JP7004618B2 - Frequency estimator and tracking receiver - Google Patents

Description

The present invention relates to a frequency estimator and a tracking receiver.

The tracking receiver estimates the frequency of the carrier wave of the input signal, performs acquisition and tracking processing, and removes excess noise by using a low-pass filter in order to reduce the amount of noise during directional calculation. It is necessary to improve the accuracy of frequency synchronization so that no residual carrier wave remains in the input signal, but the tracking receiver becomes large due to the addition of a processing circuit or the like. Further, even in a normal communication device, a device and a method for estimating and removing a residual carrier wave with a simple structure are desired.

The tracking receiver can estimate the frequency of the modulated signal using the discrete Fourier transform when a modulated signal with residual carriers is input. For example, the apparatus described in Patent Document 1 detects the power peak value of the spectral component and estimates the frequency of the residual carrier wave.

Japanese Unexamined Patent Publication No. 05-136631

However, since the apparatus described in Patent Document 1 detects the power peak value of the spectral component and estimates the reception frequency, it is modulated by the PCM (Pulse Code Modulation) / PM (Phase Modulation) method having a large degree of modulation. When the signal is received, the modulated signal component may be larger than the residual carrier wave component. As a result, the apparatus described in Patent Document 1 may erroneously detect the modulated signal component as a residual carrier wave component.

Therefore, it is an object of the present invention to provide a frequency estimation device and a tracking receiver capable of correctly estimating the frequency of a residual carrier wave.

The frequency estimator of the present invention has a discrete Fourier converter that generates a frequency spectrum that is a complex number for each frequency point by performing discrete Fourier conversion on orthogonal detection and digitized received signals, and a frequency point based on the amplitude value of the frequency spectrum. The peripheral signal level of the frequency point is calculated from the signal level calculation unit that calculates the signal level for each frequency point and the signal levels of a plurality of frequency points in a predetermined range in the vicinity of the frequency point for each frequency point. Peripheral signal level calculation unit, protrusion degree calculation unit that calculates the degree of protrusion indicating the degree of magnitude of the signal level with respect to the peripheral signal level at each frequency point, and the frequency of the carrier that remains in the received signal based on the protrusion degree. It is provided with a residual carrier frequency estimation unit for estimating.

According to the present invention, the frequency of the residual carrier can be estimated correctly.

It is a block diagram which shows the structure of the tracking receiver 1 by Embodiment 1. FIG. (A) is a diagram showing the configuration of the analog receiving unit 2S. (B) is a diagram showing the configuration of the analog receiving unit 2E. (A) is a diagram showing the configuration of the orthogonal detection unit 4S. (B) is a diagram showing the configuration of the orthogonal detection unit 4E. (A) is a figure showing the structure of the complex phase rotation part 11S. (B) is a figure showing the structure of the complex phase rotation part 11E. It is a figure which shows the structure of the direction calculation unit 12. It is a figure which shows the structure of a window function multiplier 5. (A) is a figure showing an example of a power spectrum. (B) is a diagram showing the peripheral signal level _PN for each frequency point. (C) is a diagram showing a plurality of frequency points used for calculating the peripheral signal level _PN at the frequency point A. It is a schematic diagram of the power spectrum of a PCM / PM signal. It is a schematic diagram of the power spectrum obtained by the discrete Fourier transform of the PCM / PM signal. It is a figure which shows the simulation result of the power spectrum of a PCM / PM signal. It is a schematic diagram of the protrusion degree PR for each frequency point calculated by Embodiment 1. FIG. It is a figure which shows the result of having calculated the protrusion degree PR from the simulation result of FIG. FIG. 12 is an enlarged view of a part of FIG. 12 in the direction of frequency f. It is a block diagram which shows the structure of the tracking receiver 101 by Embodiment 2. FIG. It is a figure which shows the structure of the frequency tracking part 9. It is a block diagram which shows the structure of the tracking receiver 201 by the modification 1 of Embodiment 2. It is a figure which shows the structure of the phase tracking part 209. (A) is a figure showing the structure of the complex phase rotation part 211S. (B) is a figure showing the structure of the complex phase rotation part 211E. It is a block diagram which shows the structure of the tracking receiver 301 by Embodiment 3. FIG. (A) is a figure showing a mapping point of a QPSK signal. (B) is a figure showing the mapping point of the QPSK signal to the 4th power. It is a figure which shows the simulation result which raised the QPSK signal to the fourth power. It is a figure which shows the power spectrum of a QPSK signal. It is a figure which shows the power spectrum of the QPSK signal which raised the 4th. It is a figure which shows the protrusion degree PR of the 4th power QPSK signal. It is a figure which shows the example of the hardware composition of the frequency estimation apparatus 10, 310 which concerns on Embodiments 1 to 3.

Hereinafter, embodiments will be described with reference to the drawings.
Embodiment 1.
FIG. 1 is a block diagram showing a configuration of the tracking receiver 1 according to the first embodiment. The tracking receiver 1 is mounted on a mobile body such as an aircraft, and obtains a directional error of an antenna mounted on the mobile body in order to communicate with a communication partner such as a communication satellite. The sum signal (normal signal) and the difference signal (signal corresponding to the directional error) generated by the antenna are input to the tracking receiver 1, and the tracking receiver 1 calculates and outputs the directional error. do. The directivity error is input to a drive device that changes the directivity direction of the antenna, and is controlled so that the antenna faces the direction of the communication partner.

The tracking receiver 1 uses an IF signal (Intermediate Frequency) as a sum signal and a difference signal. The IF signal is a signal whose frequency is converted from an RF signal (Radio Frequency) to a lower frequency. A BB signal (Base Band, a signal of information to be communicated) is extracted from the IF signal. As shown in FIG. 1, the tracking receiver 1 includes a complex sum signal generation unit 25S, a complex difference signal generation unit 25E, a frequency estimation device 10, complex phase rotation units 11S and 11E, and a direction calculation unit 12. Be prepared.

The complex sum signal generation unit 25S includes an analog receiving unit 2S, an ADC (Analog Digital Converter) 3S, and an orthogonal detection unit 4S. The complex difference signal generation unit 25E includes an analog receiving unit 2E, an ADC 3E, and an orthogonal detection unit 4E.

The analog receiving unit 2S executes reception processing of the sum signal SUM-IF of the intermediate frequency band generated by the antenna. FIG. 2A is a diagram showing the configuration of the analog receiving unit 2S. The analog receiving unit 2S includes a band-pass filter (BPF) 21S, an amplification unit (AMP) 22S, and a low-pass filter (LPF) 23S.

The BPF21S removes noise and spurious from the sum signal SUM-IF in the intermediate frequency band generated by the antenna. The AMP22S amplifies the level of the signal output from the BPF21S to an appropriate level. The LPF23S removes high frequency components from the signal output from the AMP22S so that aliasing noise does not occur in the subsequent ADC3S. The received and processed sum signal SUM1 in the intermediate frequency band is output from the LPF23S.

The analog receiving unit 2E executes reception processing of the difference signal ERR-IF in the intermediate frequency band generated by the antenna. FIG. 2B is a diagram showing the configuration of the analog receiving unit 2E. The analog receiving unit 2E includes a band-pass filter (BPF) 21E, an amplification unit (AMP) 22E, and a low-pass filter (LPF) 23E.

The BPF21E removes noise and spurious from the intermediate frequency band difference signal ERR-IF generated by the antenna. The AMP22E amplifies the level of the signal output from the BPF21E to an appropriate level. The LPF23E removes high frequency components from the signal output from the AMP22E so that aliasing noise does not occur in the subsequent ADC3E. The received and processed intermediate frequency band difference signal ERR1 is output from the LPF23E.

The ADC 3S converts the analog intermediate frequency band sum signal SUM1 output from the analog receiving unit 2S into the digital intermediate frequency band sum signal SUM2. The ADC 3E converts the analog intermediate frequency band difference signal ERR1 output from the analog receiving unit 2E into the digital intermediate frequency band difference signal ERR2.

The orthogonal detection unit 4S orthogonally detects the sum signal SUM2 of the digital intermediate frequency band output from the ADC 3S and converts it into a baseband complex sum signal (SUM3I + jSUM3Q). FIG. 3A is a diagram showing the configuration of the orthogonal detection unit 4S. The orthogonal detection unit 4S includes a numerically controlled oscillator (NCO) 41S, a cos / sin generator 42S, a multiplier 43S, 44S, and an LPF 45S, 46S.

The NCO41S outputs a value obtained by integrating the values specified for each sampling cycle. The cos / sin generator 42S generates a complex local signal (cosξ + jsinξ) having a phase (ξ) of the value output by the numerically controlled oscillator 41S. The multiplier 43S multiplies the sum signal SUM2 in the digital intermediate frequency band with cosξ. The multiplier 44S multiplies the sum signal SUM2 in the digital intermediate frequency band with sinξ. The LPF45S and 46S remove harmonics with respect to the multiplication result of the multipliers 43S and 44S, respectively. A baseband complex sum signal (SUM3I + jSUM3Q) is output from the LPF45S and 46S.

The orthogonal detection unit 4E orthogonally detects the digital intermediate frequency band difference signal ERR2 output from the ADC3E and converts it into a baseband complex difference signal (ERR3I + jERR3Q). FIG. 3B is a diagram showing the configuration of the orthogonal detection unit 4E. The orthogonal detection unit 4E includes a numerically controlled oscillator (NCO) 41E, a cos / sin generator 42E, multipliers 43E and 44E, and LPF45E and 46E.

The NCO41E outputs a value obtained by integrating the values specified for each sampling cycle. The cos / sin generator 42E generates a complex local signal (cosξ + jsinξ) having a phase (ξ) of the value output by the numerically controlled oscillator 41E. The multiplier 43E multiplies the difference signal ERR2 in the digital intermediate frequency band with cosξ. The multiplier 44E multiplies the difference signal ERR2 in the digital intermediate frequency band with sinξ. The LPF45E and 46E remove harmonics with respect to the multiplication result of the multipliers 43E and 44E, respectively. A base-on band complex difference signal (ERR3I + jERR3Q) is output from the LPF45E and 46E.

Let ΔF ₀ be the difference between the frequency of the complex local signal output from the cos / sin generation unit 42S in the orthogonal detection unit 4S and the frequency of the carrier wave of the sum signal SUM2 in the input digital intermediate frequency band. The complex sum signal (SUM3I + jSUM3Q) obtained by the orthogonal detection by the orthogonal detection unit 4S is expressed as {I (t) + jQ (t)} × exp (j2π × ΔF ₀ × t). I (t) + jQ (t) is the baseband signal to be demodulated. ΔF ₀ is a frequency deviation. This ΔF ₀ corresponds to the frequency of the carrier wave remaining in the complex sum signal (SUM3I + jSUM3Q) which is the received signal.

The frequency estimation device 10 estimates the frequency deviation, that is, the residual carrier frequency, which is the frequency of the carrier remaining in the complex sum signal (SUM3I + jSUM3Q), which is a received signal digitized and orthogonally detected. Let the estimated frequency be ΔF.

The complex phase rotation unit 11S is a frequency correction value (-) in which the sign of the baseband complex sum signal (SUM3I + jSUM3Q) output from the orthogonal detection unit 4S and the frequency estimation value ΔF of the residual carrier wave output from the frequency estimation device 10 are inverted. Based on the phase correction value (−Δθ) obtained from ΔF), the complex sum signal (SUM6I + jSUM6Q) obtained by removing the residual carrier wave from the complex sum signal (SUM3I + jSUM3Q) is output. By removing the residual carrier wave, the center of the frequency spectrum of the complex sum signal (SUM6I + jSUM6Q) becomes "0".

FIG. 4A is a diagram showing the configuration of the complex phase rotation unit 11S. The complex phase rotation unit 11S has a phase correction value calculation unit 231S and a phase rotation unit 232S.
The phase correction value calculation unit 231S includes a numerically controlled oscillator (NCO) 111S. The phase rotation unit 232S includes a cos / sin generation unit 112S, a complex multiplier 350S, and a low-pass filter (LPF) 115AS, 115BS. The complex multiplier 350S includes multipliers 113AS, 113BS, 113CS, 113DS and adders 114AS, 114BS.

The NCO111S receives a frequency correction value (−ΔF) obtained by inverting the sign of the frequency estimation value ΔF of the residual carrier from the frequency estimation device 10 as a frequency correction value, and integrates the frequency correction value (−ΔF) to obtain a phase correction value (−Δθ). ) Is output. The cos / sin generation unit 112S generates a complex local signal exp (−jΔθ) of the phase correction value (−Δθ) output by the NCO111S. The multipliers 113AS, 113BS, 113CS, and 113DS multiply the baseband complex sum signal (SUM3I + jSUM3Q) output from the orthogonal detection unit 4S with the complex local signal exp (−jΔθ). The adder 114AS adds the multiplication result of the multiplier 113AS and the value obtained by inverting the sign of the multiplication result of the multiplier 113BS. The adder 114BS adds the multiplication result of the multiplier 113CS and the multiplication result of the multiplier 113DS. The LPF115AS, 115BS removes extra harmonics and noise from the output of the adders 114AS, 114BS. A baseband complex sum signal (SUM6I + jSUM6Q) is output from LPF115AS and 115BS.

A method in which the NCO111S integrates the frequency correction value (−ΔF) to obtain the phase correction value (−Δθ) will be described. The input of the NCO111S is the ratio F / Fs of the frequency F to the sampling frequency Fs. Here, −0.5 ≦ F / Fs <0.5. According to the sampling theorem, a signal having only frequency components up to a frequency F of 1/2 or less of the sampling frequency Fs can be restored to the original signal even if it is sampled. The NCO111S can output a phase advanced by φ = (2π) × (F / Fs) for each sampling period (1 / Fs). Since −0.5 ≦ F / Fs <0.5, φ is −0.5 ≦ φ / (2π) <0.5. Therefore, if the frequency correction value (−ΔF) is input to the NCO111S, the phase correction value (−Δθ) obtained by integrating the frequency correction value (−ΔF) from the NCO111S is output.

The complex phase rotating unit 11E inverts the sign of the baseband complex difference signal (ERR3I + jERR3Q) output from the orthogonal detection unit 4E and the frequency estimation value ΔF of the residual carrier wave output from the frequency estimation device 10 (−ΔF). Based on the obtained phase correction value (−Δθ), the complex difference signal (ERR6I + jERR6Q) obtained by removing the residual carrier wave from the complex difference signal (ERR3I + jERR3Q) is output. By removing the residual carrier wave, the center of the frequency spectrum of the complex difference signal (ERR6I + jERR6Q) becomes "0".

FIG. 4B is a diagram showing the configuration of the complex phase rotation unit 11E. The complex phase rotation unit 11E has a phase correction value calculation unit 231E and a phase rotation unit 232E.
The phase correction value calculation unit 231E includes a numerically controlled oscillator (NCO) 111E. The phase rotation unit 232E has a cos / sin generation unit 112E, a complex multiplier 350E, and a low-pass filter (LPF) 115AE, 115BE. The complex multiplier 350E includes multipliers 113AE, 113BE, 113CE, 113DE and adders 114AE, 114BE.

The NCO111E receives a frequency correction value (-ΔF) obtained by inverting the sign of the frequency estimation value ΔF of the residual carrier from the frequency estimation device 10 as a frequency correction value, and integrates the frequency correction value (-ΔF) to obtain a phase correction value (-Δθ). ) Is output. The cos / sin generation unit 112E generates a complex local signal exp (−jΔθ) of the phase correction value (−Δθ) output by the NCO111E. The multipliers 113AE, 113BE, 113CE, and 113DE multiply the baseband complex difference signal (ERR3I + jERR3Q) output from the orthogonal detection unit 4E with the complex local signal exp (−jΔθ). The adder 114AE adds the multiplication result of the multiplier 113AE and the value obtained by inverting the sign of the multiplication result of the multiplier 113BE. The adder 114BE adds the multiplication result of the multiplier 113CE and the multiplication result of the multiplier 113DE. The LPF 115AE, 115BE removes extra harmonics and noise from the output of the adders 114AE, 114BE. A baseband complex difference signal (ERR6I + jERR6Q) is output from the LPF115AE and 115BE.

The azimuth calculation unit 12 determines the direction error (X + jY) based on the complex sum signal (SUM6I + jSUM6Q) output from the complex phase rotation unit 11S and the complex difference signal (ERR6I + jERR6Q) output from the complex phase rotation unit 11E. Perform azimuth calculation. The directivity error (X + jY) is the difference between the arrival direction in which the radio wave arrives and the directivity direction in which the antenna faces. FIG. 5 is a diagram showing the configuration of the direction calculation unit 12. The directional calculation unit 12 includes a complex division circuit 121, a low-pass filter (LPF) 122, 123, and a multiplier 124, 125.

The complex division circuit 121 divides the complex difference signal (ERR6I + jERR6Q) output from the complex phase rotation unit 11E by the complex sum signal (SUM6I + jSUM6Q) output from the complex phase rotation unit 11S, and divides the result (PX + jPY). Is output. The LPF 122 performs a smoothing process on the real part PX of the division result (PX + jPY). The multiplier 124 multiplies the signal output from the LPF 122 by the reciprocal of the normalization error sensitivity γθ [V / V / deg] (= 1 / γθ), and outputs the real part X of the directivity error. The LPF 123 performs a smoothing process on the imaginary portion PY of the division result (PX + jPY). The multiplier 125 multiplies the signal output from the LPF 123 by the reciprocal of the normalization error sensitivity γθ [V / V / deg] (= 1 / γθ), and outputs the imaginary portion Y of the directivity direction error.

Since the frequency spectra of the complex sum signal (SUM6I + jSUM6Q) and the complex difference signal (ERR6I + jERR6Q) are centered on "0" by the complex phase rotating portions 11S and 11E, they pass through the LPF 122 and 123 in a narrow band. It can be a filter. Since noise can be removed by the LPFs 122 and 123, the directivity error can be obtained accurately.

Again, with reference to FIG. 1, the configuration of the frequency estimation device 10 will be described. The frequency estimation device 10 includes a window function multiplier 5, a discrete Fourier transform unit 6, a power calculation unit 71, an LPF 72, a signal level calculation unit 73, a peripheral signal level calculation unit 74, a protrusion degree calculation unit 75, and a residual carrier frequency estimation unit. Has 8.

The window function multiplier 5 is connected to the subsequent stage of the orthogonal detection unit 4S and multiplies a window function such as a Hanning window. FIG. 6 is a diagram showing the configuration of the window function multiplier 5. The window function multiplier 5 has a window function signal generation unit 51 and multipliers 52 and 53. The window function signal generation unit 51 generates a window function signal such as a Hanning window or a Blackman window. The multipliers 52 and 53 multiply the complex sum signal (SUM3I + jSUM3Q) output from the orthogonal detection unit 4S and the window function signal output from the window function signal generation unit 51, and output the complex sum signal (SUM4I + jSUM4Q). do.

Explain the reason for using the window function. The discrete Fourier transform transforms from the time domain to the frequency domain on the premise that the waveform within the observation time is repeated. The frequency of the residual carrier component is unknown, and the frequency of the residual carrier is not always an integral multiple of 1 / T0 (Hz) when the observation time T0 used in the discrete Fourier transform is one cycle. When the frequency of the residual carrier wave is not an integral multiple of 1 / T0 (Hz), the frequency spectrum obtained by the discrete Fourier transform has many components other than the frequency of the residual carrier wave, and it becomes difficult to detect the frequency of the residual carrier wave. The window function is a function determined so that the value near the boundary of the observation time of the discrete Fourier transform becomes small. By multiplying the input signal by the window function, the frequency spectrum obtained by the discrete Fourier transform can reduce the spread of frequency components.

The discrete Fourier transform unit 6 performs a discrete Fourier transform on a complex sum signal (SUM4I + jSUM4Q) which is a digitized received signal to generate a frequency spectrum composed of a complex sum signal (SUM5I + jSUM5Q) for each frequency point.

More specifically, the discrete Fourier transform unit 6 converts the complex sum signal (SUM4I + jSUM4Q) on the time axis into the complex sum signal (SUM5I + jSUM5Q) on the frequency axis by the discrete Fourier transform process. The discrete Fourier transform unit 6 may be realized by using an algorithm of fast Fourier transform processing called FFT (Fast Fourier Transform).

The power calculation unit 71 obtains the power value of each frequency point by the sum of squares of the I component and the Q component from the discrete Fourier transform result (SUM5I + jSUM5Q) of each frequency point. Further, the power calculation unit 71 may obtain the amplitude value of each frequency point by performing a route calculation of the power value of each frequency point.

The LPF unit 72 smoothes the power value or the amplitude value of each frequency point in time series. The LPF unit 72 performs smoothing processing on the power value or amplitude value of each frequency point in time series using an IIR (Infinite Impulse Response) filter or the like. This makes it possible to reduce the dispersion of random signal components including noise.

The signal level calculation unit ₇₃ calculates the signal level PS for each frequency point based on the amplitude value of the frequency spectrum or the power value obtained by squared the amplitude value. For example, the signal level calculation unit 73 may use the power value or amplitude value of the frequency spectrum that has passed through the _LPF unit 72 as the signal level PS. FIG. 7A is a diagram showing an example of a power _spectrum representing a signal level PS for each frequency point. In the example of FIG. 7A, the power value of the power spectrum is defined as the signal level _PS .

For each frequency point, the peripheral signal level calculation unit 74 has the peripheral signal level of the frequency point from the signal level _PS of the peripheral frequency point set which is a plurality of frequency points in a predetermined range in the vicinity of the frequency point. Calculate _PN . FIG. 7B is a diagram showing the peripheral signal level _PN for each frequency point.

Here, the peripheral frequency point set is determined not to include a predetermined number of frequency points continuously arranged on a frequency axis including at least the frequency points of interest for which the peripheral signal level is calculated.

FIG. 7 (c) is a diagram showing a plurality of frequency points used for calculating the peripheral signal level _PN at the frequency point A. Note that FIG. 7 (c) enlarges the vicinity of the frequency point A in FIG. 7 (a). The peripheral signal level calculation unit 74 calculates the average value of the frequency points L1 to _L8 and the signal levels PS of R1 to R8, 8 points on each side from the second point not including the 1 point immediately adjacent to the frequency point A. It can be the peripheral signal level _PN of the point A. The frequency points L1 to L8 for calculating the average value and R1 to R8 are surrounded by a broken line in FIG. 7 (c).

The protrusion degree calculation unit 75 calculates the protrusion degree PR (= _PS / _PN ) obtained by dividing the signal level PS by the peripheral signal level _{PN at} each frequency point.

When the received signal contains a residual carrier wave, the signal level PS is high _at the frequency of the residual carrier wave, and the signal level _PN of the frequency around the frequency of the residual carrier wave is low. In order to utilize this feature, the protrusion degree PR obtained by dividing the signal level _PS by the peripheral signal level _PN is used. When calculating the peripheral signal level _PN , the frequency point and the total of 3 points immediately adjacent to it are excluded for the following reasons. The frequency width of the residual carrier is ideally 0 Hz. However, as described above, the frequency of the residual carrier wave is not always an integral multiple of 1 / T0 (Hz) when the observation time T0 used in the discrete Fourier transform is set to one cycle, and the window function. By multiplying by, the frequency of the residual carrier has some width in the frequency spectrum. Therefore, when calculating the peripheral signal level of each frequency point, the width of three points continuously arranged on the frequency axis including at least the frequency point is excluded. The frequency points to be excluded may be a fixed number of 5 points or more. The set of frequency points used when calculating the peripheral signal level is called the frequency peripheral point set.

The degree of protrusion PR is an index showing the degree of magnitude of the signal level _PS with respect to the peripheral signal level _PN . The degree of protrusion PR may be calculated by a method other than dividing the signal level _PS by the peripheral signal level _PN .

The residual carrier frequency estimation unit 8 estimates the frequency of the carrier remaining in the received signal based on the protrusion PR. The residual carrier frequency estimation unit 8 includes a frequency point determination unit 181 and a frequency calculation unit 182. The frequency point determination unit 181 detects the frequency point PO having the maximum protrusion degree PR calculated by the protrusion degree calculation unit 75. The frequency calculation unit 182 calculates the frequency estimation value ΔF of the residual carrier by multiplying the value of the frequency point PO having the maximum protrusion PR by the frequency interval between two adjacent frequency points of the discrete Fourier transform. .. When the sample frequency of the complex sum signal (SUM4I + jSUM4Q) input to the discrete Fourier transform unit 6 is 2.19 MHz and the number of samples of the discrete Fourier transform is 4096 samples, the frequency point PO is the residual carrier with respect to X (0 ≦ X ≦ 4095). The frequency estimation value ΔF of is expressed by the following equation.
ΔF = X * (2.19/4096) (0 ≤ X ≤ 2047)… (1)
ΔF = (X-4096) * {2.19/4096} (2048 ≤ X ≤ 4096)… (2)
Here, the relationship between X and the frequency is as follows.
X = 0 is the frequency point in the middle of the frequency axis X = 2047 is the rightmost frequency point on the frequency axis X = 2048 is the leftmost frequency point on the frequency axis,
X = 4095 is a frequency point one point to the left of the center on the frequency axis. As described above, the frequency estimation value ΔF of the residual carrier is calculated for each frequency point of the frequency spectrum calculated by the discrete Fourier transform, that is, the complex sum. It can be obtained for each 4096 sample of the signal (SUM4I + jSUM4Q).

In FIG. 1, the frequency calculation unit 182 outputs a frequency estimation value ΔF. In FIG. 1, the frequency estimate ΔF is not input to anything. During adjustment of the tracking receiver 1, a signal having a frequency deviation of ΔF can be input to the adjustment device, and the frequency estimation value ΔF can be confirmed on a screen or the like.

Next, the effect of the first embodiment when a signal modulated by the PCM / PM method having a large degree of modulation (hereinafter referred to as a PCM / PM signal) is received will be described.
FIG. 8 is a schematic diagram of the power spectrum of the PCM / PM signal. FIG. 8 shows the power spectrum of the PCM signal using the biphase PCM code. The power spectrum of the PCM / PM signal includes a central residual carrier component and other data signal components. The received signal received in actual operation also includes a noise component such as a white Gaussian noise component shown as a noise component.

FIG. 9 is a schematic diagram of a power spectrum obtained by performing a discrete Fourier transform on a PCM / PM signal. FIG. 10 is a diagram showing a simulation result of a power spectrum of a PCM / PM signal. FIG. 10 shows the result of applying 4096 points of FFT to the simulated signal waveform to which noise is added to the PCM / PM signal to which a large modulation factor of 1.5 rad is applied. Generally, in PCM / PM modulation, the power of the residual carrier component and the data signal component is uniquely determined with respect to the carrier level by using the modulation degree m (rad). Assuming that the power of all carriers is Pc (dB), the power of the residual carrier component is Pr (dB), and the power of the data signal component is Ps (dB), the following equation is established.
Pr = Pc + 10 * log ₁₀ (cos ² (m))… (3)
Ps = Pc + 10 * log ₁₀ (sin ² (m))… (4)
For example, when m = 1.5rad, Pr = Pc-23dB and Ps = Pc-0.02dB, and it can be seen that the component of the residual carrier wave is significantly smaller than the data component.

In FIG. 10, the power value of the power spectrum of the residual carrier component 81 is a value of about “-25” dB, but the power value of the power spectrum of the data signal component 82 is a value of about “-23” dB. The power value of the power spectrum of the data signal component 82 is larger than the power value of the power spectrum of the residual carrier wave component 81. Therefore, based on the magnitude of the power value in the power spectrum. The frequency of the residual carrier component 81 cannot be detected.

FIG. 11 is a schematic diagram of the protrusion degree PR for each frequency point calculated according to the first embodiment. As shown in FIG. 11, the protrusion PR is maximized at the frequency point including the residual carrier component. FIG. 12 is a diagram showing a result of calculating the protrusion degree PR from the simulation result of FIG. FIG. 13 is an enlarged view of a part of FIG. 12 in the direction of the frequency f. As shown in FIGS. 12 and 13, the protrusion PR of the residual carrier component 101 is a value of a little over 20 dB, but the protrusion PR of the data signal component 102 is less than 10 dB. By using the maximum value of the protrusion PR instead of the maximum value of the power value, the residual carrier component can be accurately detected.

Modification example of the first embodiment 1.
The signal level calculation unit 73 sets the power value or amplitude value of the frequency spectrum that has passed through the _LPF unit 72 as the signal level PS, and the peripheral signal level calculation unit 74 determines the vicinity of the frequency point for each frequency point. The peripheral signal level _PN of the frequency point is calculated from the signal levels of a plurality of frequency points in the above range, but the present invention is not limited to this.

For each frequency point, the peripheral signal level calculation unit 74 calculates the peripheral signal level _PN of the frequency point from the power value or amplitude value of the frequency spectrum of a plurality of frequency points in a predetermined range in the vicinity of the frequency point. It may be calculated. For each frequency point, the signal level calculation unit 73 may use a value obtained by subtracting the peripheral signal level _PN from the power value or amplitude value of the frequency _spectrum as the signal level PS of the frequency point.

Modification example of the first embodiment 2.
In the above embodiment, the peripheral signal level calculation unit 74 sets the average value of the signal level _PS of 8 points on each side, which is 2 points away from the frequency point, as the peripheral signal level _PN of the frequency point. It is not limited to this. The peripheral signal level calculation unit 74 may use the value obtained by further dividing the average value by the frequency interval between two adjacent frequency points of the discrete Fourier transform as the peripheral signal level _PN . This is because the residual carrier wave component corresponds to CW (Continuous Wave) and the level is irrelevant to the bandwidth, so that it is easier to grasp the reception status when considering PS / _{P N0} .

Modification example of the first embodiment 3.
In the above embodiment, the frequency calculation unit 182 sets the frequency corresponding to the value of the frequency point at which the protrusion PR is maximum as the frequency estimation value ΔF of the residual carrier wave, but the frequency is not limited to this. The frequency calculation unit 182 uses, for example, a frequency corresponding to the frequency point having the maximum protrusion degree PR and a frequency corresponding to the frequency points before and after the frequency point having the maximum protrusion degree PR, for example, in the primary or secondary order. The frequency estimation value ΔF of the residual carrier may be calculated by the interpolation process at a frequency higher than the frequency interval between two adjacent frequency points of the discrete Fourier transform.

Modification example of the first embodiment 4.
In the above embodiment, the complex sum signal generation unit 25S converts the sum signal SUM-IF in the intermediate frequency band generated by the antenna into a digital signal and then performs orthogonal detection, but the present invention is not limited thereto. That is, the complex sum signal generation unit 25S may perform orthogonal detection of the sum signal SUM-IF in the intermediate frequency band generated by the antenna and then convert it into a digital signal. The same applies to the complex difference signal generation unit 25E. That is, the complex sum signal generation unit 25S performs orthogonal detection of the sum signal SUM-IF in the intermediate frequency band generated by the antenna and then converts it into a digital signal, or converts it into a digital signal and then performs orthogonal detection to make it complex. Generate a sum signal. The complex difference signal generation unit 25E converts the difference signal ERR-IF in the intermediate frequency band generated by the antenna into a digital signal after orthogonal detection, or converts it into a digital signal and then performs orthogonal detection to perform a complex difference. Generate a signal.

Embodiment 2.
FIG. 14 is a block diagram showing the configuration of the tracking receiver 101 according to the second embodiment. The tracking receiver 101 of the second embodiment includes a frequency tracking unit 9 in addition to the components of the tracking receiver 1 of the first embodiment.

The frequency tracking unit 9 is a circuit that performs feedback control using a phase-locked loop (PLL) and controls the phase difference between the phase of the input signal and the phase reference value to be zero. FIG. 15 is a diagram showing the configuration of the frequency tracking unit 9. The frequency tracking unit 9 includes a thinning processing unit 91, a phase comparator 92, and a loop filter 93. FIG. 15 is a diagram when the received signal is modulated by the QPSK modulation method. In addition to QPSK, it can be similarly applied to a digital modulation method in which symbol points arranged on a complex plane have rotational symmetry N times (N is a natural number of 2 or more).
The frequency tracking unit 9 brings the residual carrier frequency close to zero based on the residual carrier frequency (ΔFinit) estimated by the frequency estimation device 10 and the phase of the complex sum signal (SUM6I + jSUM6Q) output by the complex phase rotating unit 11S. By integrating in this way, the frequency correction value (−ΔFadj) is output. As a result, the phase correction value (-Δθ) is generated in the complex phase rotation units 11S and 11S.

The thinning processing unit 91 reduces the sampling frequency of the baseband complex sum signal (SUM6I + jSUM6Q) output from the complex phase rotation unit 11S. The thinning processing unit 91 mainly uses a CIC (Cascaded Integrator Comb) filter or the like to remove high-frequency components so as not to generate aliasing noise, and when the thinning rate is K, the sampling frequency is set to (1 / K). Convert. If the amount of noise of the input complex sum signal (SUM6I + jSUM6Q) is small and the one-sided equivalent noise bandwidth BL of the PLL can be widened, the thinning process may not be performed.

The phase comparator 92 includes a calculation unit 381, a remainder calculator 382, and an adder 383. The calculation unit 381 calculates the phase θa, which is the argument (angle formed by the complex number with the real axis) of the complex sum signal (SUM6I + jSUM6Q). Specifically, the arithmetic unit 381 calculates the phase θa according to the following equation using the real part and the imaginary part of the complex sum signal (SUM6I + jSUM6Q).
θa = arctan (SUM6Q / SUM6I)… (4)

The remainder calculator 382 performs an operation (modulo operation at 2π / N) in which the addition result of the phase θa calculated by the calculation unit 381 and the phase correction value (-θadj) is divided by 2π / N to obtain the remainder. do. For example, when the received signal is a signal modulated by the QPSK modulation method (hereinafter referred to as QPSK signal), since N = 4, the mod operation is performed with π / 2 [rad]. The adder 383 further subtracts the mapping reference value. The reference value of the mapping is an angle value in order to make the phase error of the complex sum signal with respect to the reference mapping point determined according to the modulation method within the range centered on zero. In the case of QPSK, the reference value for mapping is π / 4 [rad], so π / 4 [rad] is subtracted. In FIG. 15, the modulo operation of π / 2 [rad] for the input signal is expressed as mod (π / 2).

Generally, in a digital modulation method having N rotational symmetries, the remainder calculator 382 may perform mod (2π / N) operations, and the adder 383 may subtract (−π / N). For example, in the case of BPSK, it becomes mod (π), and in the adder 383, (−π / 2) is subtracted.

The result of subtracting the reference value of the above mapping is output as the phase difference Δθdiff. The phase difference Δθdiff is a phase error of the complex sum signal (SUM6I + jSUM6Q). The phase comparator 92 is a phase error detector that detects the phase error of the complex sum signal (SUM6I + jSUM6Q).

The loop filter 93 detects that a frequency deviation ΔF exists in the frequency of the output signal based on the phase difference Δθdiff, and integrates the result of multiplying the phase difference Δθdiff by the loop gain to obtain the frequency deviation ΔF. Ask. The loop filter 93 includes a multiplier 83, a multiplier 84, an integrator 95, an adder 85, a multiplier 94, and a selective capture unit 96. The integrator 95 includes an adder 97 and a delay circuit 86.

The multiplier 83 multiplies the phase difference Δθdiff output from the phase comparator 92 by the loop gain G1 and outputs the result to the adder 85. The multiplier 84 multiplies the phase difference Δθdiff output from the phase comparator 92 by the loop gain G2, and outputs the result to the integrator 95. The output of the multiplier 84 is cumulatively added by the adder 97 and the delay circuit 86 in the integrator 95. The integrator 95 can sequentially accumulate information about the frequency deviation ΔF. The adder 85 obtains the frequency deviation ΔF by adding the output of the integrator 95 and the output of the multiplier 83. The multiplier 94 generates (-ΔF) by multiplying (-1) in order to invert the sign of the frequency deviation ΔF output from the adder 85. The multiplier 94 further multiplies (−ΔF) by K to match the sampling frequency of the complex phase rotating unit 11S, and outputs a frequency correction value (−ΔFadj).

The phase correction value calculation unit 231S of the complex phase rotation unit 11S inputs a frequency correction value (−ΔFadj) obtained by inverting the sign of the residual carrier frequency and multiplying by a coefficient. As described in the first embodiment, the phase correction value calculation unit 231S outputs the phase correction value (−Δθadj) by integrating the frequency correction value (−ΔFadj). The phase rotation unit 232S rotates the phase of the complex sum signal (SUM3I + jSUM3Q) by the phase correction value (−Δθadj). That is, the phase rotation unit 232S multiplies the complex local signal exp (−jθadj) and the complex sum signal (SUM3I + jSUM3Q). As a result, the phase of the complex sum signal (SUM3I + jSUM3Q) output from the orthogonal detection unit 4S is sequentially rotated by the phase correction value (-Δθadj). In the first embodiment, the phase correction value is expressed as (−Δθ), but in the second embodiment, the phase correction value is expressed as (−Δθadj). As a result, it is possible to output a complex sum signal (SUM6I + jSUM6Q) in which the phase difference due to the frequency deviation ΔF generated in the orthogonal detection unit 4S is removed. Further, the feedback process is realized so that the phase difference Δθdiff approaches 0.

The phase correction value calculation unit 231E of the complex phase rotation unit 11E inputs a frequency correction value (−ΔFadj) obtained by inverting the sign of the residual carrier frequency and multiplying by a coefficient. As described in the first embodiment, the phase correction value calculation unit 231E outputs the phase correction value (−Δθadj) by integrating the frequency correction value (−ΔFadj). The phase rotation unit 232E rotates the phase of the complex difference signal (ERR3I + jERR3Q) by the phase correction value (−Δθadj). That is, the phase rotation unit 232E multiplies the complex local signal exp (−jθadj) and the complex difference signal (ERR3I + jERR3Q). As a result, the phase of the complex difference signal (ERR3I + jERR3Q) output from the orthogonal detection unit 4E is sequentially rotated by the phase correction value (-Δθadj). As a result, it is possible to output a complex difference signal (ERR6I + jERR6Q) in which the phase difference due to the frequency deviation ΔF generated in the orthogonal detection unit 4S is removed.

A loop filter 93 including a phase comparator 92 in the frequency tracking unit 9, an integrator 95 in the frequency tracking unit 9, an integrator NCO111S in the complex phase rotating unit 11S, and a complex in the complex phase rotating unit 11S. The integrator 350S constitutes a PLL whose transfer function is quadratic. The phase of the complex sum signal (SUM3I + jSUM3Q) output from the orthogonal detection unit 4S is corrected by the phase correction value (-Δθadj) obtained by integrating the frequency correction value (-ΔFadj) output from the loop filter 93 by the second-order PLL. Therefore, the phase difference Δθdiff in the phase comparator 92 can be brought close to 0. Further, the frequency correction value (−ΔFadj) output from the loop filter 93 is brought closer to the frequency deviation ΔF of the complex sum signal (SUM3I + jSUM3Q) output from the orthogonal detection unit 4S by the second-order PLL. (-ΔFadj) can be adjusted.

It is possible to follow the frequency deviation ΔF accumulated in the integrator 95 with respect to the further frequency deviation of about the one-sided equivalent noise bandwidth BL determined by the loop gains G1 and G2 in the loop filter 93.

A feedback loop is formed by the complex phase rotating unit 11S and the frequency tracking unit 9. The loop gains G1 and G2 are set small so that the feedback loop does not become unstable. In order to quickly bring the frequency deviation close to zero, the selective acquisition unit 96 outputs the residual carrier frequency (frequency deviation) estimated value at the specified time point output from the residual carrier frequency estimation unit 8 as the input of the integrator 95. ΔF is taken in as an initial frequency estimation value ΔFinit and output to the delay circuit 86.

Specifically, until the frequency deviation ΔFinit output from the residual carrier frequency estimation unit 8 stabilizes, the output of the loop filter 93 is set to “0” so that the input of the NCO 1111S in the complex phase rotation unit 11S is input. It is fixed to "0". When it is determined that the frequency deviation ΔFinit output from the residual carrier frequency estimation unit 8 is stable, it is instructed to take in the frequency deviation ΔFinit, and the selective acquisition unit 96 is instructed to take in the latest frequency from the residual carrier frequency estimation unit 8. The deviation ΔFinit is acquired and output to the delay circuit 86. When the selection capture unit 96 is not instructed to capture the frequency deviation ΔFinit, the selective capture unit 96 captures the output of the adder 97 and outputs it to the delay circuit 86. As a result, ΔFinit is set as the initial value of the frequency correction value (-ΔFadj), and the frequency correction value (-) is based on the phase of the complex sum signal (SUM6I + jSUM6Q) after the phase is rotated by the complex phase rotation unit 11S. ΔFadj) is adjusted.

By doing so, the frequency tracking unit 9 can follow the fluctuation of the frequency deviation ΔF with ΔFinit as the initial value. That is, it becomes possible to calculate the frequency deviation ΔF more accurately and in sample processing units than the ΔFinit estimated on the frequency axis by the discrete Fourier transform. However, when the PLL is unlocked, the output of the loop filter 93 is set to "0" until ΔFinit is taken in again. If it is desired to maintain the state before the unlocking, the value of the integrator 95 immediately before the unlocking may be maintained instead of "0".

For example, when the sampling frequency of the data input to the discrete Fourier transform unit 6 is 2.19 MHz, the thinning rate K is 256, and the data length (number of samples) input to the discrete Fourier transform unit 6 is 4096 points, the discrete Fourier transform unit 6 is used. The frequency interval between adjacent frequency points of the frequency spectrum output from the transform unit 6 has a resolution of 534 Hz. Even if the accuracy of 1/10 is obtained by performing the interpolation processing, it is 53 Hz. The thinning processing unit 91 in the frequency tracking unit 9 has a sampling frequency of 2.19 MHz, and the sampling frequency is set to 8. Reduce to 545 kHz. When the input of the complex phase rotating unit 11S is 32 bits and the decimal point position is the 23rd bit (Q23 format), the decimal point position is equivalent to 8.545 kHz, so 0.001 Hz (= 8.545 kHz / 2 ²³ ). It is possible to calculate the frequency deviation ΔF with accurate frequency resolution.

In the first embodiment and the second embodiment, a tracking receiver that estimates and removes the frequency of the residual carrier wave has been described. Even in a normal communication device, if a frequency offset or frequency mismatch exists between the transceivers, the communication quality such as an increase in demodulation errors when demodulating the received signal deteriorates. The technique of accurately estimating and removing the frequency of the residual carrier of the first embodiment and the second embodiment can also be applied to a normal communication device.

Modification example of the second embodiment 1.
FIG. 16 is a block diagram showing the configuration of the tracking receiver 201 according to the first modification of the second embodiment. The tracking receiver 201 of the second embodiment includes a phase tracking unit 209 in addition to the components of the tracking receiver 1 of the first embodiment. The tracking receiver 201 of the second embodiment includes the complex phase rotation units 211S and 211E different from the complex phase rotation units 11S and 11E of the first embodiment.

FIG. 17 is a diagram showing the configuration of the phase tracking unit 209. The phase tracking unit 209 differs from the frequency tracking unit 9 of the second embodiment in that it includes an integrator 99, an input to the thinning processing unit 91, a configuration of the phase comparator 292, and an output of the phase tracking unit 209. Is the difference.
The phase tracking unit 209 brings the residual carrier frequency close to zero based on the residual carrier frequency (ΔFinit) estimated by the frequency estimation device 10 and the phase of the complex sum signal (SUM6I + jSUM6Q) output by the complex phase rotating unit 11S. The phase correction value (−Δθadj) is output as described above.

The thinning processing unit 91 reduces the sampling frequency of the baseband complex sum signal (SUM3I + jSUM3Q) output from the orthogonal detection unit 4S. The thinning-out processing unit 91 mainly uses a CIC (Cascaded Integrator Comb) filter or the like to convert the sampling frequency to (1 / K) times while removing high-frequency components so as not to generate aliasing noise.

The integrator 99 is a second integrator that integrates a value based on the output of the integrator 95. The integrator 99 includes an adder 87, a delay circuit 88, and a delay circuit 89. The adder 87 and the delay circuit 88 cumulatively add the frequency correction value (−ΔFadj) which is the output of the multiplier 94. The output of the delay circuit 89 is sent to the complex phase rotation units 211S and 211E as a phase correction value (-Δθadj).

The phase comparator 292 has an adder 384 that adds a phase correction value (-Δθadj) to the phase θa output by the calculation unit 381. The output of the adder 384 is input to the remainder calculator 382. The phase comparator 292 is a phase error detector that detects the phase error of the complex sum signal (SUM6I + jSUM6Q) corrected by the phase correction value (-Δθadj).

When the selective acquisition unit 96 is instructed to acquire the frequency deviation ΔFinit, the selective acquisition unit 96 acquires the latest frequency deviation ΔFinit from the residual carrier frequency estimation unit 8 and outputs the latest frequency deviation ΔFinit to the delay circuit 86. When the selection capture unit 96 is not instructed to capture the frequency deviation ΔFinit, the selective capture unit 96 captures the output of the adder 97 and outputs it to the delay circuit 86. As a result, ΔFinit is set as the initial value of the frequency correction value (-ΔFadj), and the frequency correction value (-) is based on the phase of the complex sum signal (SUM3I + jSUM3Q) before the phase is rotated by the complex phase rotation unit 211S. ΔFadj) is adjusted.

FIG. 18A is a diagram showing the configuration of the complex phase rotation unit 211S. The difference between the complex phase rotation unit 211S and the complex phase rotation unit 11S in FIG. 4A is as follows. The complex phase rotation unit 211S does not include the phase correction value calculation unit 231S composed of the NCO 111S, but includes delay circuits 116AS and 116BS.

The con / sin generation unit 112S generates a complex local signal exp (-jΔθadj) of the phase correction value (-Δθadj) sent from the phase tracking unit 209, not from the NCO111S. The delay circuits 116AS and 116BS delay the complex sum signal (SUM3I + jSUM3Q) output from the orthogonal detection unit 4S and output it to the adders 113AS, 113BS, 113CS and 113DS. The reason for the delay is that the timing of the complex multiplication in the complex phase rotating unit 211S and the timing of the phase comparison of the phase comparator 292 of the phase tracking unit 209 are the same. When the timing is not adjusted, even if the phase difference Δθdiff and the frequency difference (difference between ΔF and ΔFadj) can be removed inside the secondary PLL in the phase tracking unit 209, the phase difference occurs in the complex phase rotating unit 211S. However, there is a problem that the phase of the constellation is out of phase.

FIG. 18B is a diagram showing the configuration of the complex phase rotation unit 211E. The difference between the complex phase rotation unit 211E and the complex phase rotation unit 11E in FIG. 4B is as follows. The complex phase rotation unit 211E does not include the phase correction value calculation unit 231E composed of the NCO 111E, but includes delay circuits 116AE and 116BE. The con / sin generation unit 112E generates a complex local signal exp (-jΔθadj) of the phase correction value (-Δθadj) sent from the phase tracking unit 209, not from the NCO111E. The delay circuits 116AE and 116BE delay the complex difference signal (ERR3I + jERR3Q) output from the orthogonal detection unit 4E and output it to the adders 113AE, 113BE, 113CE and 113DE.

The second-order PLL is configured by the phase comparator 292 in the phase tracking unit 209, the loop filter 93 including the integrator 95 in the phase tracking unit 209, and the integrator 99 in the phase tracking unit 209. The complex sum signal (SUM3I + jSUM3Q) output from the orthogonal detection unit 4S by the phase correction value (-Δθadj) obtained by integrating the frequency correction value (-ΔFadj) output from the loop filter 93 by the PLL whose transfer function is quadratic. The phase of is corrected, and the phase difference Δθdiff in the phase comparator 292 can be brought close to 0. Further, the sum of the frequency deviation ΔF of the complex sum signal (SUM3I + jSUM3Q) output from the orthogonal detection unit 4S and the frequency correction value (−ΔFadj) output from the loop filter 93 is brought close to zero by the second-order PLL. In addition, the frequency correction value (-ΔFadj) can be adjusted.

Modification example of the second embodiment 2.
In the above embodiment, the output of the multiplier 94 that receives the output of the adder 85 is sent to the complex phase rotating units 11S and 11E as a frequency correction value (−ΔFadj), but the present invention is not limited to this. .. In a configuration in which the loop filter 93 does not have the multiplier 83 and the adder 85, the output of the integrator 95 is inverted in sign by the multiplier 94 and sent to the complex phase rotating units 11S and 11E as a frequency correction value (−ΔFadj). It may be done.

Embodiment 3.
FIG. 19 is a block diagram showing a configuration of the tracking receiver 301 according to the third embodiment. In the frequency estimation device 310 of the tracking receiver 301 of the third embodiment, in addition to the components of the frequency estimation device 10 of the tracking receiver 1 of the first embodiment, the N-th power detection unit 13 and the frequency 1 / N times unit 14 are added. Equipped with.

The N-th power detection unit 13 raises the complex sum signal (SUM3I + jSUM3Q) output from the orthogonal detection unit 4S to the Nth power. Where N is a positive integer. The purpose of N-th power detection is to multiply the phase of the input signal by N. The N-th power detection unit 13 may calculate the phase value θ of the complex sum signal (SIM3I + jSUMQ) and multiply the phase value θ by 4, instead of the N-th power operation. This is because the phase is multiplied by N by the Nth power operation.
The signal (I + jQ) N obtained by multiplying the complex signal (I + jQ) by ^N can be transformed as follows. Here, it is assumed that the complex signal (I + jQ) is expressed as R × exp (jθ) in polar coordinate display.
(I + jQ) ^N = ( ^R × exp (jθ)) ^N = (RN × exp (jN × θ))
The mapping points of the NPSK signal are arranged at 2π / N (rad) intervals. For example, in the case of a QPSK signal, N = 4. In the case of a BPSK signal, N = 2.

FIG. 20A is a diagram showing mapping points of the QPSK signal. For example, the mapping points of the QPSK signal are arranged at θ1, θ2, θ3, and θ4. Here, θ1 = π / 4, θ2 = 3π / 4, θ3 = -3π / 4, and θ4 = −π / 4. When there is a phase difference Δθ due to the residual carrier, each mapping point is shifted by Δθ.

FIG. 20B is a diagram showing mapping points of the squared QPSK signal. The phase of the squared QPSK signal is 4 × θ1, 4 × θ2, 4 × θ3, and 4 × θ4. Here, let mod (X, Y) be a function for obtaining the remainder obtained by dividing X by Y. It will be as follows.
mod (4 × θ1, 2π) = π,
mod (4 × θ2, 2π) = mod (3π, 2π) = π,
mod (4 × θ3, 2π) = mod (-3π, 2π) = π,
mod (4 × θ4, 2π) ＝ mod (－π, 2π) ＝ π
When there is a phase difference Δθ due to the residual carrier, each mapping point is shifted by 4 × Δθ. As shown above, when the QPSK signal is squared, it can be seen that the mapping points are concentrated at one point of π. That is, the data near the Nyquist point can be shown by one point. The component appears as a line spectral component on the frequency axis. When the frequency of the residual carrier wave is ΔF, the points gathered at this one place rotate at a speed of 4 × ΔF / Fs per sample. Fs is the sampling frequency of the complex sum signal (SUM3I + jSUM3Q) input to the Nth power detection unit 13.

FIG. 21 is a diagram showing a simulation result obtained by squaring the QPSK signal to the fourth power. FIG. 21 shows the constellation when the QPSK signal is squared under the condition of ΔF / Rs = 2/4096.

The frequency of the line spectrum component is a value obtained by multiplying the frequency ΔF ₀ of the original residual carrier by N. In order to obtain the frequency ΔF ₀ of the original residual carrier wave, the residual carrier is provided with a frequency 1 / N times portion 14 in the frequency estimation device 8. The frequency 1 / N times unit 14 multiplies the frequency point having the maximum protrusion degree PR determined by the frequency point determination unit 181 by 1 / N, and sends the result to the frequency calculation unit 182.

FIG. 22 is a diagram showing a power spectrum of a QPSK signal. In FIG. 22, the horizontal axis is the normalized frequency obtained by dividing the frequency by the symbol rate Rs. In FIG. 22, the frequency estimation value ΔF of the residual carrier wave is 0.25 × Rs. FIG. 23 is a diagram showing a power spectrum of a QPSK signal squared. In FIG. 23, the horizontal axis is the normalized frequency. The line spectrum 2001 is generated at the position where the normalized frequency is “1”. That is, the line spectrum 2001 is generated at a frequency of 1 × Rs. From FIG. 23, it can be detected that the frequency at which the power is maximized is 1 × Rs. The frequency at which the power is maximized (= 1 × Rs) is four times the frequency estimation value ΔF (= 0.25 × Rs) of the residual carrier wave. The reason why it is quadrupled is that the QPSK signal is squared. The frequency of the residual carrier can be specified by multiplying the frequency at which the power is maximum by 1/4.

FIG. 24 is a diagram showing the protrusion degree PR of the QPSK signal squared. In order to calculate the protrusion degree PR, the peripheral signal level _PN is set to the power average value for 8 points (16 points on both sides) with the position 3 points away from each frequency point as the base point. As shown in FIG. 24, as in FIG. 23, the line spectrum 2101 is generated at the position where the normalized frequency is “1”. That is, the line spectrum 2101 is generated at a frequency of 1 × Rs. From FIG. 24, it can be detected that the frequency at which the protrusion degree PR is maximum is 1 × Rs. The frequency at which the protrusion PR is maximized (= 1 × Rs) is four times the frequency estimation value ΔF (= 0.25 × Rs) of the residual carrier wave. The reason why it is quadrupled is that the QPSK signal is squared. Similar to the power example of FIG. 23, the frequency of the residual carrier can be specified by multiplying the frequency at which the protrusion PR is maximum by 1/4. Therefore, when the QPSK signal is raised to the Nth power, the frequency of the residual carrier wave can be detected by the protrusion degree PR as well as the frequency of the residual carrier wave can be detected by the power value.

Modification example of the third embodiment 1.
If the symbol point arranged on the complex plane is a digital modulation method having N rotation symmetries for a natural number N of 2 or more, the N-th power detection unit 13 and the frequency 1 / N times unit 14 can be applied. Therefore, not only the N-PSK (N-Phase Shift Keying) signal, but also the QAM (Quadrature Amplitude Modulation) signal or the N-APSK (N-Amplitude Phase Shift Keying) signal, the N power detection unit 13 and the frequency 1 / The N times portion 14 can be applied.

Modification example of the third embodiment 2.
In the above embodiment, the frequency 1 / N times unit 14 multiplies the frequency point having the maximum protrusion degree PR determined by the frequency point determination unit 181 by 1 / N, and sends the result to the frequency calculation unit 182. However, it is not limited to this. The frequency 1 / N times unit 14 may receive the frequency estimation value ΔF of the residual carrier wave calculated by the frequency calculation unit 182 and multiply the frequency estimation value ΔF by 1 / N.

Hardware configuration.
FIG. 25 is a diagram showing an example of the hardware configuration of the frequency estimation devices 10 and 310 according to the first to third embodiments. As shown in FIG. 25, the hardware of the frequency estimation devices 10 and 310 includes a processor 1100 and a memory 1200 connected to the processor 1100 by a bus 1300.

The functions of at least some of the components of the frequency estimation devices 10 and 310 are realized by the processor 1110 such as a CPU executing a program stored in the memory 1200. Further, a plurality of processors and a plurality of memories may be linked to execute the functions of the above components. Further, the functions of the above components may be realized by a processing circuit such as a system LSI. Further, a plurality of processing circuits may cooperate to execute the functions of the above components.

It should be considered that the embodiments disclosed this time are exemplary in all respects and not restrictive. The scope of the present disclosure is shown by the scope of claims rather than the above description, and is intended to include all modifications within the meaning and scope equivalent to the scope of claims.

1,101,201,301 Tracking receiver, 2S, 2E analog receiver, 3S, 3E ADC, 4S, 4E orthogonal detector, 5 window function multiplier, 6 discrete Fourier converter, 8 residual frequency estimator, 9 frequency Tracking unit, 10,310 frequency estimator, 11S, 11E, 211S, 211E complex phase rotation unit, 12 azimuth calculation unit, 13N power detection unit, 14 frequency 1 / N times unit, 21S, 21E band pass filter, 22S, 22E AMP, 23S, 23E, 45S, 45E, 46S, 46E, 72, 115AS, 115AE, 115BS, 115BE, 122,123 Low frequency pass filter, 25S complex sum signal generator, 25E complex difference signal generator, 41S, 41E , 111S, 111E Numerical control oscillator, 42S, 42E, 112S, 112E cos / sin generator, 43S, 43E, 44S, 44E, 52, 53, 83, 84, 94, 113AS, 113AE, 113BS, 113BE, 113CS, 113CE , 113DS, 113DE, 124,125 Multiplier, 51 Window function signal generator, 71 Power calculation unit, 73 Signal level calculation unit, 74 Peripheral signal level calculation unit, 75 Protrusion degree calculation unit, 85, 87, 97, 114AS, 114AE, 114BS, 114BE, 383,384 adder, 88,89,116AS, 116AE, 116BS, 116BE delay circuit, 91 thinning processing unit, 92,292 phase comparator (phase error detector), 93 loop filter, 95, 99 Integrator, 96 Selective capture unit, 121 Complex division circuit, 181 Frequency point determination unit, 182 frequency calculation unit, 209 Phase follower unit, 231S, 231E Phase correction value calculation unit, 232S, 232E Phase rotation unit, 350S, 350E Complex multiplier, 381 arithmetic unit, 382 remainder calculator, 1100 processor, 1200 memory, 1300 bus.

Claims (13)

A discrete Fourier transform unit that performs discrete Fourier transform on orthogonal detection and digitized received signals to generate a frequency spectrum that is a complex number for each frequency point.
A signal level calculation unit that calculates the signal level for each frequency point based on the amplitude value of the frequency spectrum, and
Peripheral signal for calculating the peripheral signal level of the frequency point from the signal level of the peripheral frequency point set which is a plurality of frequency points in a predetermined range in the vicinity of the frequency point for each said frequency point. Level calculation unit and
A protrusion degree calculation unit that calculates a protrusion degree indicating the degree of magnitude of the signal level with respect to the peripheral signal level at each frequency point, and a protrusion degree calculation unit.
A frequency estimation device including a residual carrier frequency estimation unit that estimates the frequency of the carrier remaining in the received signal based on the protrusion. Claimed, the peripheral frequency point set is determined not to include a fixed number of frequency points that are continuously aligned on a frequency axis that includes at least the frequency points of interest for which the peripheral signal level is to be calculated. The frequency estimation device according to 1. The peripheral signal level calculation unit according to claim 1 or 2, wherein the peripheral signal level calculation unit calculates an average value of the signal levels of the frequency points included in the peripheral frequency point set, and sets the average value as the peripheral signal level. Frequency estimator. The frequency estimation according to any one of claims 1 to 3, wherein the signal level calculation unit sets the amplitude value of the frequency spectrum or the power value obtained by squaring the amplitude value as the signal level for each frequency point. Device. The frequency estimation according to any one of claims 1 to 4, wherein the residual carrier frequency estimation unit estimates the frequency of the carrier remaining in the received signal based on the frequency point at which the protrusion is maximum. Device. The received signal is modulated by a digital modulation method in which symbol points arranged on the complex plane have rotational symmetry N times (N is a natural number of 2 or more).
The discrete Fourier transform unit performs discrete Fourier transform on a signal obtained by digitizing and orthogonally detecting the received signal and multiplying the phase of the complex digital signal by N.
The frequency estimation device according to claim 5, wherein the residual carrier frequency estimation unit estimates a frequency of 1 / N times the frequency estimated based on the protrusion as the frequency of the carrier remaining in the received signal. A complex sum signal generator that generates a complex sum signal by orthogonal detection and digitization of an analog sum signal output by an antenna that receives radio waves from a radio source.
A complex difference signal generation unit that generates a complex difference signal by orthogonal detection and digitization of the analog difference signal output by the antenna.
The frequency estimation device according to any one of claims 1 to 6, wherein the residual carrier frequency, which is the frequency of the carrier remaining in the complex sum signal, is estimated.
A first complex phase rotation unit that rotates the phase of the complex sum signal according to the phase correction value based on the residual carrier frequency, and
A second complex phase rotation unit that rotates the phase of the complex difference signal according to the phase correction value, and
Based on the complex difference signal output by the second complex phase rotation unit and the complex sum signal output by the first complex phase rotation unit, the arrival direction in which the radio wave arrives and the direction in which the antenna faces. A tracking receiver equipped with a directional error calculation unit that calculates a directional error, which is the difference from the directional direction. Integrating the residual carrier frequency so as to approach zero based on the residual carrier frequency estimated by the frequency estimator and the phase of the complex sum signal output by the first complex phase rotating unit. The tracking receiver according to claim 7, further comprising a frequency tracking unit that outputs a frequency correction value at which the phase correction value is generated. A frequency correction value obtained by inverting the sign of the residual carrier frequency and multiplying by a coefficient is input to the first complex phase rotation unit and the second complex phase rotation unit, and the frequency correction value is integrated. The tracking receiver according to claim 8, further comprising a phase correction value calculation unit that outputs a phase correction value, and a phase rotation unit that rotates the phase by the phase correction value. The frequency tracking unit is designated as an input of the integrator, a phase error detector that detects the phase error of the complex sum signal, an integrator that integrates a value obtained by multiplying the output of the phase error detector by a coefficient, and an integrator. It has a selective capture unit that selects and captures the residual carrier frequency at the time point.
The tracking receiver according to claim 8 or 9, wherein the frequency tracking unit calculates the frequency correction value based on the output of the integrator. Based on the residual carrier frequency estimated by the frequency estimator and the phase of the complex sum signal input to the first complex phase rotating unit, the phase is set so as to bring the residual carrier frequency close to zero. The tracking receiver according to claim 7, further comprising a phase tracking unit that outputs a correction value. The first complex phase rotation unit and the second complex phase rotation unit have a phase rotation unit that rotates the phase by the input phase correction value, and the first complex phase rotation unit is input. The tracking receiver according to claim 11, further comprising a delay circuit for delaying the complex sum signal, and the second complex phase rotating unit having a delay circuit for delaying the input complex difference signal. The phase tracking unit includes a phase error detector that detects the phase error of the complex sum signal corrected by the phase correction value, an integrator that integrates the value obtained by multiplying the phase error by a coefficient, and the integrator. It has a selective acquisition unit that selects and captures the residual carrier frequency at the time specified as an input, and a second integrator that integrates a value based on the output of the integrator, and the phase correction value is the second. The tracking receiver according to claim 11 or 12, which is calculated based on the output of the integrator.

Images