 degrees, and experiments were conducted.

If the difference between the angles of incidence is small in order to confirm a case in which the direction of interest and the interference signal are close, the second signal appears to have a small eigenvalue. In addition, depending on the threshold setting, the maximum eigenvalue can be obtained after the time when the signal appears is less than 2 ms.

When the initial beamforming vector estimated according to the obtained eigenvalue is compared with the existing beamforming vector, the difference in gain of the beamforming vector in the direction of the signal of interest and the interference signal decreases as the newly emerged signal is incident closer to the existing signal. Also, from the viewpoint of the beamforming vector gain in the direction of the signal of interest, it converges quickly. Therefore, as the difference in the angle of incidence decreases, the present invention reduces the difference in initial beamforming vector gain and causes a slight delay in convergence, but better convergence performance can be obtained than in the past.

The above detailed description should not be construed as limiting in all respects, but should be considered illustrative. The scope of the invention should be determined by rational interpretation of the appended claims, and all changes within the equivalent scope of the invention are included in the scope of the invention.

110,210: Signal receiving module
120~120n, 220~230n: Circulation frequency setting module
130~130n, 230~230n: Interest signal extraction module
140,240: signal processing unit
250: initial beamforming vector estimation module

Claims (7)

A signal receiving module for receiving a plurality of signals through an array antenna;
A plurality of cyclic frequency setting modules connected in parallel to the signal receiving module to set a circulating frequency for a signal of interest to be extracted;
A plurality of interest signal extraction modules for extracting a corresponding interest signal from the interference signals input through the signal reception module based on the respective circulating frequencies set in the plurality of cyclic frequency setting modules; And
When a new signal is input through the signal receiving module, an eigenvector orthogonal to the signal subspace included in the covariance matrix of the signal is calculated, and this value is the initial beam of the corresponding extraction module among the plurality of signal extraction modules of interest. And an initial beamforming vector estimation module for allocating as a forming vector value,
The initial beamforming vector estimation module calculates an eigenvector value calculated by Equation 16 as an initial beamforming vector value,

Is a covariance matrix updated per sample,

Is a covariance matrix updated with a predetermined period P determined by experiment, n is a sample unit of the signal, j is a block unit of n,

Target control system for separation of interfering signals having the relationship of
Equation 16

delete delete The method according to claim 1,
The plurality of interest signal extraction modules generate a reference signal based on the cyclic frequency, and multiple targets for separation of the interference signal updating the beamforming vector and the control vector so that the correlation between the reference signal and the beamforming output is maximized. Control system. The method according to claim 1,
The initial beamforming vector output from the initial beamforming vector estimation module is updated in block units, and the interest signal extraction module updates the covariance matrix in sample units,
When applying the initial beamforming vector output from the initial beamforming vector estimating module to the signal extraction module of interest, the orthogonality of the covariance matrix updated in units of samples is determined to have a eigenvalue greater than a set threshold. A multi-target control system for interference signal separation that assigns values to initial beamforming vectors. The method according to claim 1,
The plurality of interest signal extraction modules, a multi-target control system for the separation of the interference signal configured in parallel using a cross-SCORE algorithm utilizing the cyclic frequency of each interest signal.
The method according to claim 5,
The initial beamforming vector estimation module is a multi-target control system for segregation of interference signals using a DMP algorithm.

Images