Patent Application
20200379099
The present invention relates to a radar signal processing device used in a pulse-Doppler radar or a Doppler lidar (laser radar) that obtains the intensity of precipitation, wind speed distribution, or the like, in the air.
For devices that observe a discrete distribution type target such as precipitation droplets (rain droplets) in the air or aerosols to obtain the intensity of precipitation or wind speed distribution, a pulse-Doppler radar, a Doppler lidar, a Doppler sodar, a wind profiler, etc., are used. These devices radiate a pulsed electromagnetic wave or sound wave into space as a transmission signal, and receive a reflected wave of the transmission signal from precipitation droplets, atmospheric turbulence, aerosols, or the like. Then, the intensity of precipitation or the like is calculated from the signal strength of a reception signal, and wind speed or the like is calculated from Doppler frequency, and in general, those calculations are derived as spectral moments (signal strength: zero order, Doppler velocity: first order, and Doppler velocity width: second order) which are obtained by converting a time-domain signal to a frequency-domain signal.
Hereinafter, a Doppler lidar will be described as an example, but the same also applies to other devices.
A Doppler lidar radiates a pulsed laser beam having a predetermined time length in a predetermined direction in the air, and receives a reflected wave from aerosols that move at the same speed as wind in the air, and the reflected wave includes overall information about the aerosols in a scattering volume which is determined by a predetermined pulse width and a beam width. Namely, the signal strength which is a zero-order spectral moment is the overall signal strength of the aerosols in the scattering volume, the Doppler velocity which is a first-order spectral moment is the overall Doppler velocity of the aerosols in the scattering volume, and the Doppler velocity width which is a second-order spectral moment is the overall Doppler velocity width of the aerosols in the scattering volume.
As described above, since the Doppler lidar obtains observed values on a per given finite scattering volume, it is difficult to grasp local characteristics based on smaller spatial scale phenomena than the scattering volume. For example, in a case in which the pulse width of the Doppler lidar is 30 m, since the radius of the highest-wind-speed portion (core) of vortex turbulence called wake vortex turbulence which occurs at the rear of an aircraft is on the order of several meters even for large aircraft, when the vortex is observed at a pulse width of 30 m, not only the wind speed of the core portion but also a region therearound whose wind speed is not as high as the core portion is included in the same scattering volume. As a result, overall Doppler velocity which is obtained as an observed value is smaller than the original maximum velocity of the vortex, and there is a possibility that the hazard is underestimated. In addition, when the aircraft wingspan is less than the pulse width, two vortexes with opposite rotation directions are both included in the same scattering volume, and since Doppler velocities having different signs cancel each other out, it may be difficult to detect the vortexes.
For a reduction in an observation region to grasp local characteristics (hereinafter, referred to as implementation of high resolution), there is a possible method in which by reducing a pulse width to be transmitted, a scattering volume is reduced, but in this method, since the energy of a pulse to be transmitted decreases, there is a problem that a range that can be observed is reduced.
In addition, there is a possible method (pulse compression) in which a pulse having been subjected to band widening such as phase modulation or frequency modulation is transmitted and a reception signal is demodulated, by which implementation of high resolution is achieved, but there is a problem that a device configuration is complex and increases cost.
Hence, as a method that does not involves a reduction in observed range and that avoids a significant change (an increase in cost) in device configuration, for example, as shown in Non-Patent Literature 1, there is a method of optimizing a coefficient that is multiplied to complex time-series signals from a plurality of scattering volumes which are overlapped (which is written as “oversampled” in Non-Patent Literature 1, but “overlapped” and “oversampled” have the same meaning in a sense that signals of a predetermined scattering volume length are extracted from oversampled reception signals while the signals are partially overlapped) to increase the gain of a small region of interest.
However, in the conventional method described in the above-described related art, although characteristics of a small region of interest are obtained, in order to obtain the characteristics, an optimization technique that frequently uses inverse matrix computation or iterative computation of complex signals is adopted and thus, the amount of computation increases, causing a problem that it is difficult to perform a real-time process upon wide-area observation or upon high-speed observation.
The invention is made to solve a problem such as that described above, and an object of the invention is to provide a radar signal processing device capable of grasping local characteristics of a set region with a low amount of computation and in real time even upon wide-area observation or upon high-speed observation.
A radar signal processing device according to the invention includes: a transceiver to radiate a transmission signal made of an electromagnetic wave or a sound wave into air, and receive a reflected signal, the reflected signal being the transmission signal reflected by a target in the air; a signal cutting-out unit to cut out the reflected signal received by the transceiver by a specific length while the cut-out signals are overlapped by a set length in at least either one of a range direction and a cross-range direction; a power spectrum calculator to calculate a plurality of power spectra from the signals cut out by the signal cutting-out unit; a power spectrum reconstructor to perform power spectrum reconstruction by changing a ratio or contribution rate of power spectrum components in a set region, using the plurality of power spectra; and a spectral moment calculator to calculate a spectral moment from the power spectrum reconstructed by the power spectrum reconstructor.
The radar signal processing device of the invention cuts out a signal reflected from a target by a specific length while the cut-out signals are overlapped by a set length, to calculate a plurality of power spectra, and reconstructs a power spectrum using the plurality of power spectra. By this, local characteristics of a set region can be grasped with a low amount of computation and in real time.
To describe the invention in more detail, modes for carrying out the invention will be described below with reference to the accompanying drawings.
The radar signal processing device shown in
The control unit 1 is a control unit that controls each unit of the radar signal processing device, and is configured to output control signals such as specifications for pulse modulation, a trigger signal, and an observation direction to the signal transmission and reception controlling unit 20 in the signal transmitting and receiving unit 2, and output control signals such as specifications for pulse modulation, a trigger signal, specifications for signal cut-out, specifications for power spectrum calculation, specifications for power spectrum reconstruction, and specifications for spectral moment calculation to the signal processing control unit 30 in the signal processing unit 3, and further output a control signal about specifications for display to the display unit 4.
The signal transmission and reception controlling unit 20 in the signal transmitting and receiving unit 2 outputs, on the basis of the trigger signal obtained from the control unit 1, control signals for the pulse modulating unit 23 and the optical antenna 25 among the control signals obtained from the control unit 1. The light source 21 outputs continuous-wave light of a single frequency, and a laser beam from the light source 21 is outputted to the divider 22. The divider 22 provides a predetermined frequency shift to the laser beam from the light source 21 and divides the laser beam into two laser beams, and one of the two laser beams divided by the divider 22 is outputted to the pulse modulating unit 23 and the other is outputted to the combining unit 26. The pulse modulating unit 23 performs pulse modulation on the laser beam inputted from the divider 22. The laser beam having been subjected to the pulse modulation by the pulse modulating unit 23 is outputted to the circulator 24. The circulator 24 switches between output destinations based on an input source of a laser beam. The circulator 24 outputs a laser beam from the pulse modulating unit 23 to the optical antenna 25, and outputs a laser beam from the optical antenna 25 to the combining unit 26.
The optical antenna 25 irradiates the laser beam from the circulator 24 into the air, and collects scattered light of the laser beam from aerosols. The scattered light collected by the optical antenna 25 is outputted to the combining unit 26 through the circulator 24. The combining unit 26 combines the laser beam from the divider 22 with the laser beam outputted from the optical antenna 25 through the circulator 24. The laser beam combined by the combining unit 26 is outputted to the optical receiving unit 27. The optical receiving unit 27 performs heterodyne detection on the laser beam from the combining unit 26, and thereby converts the laser beam into an electrical signal. The electrical signal converted by the optical receiving unit 27 is outputted to the A/D converting unit 28. The A/D converting unit 28 performs Analog-to-Digital (A/D) conversion on the electrical signal from the optical receiving unit 27. A reception signal which is converted into a signal of a predetermined sampling rate by the A/D converting unit 28 is outputted to the signal processing unit 3.
The signal processing control unit 30 in the signal processing unit 3 outputs, on the basis of the trigger signal obtained from the control unit 1, control signals for the signal cutting-out unit 31 to the velocity vector calculating unit 35 among the control signals obtained from the control unit 1. The signal cutting-out unit 31 performs, on the basis of the control signal from the signal processing control unit 30, a process of cutting out the reception signal from the A/D converting unit 28 in the signal transmitting and receiving unit 2 by a specific length while the cut-out signals are overlapped by a set length in at least either one of a range direction and a cross-range direction, and outputs the cut-out signals to the power spectrum calculating unit 32. The power spectrum calculating unit 32 multiplies the signals cut out by the signal cutting-out unit 31 by a predetermined window function on the basis of the control signal from the signal processing control unit 30, and thereafter performs a Fast Fourier Transform (FFT) at a predetermined number of points to calculate a plurality of power spectra, and stores the plurality of power spectra in a storage device which is not shown. The power spectrum reconstructing unit 33 obtains the plurality of power spectra calculated by the power spectrum calculating unit 32 from the storage device which is not shown, on the basis of the control signal from the signal processing control unit 30, performs power spectrum reconstruction on the basis of a predetermined method for power spectrum reconstruction, with the aim of increasing the ratio or contribution rate of power spectrum components in a set region, and outputs a reconstructed power spectrum to the spectral moment calculating unit 34.
The spectral moment calculating unit 34 calculates, on the basis of the control signal from the signal processing control unit 30, a spectral moment (zero-order, first-order, second-order, etc.) from the reconstructed power spectrum inputted from the power spectrum reconstructing unit 33, and stores the spectral moment in a storage device which is not shown or outputs the spectral moment to the display unit 4, or outputs radial velocity which is a first-order moment to the velocity vector calculating unit 35. The velocity vector calculating unit 35 calculates, on the basis of the control signal from the signal processing control unit 30, a velocity vector using Doppler velocity (radial velocity) which is a first-order moment calculated by the spectral moment calculating unit 34, on the basis of a predetermined method for velocity vector calculation, and stores the velocity vector in a storage device which is not shown or outputs the velocity vector to the display unit 4.
The display unit 4 displays, on the basis of the control signal from the control unit 1, spectral moment information or velocity vector information which is inputted from the signal processing unit 3, on the basis of a predetermined scheme.
Next, operation of the radar signal processing device of the first embodiment will be described. First, operation of the signal transmitting and receiving unit 2 will be described.
A laser beam from the light source 21 is subjected to a predetermined frequency shift and divided by the divider 22, and one of the divided laser beams is outputted to the pulse modulating unit 23 and the other is outputted to the combining unit 26. The pulse modulating unit 23 performs pulse modulation on the laser beam. The laser beam outputted from the pulse modulating unit 23 is provided to the optical antenna 25 through the circulator 24 and irradiated into the air from the optical antenna 25. The irradiated laser beam is reflected by a target such as aerosols, and the reflected light is received by the optical antenna 25. The reflected light received by the optical antenna 25 is provided to the combining unit 26 through the circulator 24. The combining unit 26 combines the laser beam from the divider 22 with the reflected light from the circulator 24, and outputs the combined light to the optical receiving unit 27. The optical receiving unit 27 performs heterodyne detection on the output light from the combining unit 26, and thereby converts the output light into an electrical signal. The electrical signal outputted from the optical receiving unit 27 is converted into a digital signal by the A/D converting unit 28, and the digital signal is outputted to the signal processing unit 3.
Next, operation of the signal processing unit 3 will be described.
In the signal processing unit 3, first, the signal cutting-out unit 31 sets specifications for signal cut-out on the basis of a control signal from the signal processing control unit 30 (step ST101). Note that the specifications include a method, a coefficient, etc. The signal cutting-out unit 31 cuts out a signal provided from the signal transmitting and receiving unit 2, on the basis of the set signal cut-out method (step ST102). The power spectrum calculating unit 32 sets specifications for power spectrum calculation on the basis of a control signal from the signal processing control unit 30 (step ST103), and calculates power spectra on the basis of the specifications for power spectrum calculation (step ST104). In addition, the power spectrum reconstructing unit 33 sets specifications for power spectrum reconstruction on the basis of a control signal from the signal processing control unit 30 (step ST105), and reconstructs a power spectrum on the basis of the specifications for power spectrum reconstruction (step ST106). Furthermore, the spectral moment calculating unit 34 sets specifications for spectral moment calculation on the basis of a control signal from the signal processing control unit 30 (step ST107), and calculates a spectral moment on the basis of the specifications for spectral moment calculation (step ST108). Then, the velocity vector calculating unit 35 calculates a velocity vector on the basis of specifications for velocity vector calculation set on the basis of a control signal from the signal processing control unit 30 (step ST109). Operation of the units from the signal cutting-out unit 31 to the velocity vector calculating unit 35 will be described in more detail below.
The signal cutting-out unit 31 cuts out a reception signal, which is received from the signal transmitting and receiving unit 2 and sampled at a predetermined sampling rate, by a specific time length while the cut-out signals are overlapped by a preset length in at least either one of a range direction and a cross-range direction. At this time, the overlapping length is determined by the length of a set region, and the cut-out length in a time direction (range direction) is determined by transmission pulse width. These lengths are set on the signal cutting-out unit 31 by the control unit 1 through the signal processing control unit 30.
As shown in
Note that the number of signals cut out for a single set region is determined by a reconstruction method of the power spectrum reconstructing unit 33, and the reconstruction method is set by the control unit 1 through the signal processing control unit 30.
The power spectrum calculating unit 32 performs a Fourier transform (a specific technique is a Fast Fourier Transform (FFT) process) on the reception signals cut out by the signal cutting-out unit 31, on a per reception signal basis and thereby calculates a plurality of power spectra. Note that in general when an FFT is performed on a finite-length signal, a time signal is multiplied by a window function (e.g., a Hamming window or a Han (Hanning) window). In addition, since a reception signal of one pulse (corresponding to the scattering volume) is very weak and is susceptible to fluctuations such as noise, in general, a process is performed in which a predetermined number of incoherent integrations (non-coherent integrations) are performed to reduce fluctuations, improving detectability.
The power spectrum calculating unit 32 outputs the plurality of power spectra calculated in the above-described manner to a storage device which is not shown.
The power spectrum reconstructing unit 33 reconstructs a power spectrum having a high ratio or contribution rate of a set region, using each power spectrum calculated by the power spectrum calculating unit 32 and superimposed on the set region. A possible method for power spectrum reconstruction includes a method for maximizing a coefficient of a set region obtained after weighted addition of power spectra.
In this example, the scale of a set region in the range direction is ⅕ of the original scattering volume (the number of division d=5), and the number of scattering volumes superimposed on the set region (the number of scattering volumes n to be added together) is nine. For weight coefficients 401 to 405 shown in
In addition, when weighted addition is performed in the range direction (time direction), a value obtained by multiplying the weight coefficients 401 to 405 by a transmission pulse shape wt and a window function wf which is used upon a power spectrum calculation process can be used as a weight coefficient.
Another possible method for power spectrum reconstruction includes a method in which a set region is considered to be a common portion between the power spectra of adjacent scattering volumes, and a minimum value on a frequency axis (common portion) is adopted.
the number of scattering volumes: n (=2*d−1 (d: the number of divisions)),
the power spectrum of a scattering volume: Vi(f) (i=1, . . . , n),
weight coefficient: wi (i=1, . . . , n), and
the power spectrum of a set region: y(f)
y(f)=min([wi·Vi(f)]), (i=1, . . . ,n).
In an example of
The spectral moment calculating unit 34 calculates a spectral moment from the power spectrum reconstructed by the power spectrum reconstructing unit 33. For a method for spectral moment calculation, for example, a zero-order moment (reception strength), a first-order moment (Doppler velocity), or a second-order moment (Doppler velocity width) is calculated. The spectral moment calculating unit 34 outputs the calculated spectral moment to the velocity vector calculating unit 35.
The velocity vector calculating unit 35 calculates a velocity vector from the obtained spectral moment. Note that for a method for calculating a velocity vector, general methods for velocity vector calculation (e.g., a Velocity Azimuth Display (VAD) method, a 3-beam method, a 4-beam method, and a 5-beam method) can be used.
As described above, the radar signal processing device of the first embodiment includes a signal transmitting and receiving unit that radiates a transmission signal made of an electromagnetic wave or a sound wave into the air, and receives a reflected signal which is the transmission signal reflected by a target in the air; a signal cutting-out unit that cuts out the reflected signal received by the signal transmitting and receiving unit by a specific length while the cut-out signals are overlapped by a set length in at least either one of a range direction and a cross-range direction; a power spectrum calculating unit that calculates a plurality of power spectra from the signals cut out by the signal cutting-out unit; a power spectrum reconstructing unit that performs power spectrum reconstruction by changing the ratio or contribution rate of power spectrum components in a set region, using the plurality of power spectra; and a spectral moment calculating unit that calculates a spectral moment from the power spectrum reconstructed by the power spectrum reconstructing unit, and thus, can grasp local characteristics of the set region with a low amount of computation and in real time.
In addition, according to the radar signal processing device of the first embodiment, since the power spectrum reconstructing unit performs the power spectrum reconstruction by performing weighted addition of the plurality of power spectra, a reconstructed power spectrum with high accuracy can be obtained.
In addition, according to the radar signal processing device of the first embodiment, since the power spectrum reconstructing unit performs the power spectrum reconstruction by extracting a minimum value as a common portion between the plurality of power spectra, a reconstructed power spectrum with high accuracy can be obtained.
Although the first embodiment shows some methods for power spectrum reconstruction in the range direction and the cross-range direction, since the accuracy of a reconstructed power spectrum (superiority and inferiority between construction methods) depends on the actual air conditions, it is desirable to include means for evaluating the effectiveness of a power spectrum. Hence, a second embodiment shows an example in which after performing a plurality of power spectrum reconstructions, the effectiveness of reconstructed power spectra is evaluated.
Next, operation of the radar signal processing device of the second embodiment will be described.
As such, in the second embodiment, it is configured to determine and select the most ideal spectrum from a plurality of reconstructed power spectra, and thus, a reconstructed power spectrum with higher accuracy can be obtained.
As described above, according to the radar signal processing device of the second embodiment, the power spectrum reconstructing unit performs a plurality of different processes for power spectrum reconstruction, and selects a power spectrum close to a reference spectral shape from among a plurality of obtained power spectra, and thus, a reconstructed power spectrum with higher accuracy can be obtained.
In the second embodiment, after performing a plurality of power spectrum reconstructions, a power spectrum is selected with reference to whether or not the power spectrum is close to the most ideal power spectrum, but by averaging a plurality of power spectra, a power spectrum with a reduced bias component can be obtained. Hence, a third embodiment shows an example in which power spectra are subjected to an averaging process.
As such, in the third embodiment, it is configured to average a plurality of reconstructed power spectra, and thus, a bias component is reduced, making it possible to obtain a reconstructed power spectrum with higher accuracy.
As described above, according to the radar signal processing device of the third embodiment, the power spectrum reconstructing unit performs a plurality of different processes for power spectrum reconstruction, and outputs a result of averaging a plurality of obtained power spectra, and thus, a reconstructed power spectrum with higher accuracy can be obtained.
In the third embodiment, a power spectrum with a small bias is obtained by averaging a plurality of results of power spectrum reconstruction, but it is also possible to determine spectral moments for each of a plurality of power spectra and average the spectral moments, which will be described as a fourth embodiment.
As such, in the fourth embodiment, it is configured to average a plurality of spectral moments, and thus, an advantageous effect of a reduction in a bias component and an advantageous effect of a reduction in the amount of computation can be obtained.
As described above, according to the radar signal processing device of the fourth embodiment, the power spectrum reconstructing unit performs a plurality of different processes for power spectrum reconstruction, and the spectral moment calculating unit calculates a plurality of spectral moments from a plurality of power spectra obtained in the plurality of different processes for power spectrum reconstruction, and outputs a result of averaging the plurality of spectral moments, and thus, a reconstructed power spectrum with higher accuracy can be obtained.
Note that in the invention of this application, a free combination of the embodiments, modifications to any component of the embodiments, or omissions of any component in the embodiments are possible within the scope of the invention.
As described above, radar signal processing devices according to the invention relate to a configuration in which a plurality of power spectra are calculated by cutting out a signal reflected from a target by a specific length while the cut-out signals are overlapped by a set length, and a power spectrum is reconstructed using the plurality of power spectra, and are suitable for use in a pulse-Doppler radar or a Doppler lidar.
1: control unit, 2: signal transmitting and receiving unit, 3: signal processing unit, 4: display unit, 20: signal transmission and reception controlling unit, 21: light source, 22: divider, 23: pulse modulating unit, 24: circulator, 25: optical antenna, 26: combining unit, 27: optical receiving unit, 28: A/D converting unit, 30: signal processing control unit, 31: signal cutting-out unit, 32: power spectrum calculating unit, 33, 33a, 33b, 33c: power spectrum reconstructing unit, 34, 34a: spectral moment calculating unit, 35: velocity vector calculating unit, 331-1: first power spectrum reconstructing unit, 331-2: second power spectrum reconstructing unit, 331-N: Nth power spectrum reconstructing unit, 332: power spectral shape determining unit, 333: power spectrum averaging unit, 341-1: first spectral moment calculating unit, 341-2: second spectral moment calculating unit, 341-N: Nth spectral moment calculating unit, 342: spectral moment averaging unit
This application is a Continuation of PCT International Application No. PCT/JP2018/011072, filed on Mar. 20, 2018, which is hereby expressly incorporated by reference into the present application.
| Number | Date | Country | |
|---|---|---|---|
| Parent | PCT/JP2018/011072 | Mar 2018 | US |
| Child | 16998807 | US |
