JP5645211B2 - Method and apparatus for detecting direction of arrival of pulsed radio noise - Google Patents

Description

  The present invention relates to a method and an apparatus for searching the direction of arrival of pulsed radio noise, and is particularly useful when applied to identify the source of pulsed radio noise in a power facility.

  In electric power facilities, there is a concern about radio wave interference caused by pulsed radio noise radiated due to the occurrence of a spark discharge at a deterioration insulator or a poorly contacted part of a metal fitting, and a steep current flowing. Incidentally, when a spark discharge is generated by a small gap, radio noise including a frequency component of 100 MHz or more is observed, and a frequency component of 25 MHz to 2800 MHz is generated from a spark discharge generated using a deteriorated distribution insulator. Radio noise including is observed.

  Thus, the radio noise generated by the spark discharge has a wide band characteristic and may affect a wide range of wireless communications such as radio, television, amateur radio, mobile phone, and wireless LAN.

  In order to quickly eliminate the radio interference due to the radio noise, it is necessary to efficiently search for the source of the radio noise and then remove it.

  In the source noise source search according to the prior art, a method of searching for the source using a main beam of a directional microphone mainly for a twin quad antenna or a sound wave radiated during discharge has been adopted. It was.

  However, in this type of exploration technique according to the prior art, the estimation accuracy is not sufficient, and it is necessary to manually scan the main beam, so that there remains a problem in operability. On the other hand, recently, a method for estimating the arrival direction and source position of radio noise by performing signal processing on the received signals of multiple antennas has been proposed, and a method based mainly on the arrival time difference between antennas has been studied. (See, for example, Non-Patent Documents 1 to 3). In addition, when estimating the arrival direction, the arrival direction is estimated at a plurality of points, and the source position is determined by the association method (a method of estimating the arrival direction at a plurality of points and estimating the source position from the intersection of the arrival directions). Is estimated.

  However, in power facilities, 1) spark discharge may occur simultaneously at multiple locations, and 2) restart from a discontinuity point on the propagation path (impedance discontinuity point such as a bent part of the wire or a wire support). Since the estimation accuracy deteriorates due to the possibility of radiation, 3) the presence of a reflected wave from the metal surface of the power device, etc., it has not been established as a practical estimation method. Furthermore, the spark discharge generated in electric power facilities is unstable due to the influence of climate change such as wind and humidity, and radio noise cannot always be observed. For this reason, estimation with higher reliability with a small number of measurements is desired.

Electrics, Vol. 115-B, no. 10, pp. 1168-1173, 1995 Electrics, Vol. 118, no. 2, pp. 157-163, 1998 1999 IEEJ Power and Energy Competition, No. 600, 1999

  An object of the present invention is to provide a method and an apparatus for searching a direction of arrival of pulsed radio noise, which can accurately and easily estimate the direction of arrival of pulsed radio noise.

The first aspect of the present invention for achieving the above object is as follows:
While receiving pulsed radio noise with an array antenna formed by arranging N (N is a natural number of 2 or more) antennas in parallel, a first received signal representing the level of the pulsed radio noise with respect to the time axis is obtained. Generating a second received signal representing a spectrum intensity at each frequency component by Fourier-transforming the first received signal;
Thereafter, a first correlation matrix is formed for each frequency component of the second received signal, and the first correlation matrix is weighted with respect to the spectrum intensity of each frequency component, and the first matrix is further weighted by a focusing matrix . By converting the correlation matrix of 1 into a correlation matrix of a specific frequency and performing the averaging process, the existence range of the frequency spectrum of the second received signal is limited and information on all frequencies in the existence range is included. Form a correlation matrix of two,
Further, the arrival direction of the pulsed radio noise with respect to the (N-1) direction is estimated by applying a subspace method on the narrowband signal to the second correlation matrix. It is in the direction of arrival exploration method.

According to this aspect, since the first correlation matrix is weighted with respect to the spectrum intensity of each frequency component to convert it to a correlation matrix of a specific frequency and the averaging process is performed, all the narrowed down to the necessary frequency band is performed. A second correlation matrix including frequency information can be formed. Moreover, since the arrival direction of the radio wave is searched by applying the subspace method for the narrowband signal to the second correlation matrix in combination with the narrowing of the frequency band, the pulsed radio wave having a wide band is obtained. Even for noise, the direction of arrival can be estimated with high accuracy by the subspace method.
In the present embodiment, since the focusing matrix is used when generating the second correlation matrix, the second correlation matrix can be calculated easily and appropriately.

The second aspect of the present invention is:
An array antenna comprising N antennas (N is a natural number of 2 or more) arranged in parallel at intervals corresponding to a predetermined center frequency so as to receive pulsed radio noise;
A signal receiving unit for detecting a first received signal representing a level with respect to a time axis of the pulsed radio noise received via each antenna;
A Fourier transform unit that performs a Fourier transform on the first received signal to generate a second received signal representing a spectrum intensity at each frequency component;
A first correlation matrix computing unit that computes a first correlation matrix for each frequency component of the second received signal;
The first correlation matrix is weighted with respect to the spectrum intensity of each frequency component, and the first correlation matrix is converted into a correlation matrix of a specific frequency by a focusing matrix and an averaging process is performed. A second correlation matrix calculation unit that limits the existence range of the frequency spectrum of the received signal and calculates a second correlation matrix including information on all frequencies in the existence range;
An arrival direction estimation unit that estimates an arrival direction of the pulsed radio noise with respect to the (N-1) direction by applying a subspace method for a narrowband signal to the second correlation matrix; It is in the arrival direction exploration device of the pulsed radio noise.

According to this aspect, the second correlation matrix calculation unit weights the first correlation matrix with respect to the spectrum intensity of each frequency component and performs a predetermined averaging process. A second correlation matrix containing all frequency information can be formed. In addition, in the arrival direction estimation unit in combination with the narrowing of the frequency band, the arrival direction of the radio wave is explored by applying the subspace method for the narrowband signal to the second correlation matrix. Even for pulsed radio noise having a band, the direction of arrival can be estimated with high accuracy by the subspace method.
In the present embodiment, since the focusing matrix is used when generating the second correlation matrix, the second correlation matrix can be calculated easily and appropriately.

The third aspect of the present invention is:
In the arrival direction exploration device for pulsed radio noise described in the second aspect,
The specific frequency is the central frequency, and the arrival direction search apparatus for pulsed radio noise is characterized.

  According to this aspect, the conversion to the correlation matrix can be made suitable.

The fourth aspect of the present invention is:
In the arrival direction exploration device for pulsed radio noise described in the second or third aspect ,
The pulsed radio noise is radiated from a power facility.

  According to this aspect, it is easy and reliable even for pulsed radio noise generated from a power facility that may generate spark discharge or the like at a plurality of locations at the same time or may have a reflected wave from a metal surface. Each arrival direction can be specified.

  According to the present invention, since the weighting process and the broadband subspace method are combined, the direction of arrival can be determined without degrading the accuracy of direction-of-arrival exploration due to the frequency characteristics of pulsed radio noise that may be different for each occurrence. The arrival direction of pulsed radio noise can be probed with high accuracy, even when there are multiple.

It is a block diagram which shows the arrival direction search apparatus of the pulsed radio noise which concerns on embodiment of this invention. It is a wave form diagram which shows an example of the waveform of the pulsed radio noise which is the 1st received signal in the embodiment of the present invention. It is a wave form diagram which shows the 2nd received signal which carried out the Fourier transform of the 1st received signal in embodiment of this invention. It is a figure which shows notionally the aspect of the experiment for simulating the arrival direction of the pulse radio wave noise using the arrival direction search apparatus shown in FIG. 1, (a) is explanatory drawing shown in the state seen planarly, (B) is explanatory drawing shown in the state seen from the side surface. It is a characteristic view which shows an example of the simulation result of the arrival direction of pulsed radio noise. It is a characteristic view which shows the other example of the simulation result of the arrival direction of pulsed radio noise.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a block diagram showing an arrival direction searching apparatus for pulsed radio noise according to an embodiment of the present invention. As shown in the figure, the array antenna 1 includes four antennas 3A, 3B, 3C, and 3D arranged on a horizontal base 2. Here, the gantry 2 is horizontally disposed on the leg portion 4 and is made of wood in order to reduce reflection of radio waves. The antennas 3 </ b> A to 3 </ b> D are dipole antennas, which are fixed in a state in which the antennas 3 </ b> A and 3 </ b> D are arranged at equal intervals in the horizontal direction on the gantry 2 and are vertically set. The dimensions and intervals of the antennas 3A to 3D are determined according to the center frequency of the received radio wave. In this embodiment, since the center frequency is 500 MHz, the dimensions and intervals of the antennas 3A to 3D are set to 0.3 m (λ / 2) based on the wavelength λ of the radio wave of 500 MHz.

  A reception signal received by the array antenna 1 is supplied to the signal reception unit 6 through the coaxial cable 5. The signal receiving unit 6 detects a first received signal representing a level of the received radio wave with respect to the time axis, such as pulsed radio noise. The waveform of the first signal in this embodiment, which is an example of pulsed radio noise radiated from the power equipment, is shown in FIG. 2, and this first received signal is supplied to the signal processing unit 7.

  The signal processing unit 7 includes a Fourier transform unit 8, a first correlation matrix calculation unit 9, a second correlation matrix calculation unit 10, and an arrival direction estimation unit 11, and performs predetermined signal processing in the broadband subspace method. Later, the arrival direction of radio noise is probed by performing predetermined signal processing by a subband method of a narrow band.

  Here, the subspace method considers the vector space of the received signal, and the subspace to which the signal belongs (referred to as signal subspace) and the subspace that is complementary to the signal subspace (referred to as noise subspace). This is a method of estimating the direction of arrival by using it. In this embodiment, CSSM (Coherent Signal Subspace Method) that can ensure good estimation accuracy even when the S / N is low or when a plurality of signals arrive from close directions is a wideband partial space. Applicable as a law. This CSSM is a method of obtaining a correlation matrix including information on all frequencies by converting a correlation matrix for each frequency into a correlation matrix of a specific frequency and performing an averaging process on the converted correlation matrix. In addition, as a narrow-band subspace method, a MUSIC (Multiple Signal Classification) method, an ESPRIT (Estimation of Signal Parameter via Rotational Induction Technique) method, and an EM (Expectation-Mix method) are used. Although applicable, the MUSIC method is applied in this embodiment.

  The Fourier transform unit 8 of the signal processing unit 7 performs a Fourier transform on the first received signal to generate a second received signal representing the spectrum intensity at each frequency component. In the case of this embodiment, the second received signal is a broadband signal as shown in FIG. 3 having a center frequency in the vicinity of 500 MHz.

  The first correlation matrix calculator 9 calculates a first correlation matrix for each frequency component of the second received signal. That is, the following equation (1) is calculated.

R _xx (f _j ) = X (f _j ) X ^H (f _j ) (1)
(J = 1 to J)
In the above equation (1), R _xx (f _j ) is a correlation matrix, and H is a complex conjugate transpose.

  The second correlation matrix calculation unit 10 weights the spectrum intensity of each frequency component to the first correlation matrix and uses the focusing matrix to specify the first correlation matrix at a specific frequency (here, the center frequency of 500 MHz is used). The second correlation matrix is calculated by performing the averaging process after converting to the correlation matrix. That is, the calculation based on the following equation (2) is performed.

R _xx (f ₀ ) = Σw _j T _j R _xx (f _j ) X (f _j ) T _j ^H (2)
In the above equation (2), f ₀ is a specific frequency (500 MHz), and w _j is a weight for the correlation matrix.

  The second correlation matrix including information on all frequencies in the existence range can be obtained while limiting the existence range of the frequency spectrum of the second received signal by the calculation of the above equation (2).

Here, there is no special restriction on the weighting method. For example, a value obtained by dividing the spectrum intensity in each frequency component by the sum of the frequency spectrum intensity in all frequency bands can be used as a weighting coefficient, or a threshold value corresponding to a predetermined value of the spectrum intensity may be used. In the present embodiment, the focusing matrix is used when converting the first correlation matrix into the correlation matrix of a specific frequency. However, the present invention is not limited to this as long as it is a conversion matrix. Various methods have been proposed as a method for creating a focusing matrix. When the product of the focusing matrix (T _j T _j ^H ) does not depend on the frequency, the matrix is converted before and after the correlation matrix is converted by the focusing matrix. A decrease in S / N can be avoided later. As such a focusing matrix, a matrix called SST (Signal Subspace Transform) or MTLS (Modification to Total Least Squares) has been proposed. Also, as a special case of the SST matrix, there is a technique called RSS (Rotational Signal Subspace) that uses a unitary matrix as a focusing matrix. In this embodiment, a focusing matrix is created by RSS that is easy to apply. Further, although the center frequency is used as the specific frequency when the first correlation matrix is converted to the second correlation matrix, the present invention is not limited to this.

  The arrival direction estimation unit 11 estimates the arrival direction of pulsed radio noise by applying a MUSIC method, which is a subspace method for narrowband signals, to the second correlation matrix. Here, in this embodiment, since the four antennas 3A to 3D of the array antenna 1 are formed, it is possible to search for arrival directions in up to three directions. In general, when the number of antennas 3A to 3D is N (N is a natural number of 2 or more), the arrival direction with respect to the maximum (N-1) direction can be estimated.

  Here, the operation and effect of the present embodiment will be described together with a numerical simulation result obtained by exploring the arrival direction of the pulsed radio noise caused by the spark discharge simulating the pulsed radio noise radiated from the power equipment.

  FIG. 4 is a diagram conceptually showing a mode of numerical simulation for estimating the arrival direction of pulsed radio noise using the arrival direction searching apparatus according to this embodiment, and (a) is shown in a plan view. Explanatory drawing, (b) is explanatory drawing shown in the state seen from the side surface. In this embodiment, as shown in FIG. 4A, a one-dimensional direction-of-arrival estimation is performed using an equally spaced linear array composed of four antennas 3A to 3D (assuming an omnidirectional antenna). The distance d between 3A and 3D is set to 0.3 m which is a half wavelength of 500 MHz as described above.

  In the numerical simulation, the generation source of the radio noise is assumed to be one place, and the reflected wave generated at one place is further considered. Accordingly, the direction of arrival is estimated with two pulsed radio noises arriving at the array antenna 1.

  First, a method for simulating the first received signal in the antennas 3A to 3D will be described. The waveform of FIG. 2 described above was used to simulate the first received signal in which the pulsed radio noise radiated from the source is received by the antenna when the source of the pulsed radio noise is one. However, it is assumed that the reflected wave is not included in this waveform. Here, the waveform shown in FIG. 2 is linearly interpolated into a received signal waveform measured by a single half-wave dipole antenna by applying a voltage to the sphere gap 12 to actually generate a spark discharge in the experimental arrangement shown in FIG. (Time interval 0.2 ns, score 4096).

The specifications of the experimental apparatus are as follows.
1) A ball gap 12 is hung on a track having a height of about 10 m, a voltage of 70 kV (effective value) is applied between the track and the ground, and a voltage is applied to the ball gap 12 by a partial pressure due to stray capacitance to cause a spark discharge. Generated.
2) The pulsed radio noise radiated by the spark discharge was received by the antennas 3A to 3D which are half-wave dipole antennas arranged at a position about 20 m away.
3) The center frequency of the half-wave dipole antenna was set to 500 MHz, and the received waveform of the first received signal was observed by the signal receiving unit 6 that was an oscilloscope.
4) The half-wave dipole antenna and the signal receiving unit 6 were connected by a coaxial cable 5 having a length of 7 m. The time sampling interval in the signal receiving unit 6 was 0.02 ns.

The frequency spectrum of the waveform of FIG. 2 is as shown in FIG. 3, and it can be seen from FIG. 2 that most of the frequency components of the waveform of FIG. 2 are 1 GHz or less. The waveform x _i (t) (i = 1 to 4) of the first reception signals of the antennas 3A to 3D was given as follows, with the waveform of FIG. 2 being S (t).

x _i (t) = s (t−τ _i1 ) + s (t−τ _i2 −t _d ) + n _i (t) (3)

Here, in consideration of the reflected wave, the first term on the right side of the above equation (3) represents the direct wave, and the second term on the right side represents the reflected wave. However, the attenuation at the time of reflection was not considered, and the waveform of the reflected wave was the same as the direct wave. t _d represents the time delay of the reflected wave with respect to the direct wave, and here, t _d = 20 ns. τ _i1 and τ _i2 represent the arrival time difference between the direct wave and the reflected wave of the antennas 3B to 3D with respect to the antenna 3A, and are given by the following equation (4).

τ _ij = (i−1) dsin θ _j / c (i = 1 to 4, j = 1, 2) (4)
Here, θ ₁ and θ ₂ are the arrival directions of the direct wave and the reflected wave, respectively, and c is the speed of light.

Further, as shown in Equation (3), n _{i (t)} is added to the received signal of each antenna 3A~3D to simulate measurement error due to thermal noise. n _i (t) is random noise, and is a Gaussian random number generated with an average of 0 mV and a standard deviation of the maximum value of s (t) × 0.02 mV.

First, a numerical simulation of direction-of-arrival estimation was performed when θ ₁ = 30 ° and θ ₂ = 60 ° and the subspace method was applied. That is, the direction of arrival was estimated using CSSM, which is one of the high resolution estimation methods for wideband signals. Here, RSS was used for calculation of the focusing matrix. The specific frequency used for the calculation of the focusing matrix is set to 500 MHz as the center frequency, the correlation matrix of the frequency of 300 MHz to 500 MHz is converted using the focusing matrix, and the correlation matrix at frequencies other than 300 MHz to 500 MHz is excluded. In addition, the calculation of the focusing matrix using RSS requires prior information on the arrival direction, and the focusing matrix was calculated for directions of 20 °, 40 °, and 70 °. The result of the direction of arrival estimated by the CSSM method is shown in FIG. As is apparent from FIG. 5, the estimation error is 1.5 ° in the direction of 30 ° and 0.5 ° in the direction of 60 °, and a good result is obtained.

Next, FIG. 6 shows the results when CSSM is applied to the closer arrival directions with θ ₁ = 30 ° and θ ₂ = 40 °. Here, a focusing matrix was calculated for directions of 25 °, 35 °, and 45 °.

In the case shown in FIG. 6, the direction of arrival can be estimated with an error of 1 ° with respect to the directions of θ ₁ = 30 ° and θ ₂ = 40 °.

  As described above, according to the present embodiment, prior information on the arrival direction is required, but arrival direction estimation for pulsed radio noise from a plurality of directions can be performed with extremely high accuracy.

  The present invention can be effectively used in the industrial field of managing and operating power equipment.

DESCRIPTION OF SYMBOLS 1 Array antenna 3A-3D antenna 6 Signal receiving part 8 Fourier transform part 9 1st correlation matrix calculating part 10 2nd correlation matrix calculating part 11 Arrival direction estimation part

Claims (4)

While receiving pulsed radio noise with an array antenna formed by arranging N (N is a natural number of 2 or more) antennas in parallel, a first received signal representing the level of the pulsed radio noise with respect to the time axis is obtained. Generating a second received signal representing a spectrum intensity at each frequency component by Fourier-transforming the first received signal;
Thereafter, a first correlation matrix is formed for each frequency component of the second received signal, and the first correlation matrix is weighted with respect to the spectrum intensity of each frequency component, and the first matrix is further weighted by a focusing matrix . By converting the correlation matrix of 1 into a correlation matrix of a specific frequency and performing the averaging process, the existence range of the frequency spectrum of the second received signal is limited and information on all frequencies in the existence range is included. Form a correlation matrix of two,
Further, the arrival direction of the pulsed radio noise with respect to the (N-1) direction is estimated by applying a subspace method on the narrowband signal to the second correlation matrix. Direction of arrival exploration method. An array antenna comprising N antennas (N is a natural number of 2 or more) arranged in parallel at intervals corresponding to a predetermined center frequency so as to receive pulsed radio noise;
A signal receiving unit for detecting a first received signal representing a level with respect to a time axis of the pulsed radio noise received via each antenna;
A Fourier transform unit that performs a Fourier transform on the first received signal to generate a second received signal representing a spectrum intensity at each frequency component;
A first correlation matrix computing unit that computes a first correlation matrix for each frequency component of the second received signal;
The first correlation matrix is weighted with respect to the spectrum intensity of each frequency component, and the first correlation matrix is converted into a correlation matrix of a specific frequency by a focusing matrix and an averaging process is performed. A second correlation matrix calculation unit that limits the existence range of the frequency spectrum of the received signal and calculates a second correlation matrix including information on all frequencies in the existence range;
An arrival direction estimation unit that estimates an arrival direction of the pulsed radio noise with respect to the (N-1) direction by applying a subspace method for a narrowband signal to the second correlation matrix; An arrival direction exploration device for pulsed radio noise. In the arrival direction exploration device for pulsed radio noise according to claim 2 ,
The arrival direction search apparatus for pulsed radio noise, wherein the specific frequency is the center frequency. In the arrival direction exploration device for pulsed radio noise according to claim 2 or claim 3 ,
An apparatus for searching the direction of arrival of pulsed radio noise, wherein the pulsed radio noise is radiated from power equipment.

Images