JP2016151469A - Signal processing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a signal processing device with which it is possible to suppress an arithmetic load and also detect a signal that does not necessarily have a linear shape.SOLUTION: A signal candidate detection unit 1 detects a point having a value in three-dimensional directions greater than or equal to a discretionary threshold as a signal candidate where the value of a point represented by two-dimensional characteristics is made a value in three-dimensional directions. A signal area estimation unit 2 estimates an area in which a signal exists on the basis of the density of points distributed on a plane that is equivalent to the threshold in the signal candidate detection unit 1. A signal determination unit 3 determines whether a point in the area estimated by the signal area estimation unit 2 is a signal or not.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a signal processing apparatus that detects a signal from signal candidates including noise.

  Conventionally, as a signal processing apparatus for detecting a true signal from signal candidates including noise, for example, as shown in Non-Patent Document 1, assuming that the gradient passing through the cell is brute force, Hough transform There is a method in which cells existing on the stacked straight lines are used as signals. Further, for example, as shown in Non-Patent Document 2, there is a method in which PDI is performed assuming that the starting point and inclination of a straight line are brute force and cells existing on the stacked straight lines are used as signals.

“Search Radar Detection and Track with the Hough Transform, I to III”, IEEE Trans. On Aerospace and Electronic Systems, 1994. “Improvement of acceleration target detection performance of Doppler radar using PDI method”, IEICE Japanese Journal B, 1999.

However, the conventional method has the following problems.
・ Problem 1: The calculation load is high because a straight line is assumed in round robin.
Problem 2: Since a straight line is estimated, it is difficult to detect a signal that changes in a curved line due to Doppler shift, a signal having a modulation bandwidth, and a linear signal such as a pulse instead of a straight line.

  The present invention has been made in order to solve the above-described problems, and it is an object of the present invention to obtain a signal processing device that can reduce the calculation load and can detect even a signal that does not necessarily have a linear shape. Objective.

  The signal processing apparatus according to the present invention detects a point having a value greater than or equal to an arbitrary threshold in the three-dimensional direction as a signal candidate when the value of the point represented by the two-dimensional characteristic is a value in the three-dimensional direction. A detection unit, a signal region estimation unit for estimating a region where a signal exists based on a density of points distributed on a plane corresponding to a threshold value in the signal candidate detection unit, and a point in the region estimated by the signal region estimation unit And a signal determination unit for determining whether or not is a signal.

  The signal processing apparatus according to the present invention detects a point having a value equal to or larger than an arbitrary threshold value in a three-dimensional direction as a signal candidate, and the presence of a signal is present based on the density of points distributed on a plane corresponding to the threshold value of the signal candidate. Since the region is estimated and the presence / absence of a signal is determined in this region, it is possible to detect even a signal that suppresses the computation load and does not necessarily have a linear shape.

BRIEF DESCRIPTION OF THE DRAWINGS It is a block diagram which shows the signal processing apparatus by Embodiment 1 of this invention. It is a flowchart which shows operation | movement of the signal processing apparatus by Embodiment 1 of this invention. It is explanatory drawing of the threshold value in the signal candidate detection part of the signal processing apparatus by Embodiment 1 of this invention. It is explanatory drawing which shows the concept of the signal region estimation in the signal region estimation part of the signal processing apparatus by Embodiment 1 of this invention. It is a flowchart which shows the detail of operation | movement in the signal area estimation part of the signal processing apparatus by Embodiment 1 of this invention. It is explanatory drawing of the processing formula corresponding to each step of FIG. It is explanatory drawing which shows the example of a process result of the signal processing apparatus by Embodiment 1 of this invention. It is an enlarged view of the processing result of FIG. It is explanatory drawing which shows the arithmetic expression corresponding to each process of the signal processing apparatus by Embodiment 2 of this invention. It is explanatory drawing which shows the concept of the coherent integration in the signal determination part of the signal processing apparatus by Embodiment 3 of this invention. It is explanatory drawing which shows the cell in the ellipse of the signal processing apparatus by Embodiment 4 of this invention.

Embodiment 1 FIG.
1 is a block diagram showing a signal processing apparatus according to Embodiment 1 of the present invention.
The signal processing apparatus shown in FIG. 1 includes a signal candidate detection unit 1, a signal region estimation unit 2, and a signal determination unit 3. The signal candidate detection unit 1 is a processing unit that detects a point having a value equal to or larger than an arbitrary threshold value in the three-dimensional direction as a signal candidate when the value of the point represented by the two-dimensional characteristic is a value in the three-dimensional direction. . The signal region estimation unit 2 is a processing unit that estimates a region where a signal exists based on the density of points distributed on a plane corresponding to the threshold value in the signal candidate detection unit 1. The signal determination unit 3 is a processing unit that determines whether or not a point in the region estimated by the signal region estimation unit 2 is a signal.

FIG. 2 is a flowchart showing the operation of the signal processing apparatus according to the first embodiment.
First, the signal candidate detection unit 1 detects points that are equal to or greater than a predetermined threshold as signal candidates (step ST1).
FIG. 3 is an explanatory diagram of threshold values in the signal candidate detection unit 1 and shows a relationship between a time-frequency plot of a received signal and a detection threshold value in a low SNR environment. Here, a point represented by a two-dimensional characteristic of time and frequency is defined as a cell. As shown in FIG. 3A, when the threshold is set high so that false alarms (noise other than signals, etc.) do not exceed the threshold, the low SNR signal cannot exceed the threshold and signal detection becomes difficult. . Therefore, in the signal candidate detection unit 1 of the present invention, as shown in FIG. 3B, the threshold value is set low so that a signal that has not exceeded the detection threshold value can exceed the threshold value. For example, in the example shown in FIG. 3A, the detection threshold is about 3.8, but in the example shown in FIG. 3B, it is about 2.1. As a result, the false alarm also exceeds the threshold, but the signal candidate detection unit 1 outputs these cells as signal candidate cells.

Next, as described in detail below, the signal region estimation unit 2 sets the density of the points distributed on the threshold plane, that is, the density of cells distributed on the plane corresponding to the threshold in cells having a value equal to or greater than the threshold. A region where a signal exists is estimated based on an elliptical region equal to or greater than the value (step ST2).
FIG. 4 is an explanatory diagram showing the concept of signal region estimation in the signal region estimation unit 2. As shown in the figure, a convergence operation is performed on the ellipse area and center set in the initial PCA (principal component analysis) to calculate an area and an ellipse center with a high possibility of signal components.

FIG. 5 is a flowchart showing details of the operation in the signal region estimation unit 2. FIG. 6 is an explanatory diagram showing a processing expression corresponding to each step.
First, as an initial ellipse, an average value and a covariance matrix of all cells are used as a matrix for determining an ellipse center and an ellipse shape, respectively (step ST21). Next, the inside / outside determination of an ellipse is performed with respect to all the cells (step ST22). Next, the distance from the center of the ellipse of the cell that entered the ellipse, that is, the residual is calculated (step ST23). Next, the reliability of each cell is calculated (step ST24). Next, each cell is weighted by reliability, and the covariance matrix that determines the shape of the ellipse is updated (step ST25). By doing so, since the reliability of the cell greatly deviating from the center of the ellipse is small, the contribution to the update of the covariance matrix is reduced. The updated covariance matrix is updated to a shape including cells that are relatively close to the center of the ellipse.

  Next, the center of the ellipse is updated with the cells that have entered the ellipse (step ST26). Here, the average value of the cells in the ellipse is the new center. By doing so, the center of the ellipse moves in the direction of high density (the direction in which the signal exists). Further, it is determined whether or not it has converged (step ST27). As a criterion, if the cells included in the ellipse are the same in the pre-process and the current process, it is assumed that the cells have converged, and the calculation is terminated. “Same” means that only the cells at the same position are included in the pre-processing and the current processing. Otherwise, it is assumed that it has not converged. If not converged, the process returns to step ST22 and a series of processing is repeated.

FIG. 7 shows an example of the processing result of the first embodiment, and FIG. 8 shows an enlarged view of the processing result of FIG. It is assumed that the false alarm of Pfa = 10 ⁻² has risen as a result of lowering the threshold to the extent that the pulse signal (cell in the area indicated by the elliptical area 10) exceeds the threshold. 7 and 8, it can be seen that it converges to an ellipse that includes only signal cells and does not include false alarms.

  When the signal region estimation processing in the signal region estimation unit 2 is completed, the signal determination unit 3 determines whether or not a point in the region estimated by the signal region estimation unit 2 is a signal (step ST3). This process will be described in detail in the third and fourth embodiments.

  As described above, in the signal processing apparatus according to the first embodiment, the signal is not obtained by “narrowing down a region where cells are densely present”, instead of “brute force” as illustrated in the conventional non-patent documents 1 and 2. Estimate "region" instead of "straight line". Therefore, since it is not brute force, the calculation load can be reduced. In addition, since it is an “area” instead of a “straight line”, a signal component exists even for a signal that changes in a curved line due to Doppler shift, a signal that has a modulation bandwidth, and a linear signal that is not a straight line. Only cells can be identified and integrated.

  As described above, according to the signal processing apparatus of the first embodiment, when the value of a point represented by a two-dimensional characteristic is a value in a three-dimensional direction, a value equal to or greater than an arbitrary threshold value in the three-dimensional direction. A signal candidate detection unit that detects points as signal candidates, a signal region estimation unit that estimates a region where a signal exists based on a density of points distributed on a plane corresponding to a threshold in the signal candidate detection unit, and a signal region estimation And a signal determination unit that determines whether or not a point in the region estimated by the unit is a signal, so that even a signal that does not necessarily have a linear shape can be detected while reducing the computation load. can do.

  In addition, according to the signal processing device of the first embodiment, the signal region estimation unit estimates the region where the signal exists based on the elliptic region where the density distributed on the plane is equal to or higher than the set value. Processing with a reduced load can be performed.

  Further, according to the signal processing apparatus of the first embodiment, the signal region estimation unit performs a convergence calculation of the size and center position of the elliptic region using the likelihood and average value of the points in the elliptic region, and performs signal convergence processing. Since the existing area is estimated, it is possible to perform processing with reduced calculation load.

Embodiment 2. FIG.
In the first embodiment, the average position of the cells in the ellipse is set as the new ellipse center. However, in the second embodiment, the center of the ellipse is moved faster by the dense cell. A weighted average with a large weight placed on is performed. Here, the configuration of the signal processing apparatus according to the second embodiment in the drawing is the same as that in FIG. 1, and will be described using the configuration in FIG. 1.

  In addition to the function of the signal region estimation unit 2 of the first embodiment, the signal region estimation unit 2 of the second embodiment weights the center position of the ellipse region by the likelihood according to the distance between points in the ellipse region. It is configured to calculate the average. Since the signal candidate detection unit 1 and the signal determination unit 3 are the same as those in the first embodiment, description thereof is omitted here.

最後に、

Next, the following [Process 1] to [Process 4] will be described as operations of the signal region estimation unit 2 of the second embodiment. FIG. 9 shows an arithmetic expression corresponding to each process.

Finally,

  As described above, according to the signal processing device of the second embodiment, the signal region estimation unit performs weighted average calculation based on the likelihood according to the distance between points in the elliptic region, and calculates the center position of the elliptic region. Since the convergence calculation is performed, the convergence calculation can be performed promptly, and thus the speed of the signal detection process can be increased.

Embodiment 3 FIG.
The third embodiment relates to the signal determination unit 3 shown in FIG. 1, and the signal candidate detection unit 1 and the signal region estimation unit 2 are the same as those in the first embodiment or the second embodiment. The description in is omitted.

また、雑音が付加された場合、次式（２）

となる。ここで、μはチャープ係数と呼ばれ、単位時間当たりの周波数の変動量を示す。
ここでは、楕円内にチャープパルス信号が得られた場合、コヒーレントに積分する方法を示す。 The signal determination unit 3 of the third embodiment assumes that the cell indicates the value of the chirp pulse, estimates the frequency displacement speed of the chirp from the inclination of the major axis of the ellipse, performs coherent integration along the major axis, and the result Based on the above, it is configured to determine whether or not a signal exists in the elliptical region.
FIG. 10 is an explanatory diagram illustrating the concept of coherent integration in the signal determination unit 3. The figure shows an ellipse obtained by the region estimation algorithm.
Here, a linear chirp pulse signal is assumed. That is, the signal can be expressed by the following equation (1).

When noise is added, the following equation (2)

It becomes. Here, μ is called a chirp coefficient, and indicates a fluctuation amount of the frequency per unit time.
Here, a method of coherent integration when a chirp pulse signal is obtained within an ellipse is shown.

Here, t ₂ -t ₁ and f ₂ -f ₁ are the chirp time width and frequency width shown in FIG. 10, respectively, and can be obtained from the length and direction of the axis of the ellipse. Finally, coherent integration is performed using the following equation. For example, when coherent integration is performed using the sum of squares, the integration formula is expressed by the following formula (4).

閾値をthresholdとすると、

の場合雑音、

の場合信号と判定する。
このように、収束演算で得られた楕円内のセルを積分することにより、信号が存在するか否かを判定することができる。

If the threshold is threshold,

In case of noise,

In the case of, it is determined as a signal.
In this way, it is possible to determine whether or not a signal exists by integrating the cells in the ellipse obtained by the convergence calculation.

  As described above, according to the signal processing apparatus of the third embodiment, the signal determination unit determines whether or not a signal exists in the region by performing coherent integration along the long axis of the ellipse. As a result, the presence or absence of a signal in the region can be reliably determined.

  Further, according to the signal processing device of the third embodiment, the signal determination unit uses the point as a point indicating the value of the chirp pulse, estimates the frequency displacement speed of the chirp by the inclination of the major axis of the ellipse, and follows the major axis. The coherent integration is performed, and based on the result, it is determined whether or not there is a signal in the elliptical region. Therefore, the presence or absence of the signal in the region can be reliably determined.

Embodiment 4 FIG.
The fourth embodiment relates to the signal determination unit 3 shown in FIG. 1, and the signal candidate detection unit 1 and the signal region estimation unit 2 are the same as those in the first embodiment or the second embodiment. The description in is omitted.

すなわち、図１１に示すように、楕円内の全てのセルの振幅｜ｙ_ｎ｜をインコヒーレント積分する。 The signal determination unit 3 according to the fourth embodiment is configured to determine whether or not a signal exists in the region based on the incoherent integration results of all points in the region estimated by the signal region estimation unit 2. Yes. Hereinafter, the operation of the signal determination unit 3 will be described.
In the fourth embodiment, as a method for determining whether or not a signal really exists in the ellipse obtained as a result of the convergence calculation in step ST3 shown in FIG. 2, the cells in the region are integrated incoherently and the size thereof is determined. The presence or absence of a signal is determined by The cells in the ellipse are incoherently integrated as in the following equation (9), and the presence or absence of a signal is determined based on the magnitude.

That is, as shown in FIG. 11, the amplitudes | y _n | of all the cells in the ellipse are incoherently integrated.

一方、雑音のみの場合は次式（１１）となる。

When there is a signal, the following equation (10) is obtained.

On the other hand, in the case of only noise, the following equation (11) is obtained.

の場合雑音、

の場合信号と判定する。
このように、収束演算で得られた楕円内のセルを積分することにより、信号が存在するか否かを判定することができる。 Comparing equation (10) and equation (11), it can be seen that the value is greater when a signal is present. Since the sizes are different, the presence / absence of a signal can be finally determined by the height of the stack. The determination threshold can be obtained, for example, by assuming a noise probability distribution. A threshold can be set for the integral value. If the threshold is threshold,

In case of noise,

In the case of, it is determined as a signal.
In this way, it is possible to determine whether or not a signal exists by integrating the cells in the ellipse obtained by the convergence calculation.

  As described above, according to the signal processing apparatus of the fourth embodiment, the signal determination unit has a signal in the region based on the incoherent integration results of all points in the region estimated by the signal region estimation unit. Since it is determined whether or not to perform, it is possible to reliably determine the presence or absence of a signal in the region.

  In each of the above embodiments, time and frequency are set as the two-dimensional characteristics, and amplitude is set as the value in the three-dimensional direction. However, the values are not limited to these values. For example, the two-dimensional characteristics include time and direction. It can be applied to all values such as frequency and direction.

  Further, within the scope of the present invention, the invention of the present application can be freely combined with each embodiment, modified with any component in each embodiment, or omitted with any component in each embodiment. .

  1 signal candidate detection unit, 2 signal region estimation unit, 3 signal determination unit.

Claims (7)

When a value of a point represented by a two-dimensional characteristic is a value in a three-dimensional direction, a signal candidate detection unit that detects a point having a value equal to or larger than an arbitrary threshold in the three-dimensional direction as a signal candidate
A signal region estimation unit that estimates a region where a signal exists based on a density of points distributed on a plane corresponding to the threshold value in the signal candidate detection unit;
A signal processing apparatus comprising: a signal determination unit that determines whether or not a point in the region estimated by the signal region estimation unit is a signal.   The signal processing apparatus according to claim 1, wherein the signal region estimation unit estimates a region in which a signal exists based on an elliptic region having a density distributed on the plane equal to or higher than a set value.   The signal region estimation unit estimates a region where the signal exists by performing a convergence calculation of the size and center position of the elliptic region using the likelihood and average value of points in the elliptic region. Item 3. The signal processing device according to Item 2.   4. The signal processing according to claim 3, wherein the signal region estimation unit performs a weighted average calculation with a likelihood according to a distance between points in the elliptic region and performs a convergence calculation of a center position of the elliptic region. apparatus.   The said signal determination part determines whether a signal exists in an area | region by performing a coherent integration along the long axis of an ellipse, The any one of Claims 2-4 characterized by the above-mentioned. The signal processing device according to item.   The signal determination unit uses the point as a point indicating the value of the chirp pulse, estimates the chirp frequency displacement speed by the inclination of the major axis of the ellipse, performs coherent integration along the major axis, and based on the result 5. The signal processing apparatus according to claim 2, wherein it is determined whether or not a signal exists in an elliptical area.   The signal determination unit determines whether a signal exists in the region based on an incoherent integration result of all points in the region estimated by the signal region estimation unit. The signal processing device according to claim 6.

Images