Calibration device, conversion device, calibration method, and calibration program

Abstract

SOLUTION: A calibration device is provided that comprises: a calibration signal supply unit obtaining each of a plurality of bandpass signals by multiplying an input signal with each of a plurality of signal patterns to limit a band region, and supplying, as a calibration input signal, a multitone signal having a tone in a plurality of frequency band regions to a converter reconstructing an output signal in accordance with an input signal from the plurality of bandpass signals; a calibration bandpass signal acquisition unit acquiring a plurality of calibration bandpass signals to be obtained by the converter in accordance with the multitone signal; and a calibration processing unit calibrating a reconstructed parameter in the converter on the basis of the plurality of calibration bandpass signals.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a calibration device, a conversion device, a calibration method, and a calibration program. [Background technology]

[0002] Compressed sensing is a technique for sampling narrowband signals sparsely present in a wide frequency band. A modulated wideband converter (MWC) is known as a converter that performs compressed sensing (Non-Patent Documents 1 and 2). MWC reconstructs the original input signal using multiple samples obtained by multiplying a wideband input signal, in which narrowband signals are sparsely present, by multiple periodic sign functions (PSFs). Non-Patent Documents 3 to 8 disclose techniques for calibrating such MWCs. Non-Patent Documents 10 to 12 demonstrate that the signal reconstruction ability of MWCs deteriorates when a non-ideal low-pass filter is used. Non-Patent Documents 13 to 17 disclose techniques for correcting the frequency characteristics of low-pass filters. [Prior art document] [Non-patent literature] [Non-Patent Document 1] Mishali and YC Eldar, "From theory to practice: Sub-nyquist sampling of sparse wideband analog signals", IEEE Journal of Selected Topics in Signal Processing, vol. 4, no. 2, pp. 375-391, April 2010 [Non-Patent Document 2] DL Donoho, "Compressed sensing," in IEEE Transactions on Information Theory, vol. 52, no. 4, pp. 1289-1306, April 2006 [Non-Patent Document 3] Peng Wang, Fei You, and Songbai He, 「An Improved Signal Reconstruction of Modulated Wideband Converter Using a Sensing Matrix Built upon Synchronized Modulated Signals」, Circuits Syst. Signal Process. 38, 7, July 2019 [Non-Patent Document 4] J. Park, J. Jang and H. Lee, 「A calibration for the modulated wideband converter using sinusoids with unknown phases」, 2017 Ninth International Conference on Ubiquitous and Future Networks(ICUFN), Milan, 2017, pp. 951-955 [Non-Patent Document 5] L. Chen, J. Jin and Y. Gu, 「A calibration system and perturbation analysis for the Modulated Wideband Converter」, IEEE 10th INTERNATIONAL CONFERENCE ON SIGNAL PROCESSING PROCEEDINGS, Beijing, 2010, pp. 78-81 [Non-Patent Document 6] Liu, Weisong et al., 「Design of a Single Channel Modulated Wideband Converter for Wideband Spectrum Sensing: Theory, Architecture and Hardware Implementation」, Sensors(Basel, Switzerland) vol. 17, 5 1035.4 May. 2017 [Non-Patent Document 7] E. Israeli et al., "Hardware calibration of the modulated wideband converter," 2014 IEEE Global Communications Conference, Austin, TX, 2014, pp. 948-953 [Non-patent document 8] N. Fu, S. Jiang, L. Deng and L. Qiao, “Successive-phase correction calibration method for modulated wideband converter system”, in IET Signal Processing, vol. 13, no. 6, pp. 624-632, 8 2019 [Non-Patent Document 9] S. Boyd, "Multitone signals with low crest factor," IEEE Trans. Circuits Syst., vol. CAS-33, no. 10, pp. 1018-1022, October 1986 [Non-patent document 10] J. Zhang, N. Fu, W. Yu, and X. Peng, "Analysis of non-idealities of low-pass filter in random demodulator", Proc. SPIE, vol. 8759, 01 2013 [Non-Patent Document 11] PJ Pankiewicz, T. Arildsen, and T. Larsen, “Sensitivity of the random demodulation framework to filter tolerances,” in 2011 19th European Signal Processing Conference, 2011, pp. 534-538 [Non-Patent Document 12] S. Smaili and Y. Massoud, "Accurate and efficient modeling of random demodulation based compressive sensing systems with a general filter", in 2014 IEEE International Symposium on Circuits and Systems (ISCAS), 2014, pp. 2519-2522 [Non-Patent Document 13] Zhao Yijiu, Long Ling, Zhuang Xiaoyan, and Dai Zhijian, "Model calibration for compressive sampling system with non-ideal lowpass filter", in 2015 12th IEEE International Conference on Electronic Measurement Instruments (ICEMI), vol. 02, July 2015, pp. 808-812 [Non-Patent Document 14] Y. Zhao, H. Wang, and Z. Dai, "Model calibration of non-ideal lowpass filter in modulated wideband converter for compressive sampling", Compel-the International Journal for Computation and Mathematics in Electrical and Electronic Engineering, vol. 34, pp. 941-951, 2015 [Non-Patent Document 15] Y. Chen, M. Mishali, Y. C. Eldar, and A. O. Hero, "Modulated wideband converter with non-ideal lowpass filters", in 2010 IEEE International Conference on Acoustics, Speech and Signal Processing, 2010, pp. 3630-3633 [Non-Patent Document 16] L.-L. Nguyen, R. Gautier, A. Fiche, G. Burel, and E. Radoi, "Digital compensation of lowpass filters imperfection in the modulated wideband converter compressed sensing scheme for radio frequency monitoring", Signal Processing, vol. 152, pp. 292-310, 2018, [Online] http: / / www.sciencedirect.com / science / article / pii / S0165168418302081 [Non-Patent Document 17] L. Nguyen, A. Fiche, R. Gautier, C. Canaff, E. Radoi, and G. Burel, "Implementation of modulated wideband converter compressed sensing scheme based on cots lowpass filter with amplitude and phase compensation for spectrum monitoring", in 2018 15th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS), 2018, pp. 1-6

Summary of the Invention

[0003] A first aspect of the present invention provides a calibration device. The calibration device may include a calibration signal supply unit that supplies a multi-tone signal having tones at multiple frequencies as a calibration input signal to a converter that obtains multiple band-pass signals by multiplying an input signal by each of multiple signal patterns to band-limit the signal, and reconstructs an output signal corresponding to the input signal from the multiple band-pass signals. The calibration device may include a calibration band-pass signal acquisition unit that acquires the multiple calibration band-pass signals obtained by the converter in response to the calibration input signal. The calibration device may include a calibration processing unit that calibrates a correction filter that corrects each of the multiple band-pass signals based on each of the multiple calibration band-pass signals.

[0004] The calibration input signal may have a center frequency that is an integer multiple of the repetition frequency of the plurality of signal patterns, and may have tones at each of a plurality of frequencies that are arranged asymmetrically with respect to the center frequency within a frequency slot having a width of the repetition frequency.

[0005] Each tone of the calibration input signal may have a frequency symmetrical to the frequency of the tone about a center frequency at which no other tone of the calibration input signal is present.

[0006] Each tone of the calibration input signal may be arranged at a predetermined frequency interval within the frequency slot.

[0007] Each tone of the calibration input signal may have the same signal strength.

[0008] The calibration processing unit may estimate the frequency characteristics of at least a portion of the path from the calibration input signal to obtaining the first calibration bandpass signal within a frequency slot in accordance with changes in signal strength among multiple tones included in the frequency slot in the first calibration bandpass signal.

[0009] The calibration processing unit may calibrate the correction filter that corrects the first calibration bandpass signal to characteristics according to the inverse characteristics of the frequency characteristics within one frequency slot.

[0010] The calibration processing unit may calibrate a correction filter that corrects the first calibration bandpass signal, and for each frequency slot up to the cutoff frequency of the band limit, may have characteristics according to the inverse characteristics of the frequency characteristics in that frequency slot.

[0011] The calibration processing unit may perform an adjustment at the boundary between adjacent first and second frequency slots in the first calibration bandpass signal to bring the frequency characteristics of the path on the first frequency slot side closer to the frequency characteristics of the path on the second frequency slot side.

[0012] In a second aspect of the present invention, there is provided a conversion device. The conversion device may include a converter that obtains a plurality of band-pass signals by multiplying an input signal by each of a plurality of signal patterns to band-limit the signal, and reconstructs an output signal corresponding to the input signal from the plurality of band-pass signals. The conversion device may include a calibration device.

[0013] The converter may include a plurality of mixers that multiply an input signal by each of a plurality of signal patterns. The converter may include a plurality of band-pass filters that band-limit each of a plurality of signals output from the plurality of mixers. The converter may include a plurality of AD converters that output a plurality of band-pass signals obtained by sampling signals that have passed through the plurality of band-pass filters. The converter may include a plurality of correction filter units that output a plurality of corrected band-pass signals obtained by correcting each of the plurality of band-pass signals. The converter may include a reconstruction unit that reconstructs an output signal from the plurality of corrected band-pass signals.

[0014] Each of the plurality of correction filter units may include a low-pass filter that performs low-pass filtering on a band-pass signal input to the correction filter unit among the plurality of band-pass signals. Each of the plurality of correction filter units may include a correction filter that corrects the band-pass signal that has passed through the low-pass filter.

[0015] A third aspect of the present invention provides a calibration method. In the calibration method, a multi-tone signal having tones at multiple frequencies may be supplied as a calibration input signal to a converter that obtains multiple band-pass signals by multiplying an input signal by each of multiple signal patterns to band-limit the input signal, and reconstructs an output signal corresponding to the input signal from the multiple band-pass signals. In the calibration method, multiple calibration band-pass signals may be obtained by the converter in response to the calibration input signal. In the calibration method, a correction filter that corrects each of the multiple band-pass signals may be calibrated based on each of the multiple calibration band-pass signals.

[0016] A fourth aspect of the present invention provides a calibration program executed by a computer. The calibration program may cause the computer to function as a calibration signal supplying unit that supplies a multi-tone signal having tones at multiple frequencies as a calibration input signal to a converter that obtains multiple band-pass signals by multiplying an input signal by each of multiple signal patterns to band-limit the input signal and reconstructs an output signal corresponding to the input signal from the multiple band-pass signals. The calibration program may cause the computer to function as a calibration band-pass signal acquiring unit that acquires multiple calibration band-pass signals obtained by the converter in response to the calibration input signal. The calibration program may cause the computer to function as a calibration processing unit that calibrates a correction filter that corrects each of the multiple band-pass signals based on each of the multiple calibration band-pass signals.

[0017] The above summary of the invention does not list all of the necessary features of the present invention, and subcombinations of these features may also constitute inventions. [Brief explanation of the drawings]

[0018] [Figure 1] 1 shows the configuration of a conversion device according to this embodiment. [Figure 2] 10 shows an example of a sampling flow of the conversion device according to the present embodiment. [Figure 3]1 shows an example of the waveform of an input signal X(f) in the frequency domain. [Figure 4] 1 shows an example of the waveform of a signal pattern Pi(f) in the frequency domain. [Figure 5] 1 shows an example of the waveform of the input signal and the signal Ypi(f) multiplied by the signal pattern Pi(f) in the frequency domain. [Figure 6] 10 shows an example of the waveform of a bandpass signal Yi(f) obtained by band-limiting an input signal and a signal multiplied by a signal pattern in the frequency domain. [Figure 7] 1 shows an example of the waveform of an output signal Xo(f) reconstructed from a bandpass signal Yi(f) in the frequency domain. [Figure 8] 10 shows an example of a calibration flow of the conversion device according to the present embodiment. [Figure 9] 1 shows an example of waveforms of a calibration input signal Xcal(f) and a signal Ypi(f) multiplied by a signal pattern Pi(f) in the frequency domain. [Figure 10] 3 shows the configuration of a calibration signal supply unit according to the present embodiment. [Figure 11] 2 shows the configuration of a configuration processing unit according to the present embodiment. [Figure 12] 1 shows examples of ideal and actual frequency characteristics of a low-pass filter. [Figure 13] 1 shows an example of an unwanted image that appears in an output signal obtained by reconstructing a multi-tone signal. [Figure 14] 1 shows a model of the front-end part of the converter. [Figure 15] 10 shows the configuration of a conversion device according to a modified example of the present embodiment. [Figure 16] 10 shows the configuration of a correction filter unit according to a modified example of the present embodiment. [Figure 17] 10 shows an example of a calibration flow of a correction filter according to a modified example of the present embodiment. [Figure 18] 10 shows the configuration of a calibration processing unit according to a modified example of the present embodiment. [Figure 19] 10 shows an example of an operation flow of a calibration processing unit according to a modified example of the present embodiment. [Figure 20] 10 shows an example of a baseband spectrum of a calibration input signal Ẑk(f) according to a modified example of this embodiment. [Figure 21] 10 shows an example of the spectrum of a calibration input signal X(f) according to a modification of this embodiment. [Figure 22] 10 shows an example of the spectrum of a signal pattern Pi(f) according to a modified example of this embodiment. [Figure 23] 10 shows an example of the spectrum of a signal input to an AD converter according to a modified example of the present embodiment. [Figure 24] 10 shows an example of the spectrum of an output signal before and after correction according to a modified example of the present embodiment. [Figure 25] 10 shows an example of the spectrum of the output signal output by the AD converter and the correction filter unit in response to input of the calibration input signal X(f) before and after correction according to a modification of this embodiment. [Figure 26] 10 shows an example of frequency characteristics of a low-pass filter and a correction filter according to a modified example of the present embodiment. [Figure 27] 1 illustrates an example computer in which aspects of the present invention may be embodied in whole or in part. DETAILED DESCRIPTION OF THE INVENTION

[0019] The present invention will be described below through embodiments of the invention, but the following embodiments do not limit the scope of the invention according to the claims. Furthermore, not all of the combinations of features described in the embodiments are necessarily essential to the solution of the invention.

[0020] FIG. 1 shows the configuration of a conversion device 100 according to this embodiment. The conversion device 100 samples an analog input signal x(t), which may be a narrowband signal sparsely present in a wide frequency band, and outputs the sampled signal as a digital output signal xo(t), where t represents time. The conversion device 100 also has a function of performing calibration using a calibration input signal xcal(t) instead of the external input signal x(t).

[0021] The conversion device 100 includes a selector 110, periodic signal generators 120-1 to 120-m, mixers 130-1 to 130-m, band-pass filters 140-1 to 140-m, AD converters 150-1 to 150-m, a reconstruction unit 160, a calibration signal supply unit 170, a calibration band-pass signal acquisition unit 180, and a calibration processing unit 190. The periodic signal generators 120-1 to 120-m, mixers 130-1 to 130-m, band-pass filters 140-1 to 140-m, AD converters 150-1 to 150-m, and the reconstruction unit 160 apply a plurality of signal patterns p to an analog input signal x(t). i (t) (where i is an integer in the range of 1≦i≦m) and band-limited to generate multiple band-pass signals y i [n], and obtain multiple bandpass signals y i It functions as a converter that reconstructs an output signal xo(t) corresponding to an input signal x(t) from [n]. This converter may be an MWC.

[0022] Each of the multiple periodic signal generators 120-1 to 120-m (hereinafter also referred to as "periodic signal generators 120") generates a different signal pattern p i (t) is generated. i (t) is a periodic sign function (PSF) having a predetermined period Tp and having a positive or negative sign (for example, ±1) for each of M intervals of the period Tp. Here, the repetition frequency fp (Hz) of the signal pattern is 1 / Tp. The multiple periodic signal generators 120 generate the signal pattern p at the same timing using a clock of the period Tp as a trigger. i It may be configured to initiate the occurrence of (t).

[0023] The plurality of mixers 130-1 to 130-m (hereinafter also referred to as "mixers 130") are connected to the selector 110 and the plurality of periodic signal generators 120, and multiply the input signal x(t) by each of the plurality of signal patterns pi(t). i (t)=x(t)·p i Output (t).

[0024] The plurality of bandpass filters 140-1 to 140-m (hereinafter also referred to as "bandpass filters 140") are connected to the plurality of mixers 130, and are used to filter the plurality of signals yp i Then, the plurality of band-pass filters 140 band-limit each of the plurality of band-limited signals y i Output (t).

[0025] The AD converters 150-1 to 150-m (hereinafter also referred to as "AD converters 150") are connected to the band-pass filters 140, and the signal y i (t) are sampled digital bandpass signals y i Here, the sampling frequency of each of the AD converters 150 is represented as fs (Hz). The sampling frequency fs is the frequency of the signal pattern p i (t) is higher than the frequency fp of the code, for example, fs≧fp. n indicates the discrete time for each sampling timing, and t=n·Ts.

[0026] The reconstruction unit 160 is connected to the AD converters 150 and receives the bandpass signals y i [n], the reconstructor 160 reconstructs an output signal xo(t) corresponding to the input signal x(t). Here, the output signal xo(t) is a signal obtained by converting analog x(t) into digital. Alternatively, the reconstructor 160 may be an analog or digital signal obtained by down-converting the analog x(t). The reconstructor 160 may be realized by a dedicated circuit or a programmable circuit. Alternatively, the reconstructor 160 may be realized by executing a reconstruction program on a computer.

[0027] The selector 110, the calibration signal supply unit 170, the calibration bandpass signal acquisition unit 180, and the calibration processing unit 190 function as a calibration device that calibrates the converter that converts the input signal x(t) into the output signal xo(t). Such a calibration device may be realized by a dedicated circuit or a programmable circuit. Alternatively, the calibration device may be realized by executing a reconfiguration program on a computer.

[0028] The selector 110 switches between inputting an external input signal x(t) to the converter or inputting a calibration input signal xcal(t) to the converter. The selector 110 is connected to a calibration signal supply unit 170, and inputs the external input signal x(t) to the multiple mixers 130 during normal operation of the conversion device 100, and inputs the calibration input signal xcal(t) from the calibration signal supply unit 170 to the multiple mixers 130 during calibration operation of the conversion device 100.

[0029] The calibration signal supply unit 170 is connected to the calibration processing unit 190, generates a calibration input signal xcal(t) in response to an instruction from the calibration processing unit 190, and supplies it to the converter via the selector 110. The calibration input signal xcal(t) is converted into a plurality of calibration bandpass signals y by a plurality of mixers 130, a plurality of bandpass filters 140, and a plurality of filter filters 150, in the same manner as the input signal x(t). i Converted to [n].

[0030] The calibration bandpass signal acquisition unit 180 is connected to the plurality of AD converters 150. The calibration bandpass signal acquisition unit 180 acquires a plurality of calibration bandpass signals y obtained by the converters in response to the calibration input signal xcal(t). i Get [n].

[0031] The calibration processing unit 190 is connected to the calibration band-pass signal acquisition unit 180 and performs a calibration operation. In the calibration operation, the calibration processing unit 190 instructs the selector 110 to switch so as to supply the calibration input signal xcal(t) to the converter, and instructs the calibration signal supply unit 170 to generate the calibration input signal xcal(t). The calibration processing unit 190 calculates the calibration band-pass signals y acquired by the calibration band-pass signal acquisition unit 180 in response to the calibration input signal xcal(t). i The calibration processing unit 190 according to this embodiment sets or adjusts the reconstruction parameters used by the reconstruction unit 160.

[0032] 2 shows an example of a sampling flow of the conversion device 100 according to this embodiment. In step 200 (S200), the conversion device 100 receives an input signal x(t). In normal operation, the selector 110 supplies the input signal x(t) to multiple mixers 130.

[0033] In S210, the plurality of periodic signal generators 120 generate a plurality of signal patterns p i In S220, the plurality of mixers 130 mix the input signal x(t) with the plurality of signal patterns p i (t) and multiple signals yp i Generate and output (t).

[0034] In S230, the plurality of bandpass filters 140 receives the plurality of signals yp i (t) is a band-limited multi-bandpass signal y i In S240, the AD converters 150 generate and output the band-pass signals y i By sampling (t) and AD (Analog-Digital) converting it, multiple digital bandpass signals y i Get [n].

[0035] In S250, the reconstruction unit 160 reconstructs a plurality of bandpass signals y iThe output signal xo(t) is reconstructed based on [n]. The sampling operation of the conversion device 100 will be described in more detail below with reference to FIGS.

[0036] 3 shows an example of the waveform of input signal X(f) in the frequency domain. Input signal x(t) is a narrowband signal, such as 20 MHz, that exists in a portion of a wide frequency band, such as 2.4 GHz. Therefore, in the frequency domain, where the horizontal axis represents frequency and the vertical axis represents the amplitude of each frequency component, input signal x(t) becomes input signal X(f) that has narrowband signals 310-1-2 and corresponding negative frequency component signals 320-1-2 in a portion of the wide frequency band, as shown in FIG.

[0037] Figure 4 shows the signal pattern P i An example of the waveform of (f) is shown below. i (t) has a period Tp, and can therefore be expanded by Fourier expansion as shown in the following equation (1).

number

[0038] From equation (1), the signal pattern p i (t) is the signal pattern P i (f) is expressed in the range of l = -L0 to L0 as shown in Figure 4. Here, c i,l is the frequency domain signal pattern P i (f) shows the frequency component of frequency l·fp.

[0039] FIG. 5 shows the input signal X(f) and the signal pattern P i Signal Yp multiplied by (f) i An example of the waveform of the signal Yp in the frequency domain is shown in (f). i (f) is the signal yp in the time domain i (t)=x(t)·p i (t) is converted into the frequency domain and is expressed by the Fourier transform shown in the following equation (2).

number

[0040] From equation (2), the signal Yp i As shown in Figure 5, (f) is obtained by shifting the input signal X(f) by an integer multiple lfp of fp. i Each frequency component c of (f) i,l In this way, the converter can frequency-convert narrow-band signals sparsely present in a wide frequency band to near baseband.

[0041] FIG. 6 shows the input signal X(f) and the signal pattern P i Signal Yp multiplied by (f) i (f) is a band-limited bandpass signal Y i In this embodiment, the bandpass filter 140 is configured to filter the signal at a frequency f bpf Low-pass filtering is performed on frequency components below frequency f bpf In this case, the band pass filter 140 is a low pass filter that cuts the above frequency components. i (f) at frequency -f bpf The following frequency components and frequency f bpf The frequency components above the frequency -f bpf Exceeding frequency f bpf Low-pass filtering is performed on frequency components below frequency f bpf may be twice the sampling frequency fs or more to achieve sub-Nyquist sampling.

[0042] Bandpass signal Y i is the digital bandpass signal y i It can be approximated by the discrete Fourier transform of [n] as shown in the following equation (3).

number

[0043] where L0 is the sum of all non-zero components of X(f) and the bandpass signal Y i L0 is a positive integer selected so as to be included in the range of -fs / 2≦f≦fs / 2. L0 may be the smallest positive integer that satisfies this condition.

[0044] Equation (3) can be expressed in the form of a matrix vector product as shown in equation (4) below.

number

[0045] Figure 7 shows the bandpass signal Y in the frequency domain. i An example of the waveform of the output signal Xo(f) reconstructed from (f) is shown below. i,l Let A be the matrix of the above. The matrix A has m rows and 2L0+1 columns. Since 2L0+1 is very large compared to m, the left side of Y i However, if the input signal X(f) is sufficiently sparse, most of the X(f-lfp) on the right side will be 0, and the multiple parameters c i,l Using Y on the left side i The vector X(f-lfp) on the right-hand side can be uniquely calculated from the vector.

[0046] Based on this principle, the reconstruction unit 160 reconstructs the plurality of bandpass signals y i [n] into the frequency domain bandpass signal Y i (f), and each frequency component c of the signal pattern shown in Fig. 4 i,l Reconstruct multiple bandpass signals Y in the frequency domain using the reconstruction parameters calculated from i 3 from (f), and then inversely transforming the frequency domain output signal Xo(f) into the time domain output signal xo(t), thereby obtaining the output signal xo(t). Here, the reconstruction unit 160 may output the output signal xo(t) in the form of a digital discrete-time signal xo[n].

[0047] In the above description, the frequency domain bandpass signal Y i Alternatively, the transform device 100 may use time-domain operations equivalent to the above frequency-domain operations to reconstruct the time-domain bandpass signal y i It is also possible to reconstruct the time domain output signal xo(t) from [n].

[0048] In the above description, each bandpass filter 140 is a lowpass filter, but the above processing can be performed in the same way as long as a signal in a frequency range sufficient to reconstruct the output signal xo(t) can be bandpass filtered, so each bandpass filter 140 may be a bandpass filter.

[0049] The conversion device 100 described above can sample narrowband signals in a wide frequency band that includes frequencies from -L0fp to L0fp using multiple AD converters 150 having a sampling frequency fs that is very small compared to frequency L0fp. This makes it possible for the conversion device 100 to sample, for example, narrowband signals of 20 MHz that are sparsely present in a frequency band from several GHz to several tens of GHz using multiple AD converters 150 with a sampling frequency of 80 MHz.

[0050] 8 shows an example of a calibration flow of the conversion device according to this embodiment. In S800, the calibration signal supplying unit 170 generates a calibration input signal xcal(t) in response to an instruction to start a calibration operation from the calibration processing unit 190. In the calibration operation, the selector 110 supplies the calibration input signal xcal(t) to the multiple mixers 130.

[0051] In S810, the plurality of periodic signal generators 120 generate the plurality of signal patterns p iIn S820, the plurality of mixers 130 generate the calibration input signal xcal(t) and the plurality of signal patterns p i (t) and multiple signals yp i Generate and output (t).

[0052] In S830, the plurality of band-pass filters 140 are configured to filter the plurality of signals yp i (t) is band-limited to generate multiple calibration band-pass signals y i In S840, the AD converters 150 generate and output the calibration bandpass signals y (t), similar to S240 during normal operation. i (t) is sampled and converted to analog signals to generate multiple digital calibration bandpass signals y i The calibration bandpass signal acquisition unit 180 outputs the calibration bandpass signals y [n] output by the AD converters 150. i Get [n].

[0053] In S850, the calibration processing unit 190 generates a plurality of calibration bandpass signals y i Based on [n], the reconstruction parameters used by the reconstruction unit 160 are calibrated. The calibration operation of the conversion device 100 will be described in more detail below with reference to FIG.

[0054] FIG. 9 shows the calibration input signal Xcal(f) and the signal pattern P i Signal Yp multiplied by (f) i 1 to 7, the conversion device 100 ideally generates a waveform for each signal pattern p i Frequency component c of (t) i,l is given, the output signal xo(t) corresponding to the input signal x(t) can be correctly reconstructed. However, in reality, the conversion device 100 is very sensitive and is significantly affected by the characteristics of the components and wiring actually used. Therefore, in order to correctly reconstruct the output signal xo(t), calibration of the conversion device 100 is required.

[0055] Non-Patent Documents 3 to 8 disclose techniques for calibrating reconstruction parameters by sweeping the frequency domain using a single-tone calibration signal. However, such techniques require a long time to calculate all reconstruction parameters from measurement results using single-tone calibration signals generated at different timings. Furthermore, since single-tone calibration signals are generated at multiple different timings, these calibration signals may be affected by phase shifts due to random jitter superimposed on the start timing, which may make it difficult to set reconstruction parameters accurately.

[0056] In contrast, the calibration signal supplying section 170 according to this embodiment supplies a multi-tone signal having tones in multiple frequency bands as the calibration input signal xcal(t) to the converter, as illustrated in the frequency spectrum in the upper part of Fig. 9. The calibration input signal xcal(t) is a signal having multiple signal patterns p i The mixer 130 may have a tone in each of a plurality of different frequency bands, each of which is centered at a frequency l·fp that is an integer multiple l of the repetition frequency fp of the calibration input signal xcal(t) and has a width of the repetition frequency fp. i By multiplying by (t), the signal pattern p i (t) i,l As a result, the bandpass filter 140 can convert the frequency components c i,l and the AD converter 150 passes the different frequency components c i,l A bandpass signal y containing i [n] can be sampled.

[0057] The calibration input signal xcal(t) may have tones in multiple frequency bands obtained by dividing the frequency range to be detected in the input signal x(t) for each repetition frequency fp. When the frequency range to be detected is from 0 to (L0+1 / 2)fp, the calibration input signal xcal(t) may have a tone in each frequency band centered on frequency l·fp (l=0 to L0), obtained by dividing the frequency range to be detected for each repetition frequency fp, as shown in the upper part of Figure 9. Alternatively, the calibration input signal xcal(t) may have a tone in each of the frequency bands corresponding to at least two l's.

[0058] Here, the tones in two different frequency bands among the above-described multiple frequency bands may have a frequency difference that is offset from an integer multiple of the repetition frequency fp. By providing such a frequency difference between the two tones, the bandpass filter 140 applies the signal pattern p i As a result of multiplying by (t), the different frequency components c i,l to a position shifted by this frequency difference in the vicinity of the baseband. i,l can be separated and detected.

[0059] The calibration input signal xcal(t) shown in Figure 9 has tones with frequencies f0 + l·fp + lΔf within the target frequency range, where f0 is the fundamental frequency, fp is the repetition rate, Δf is the offset frequency, and l is an integer. This calibration input signal xcal(t) has a frequency difference of fp + Δf between a tone in a frequency band and a tone in an adjacent high-frequency band. Here, the fundamental frequency f0 may be -½fp ​​≦ f0 ≦½fp, and the offset frequency Δf may be a positive or negative frequency so that all tones do not fall outside the corresponding frequency band.

[0060] When Δf is a positive infinitesimal frequency, the calibration input signal xcal(t) is given a signal pattern p iWhen multiplied by (t), the signal pattern P i (f) Frequency component c of each frequency l·fp i,l (l=0 to L0) are compressed and arranged in the same order on the frequency axis near the baseband, and the bandpass signal Yp i Therefore, the calibration processing unit 190 obtains the bandpass signal Yp i Frequency component c in frequency order from (f) i,1 When Δf is a negative minute frequency, the band-pass signal Yp i Frequency component c in (f) i,l The order is reversed.

[0061] 9, the calibration input signal xcal(t) may have a tone with a frequency of f0+l·fp+qΔf within the frequency range to be detected, where q may be an integer uniquely selected to satisfy 0≦q≦L0 for multiple frequency bands, or may be an integer randomly selected under this condition.

[0062] In this way, the conversion device 100 calculates the bandpass signal y obtained in response to the multi-tone calibration input signal xcal(t) for the i-th front end including the periodic signal generator 120-i, the mixer 130-i, the bandpass filter 140-i, and the AD converter 150-i. i From [n], P i (f) Multiple frequency components c i,l (l=0 to L0). Therefore, the calibration processing unit 190 can obtain these multiple frequency components c i,l It is possible to calibrate a plurality of reconstruction parameters corresponding to the above-mentioned equations collectively.

[0063] Furthermore, the conversion device 100 may generate a bandpass signal y obtained in response to a multi-tone calibration input signal xcal(t) supplied at the same time for two or more or all of the front ends. i From [n], P i(f) Multiple frequency components c i,l (l=0 to L0) can be obtained. In this case, the calibration processing unit 190 can obtain these multiple frequency components c i,l It is possible to calibrate multiple reconstruction parameters corresponding to , or all reconstruction parameters at once.

[0064] Here, the multiple frequency components c acquired in response to the multi-tone calibration input signal xcal(t) supplied at the same time are i,l is not affected by random jitter that occurs when the calibration input signal xcal(t) is supplied at different timings. Therefore, the calibration processing unit 190 can calibrate the reconstruction parameters of the reconstructor 160 with high accuracy.

[0065] Note that the above description has been given on the simple case where each tone of the calibration input signal xcal(t) has an amplitude of 1 and a phase of 0. If each tone has at least one of an amplitude other than 1 and a phase other than 0, the bandpass signal Yp i Each frequency component in (f) corresponds to a corresponding tone of the calibration input signal X(f) in the frequency domain and a signal pattern P i The corresponding frequency component c in (f) i,l Therefore, the calibration processing unit 190 converts the bandpass signal Yp i The signal pattern P is obtained by dividing each frequency component in (f) by the frequency component of the corresponding tone in the calibration input signal X(f). i The corresponding frequency components of (f) can be calculated.

[0066] Fig. 10 shows the configuration of a calibration signal supplying unit 1000 according to this embodiment. The calibration signal supplying unit 1000 may be used as the calibration signal supplying unit 170 in Fig. 1, or may have a different configuration from the calibration signal supplying unit 170 in Fig. 1.

[0067] The calibration signal supplying section 1000 includes a trigger generator 1010, a plurality of sine wave generators 1020-1 to 1020-k, a plurality of attenuators 1030-1 to 1030-k, a multiplexer 1040, and a phase / amplitude setting section 1050. The trigger generating section 1010 generates a trigger that indicates the start timing of the calibration input signal xcal(t). The trigger generator 1010 may periodically generate the trigger to periodically generate the calibration input signal xcal(t).

[0068] A plurality of sine wave generators 1020-1 to k (also referred to as "sine wave generators 1020") are connected to the trigger generator 1010 and generate respective tones to be included in the calibration input signal xcal(t). For example, in FIG. 9, sine wave generator 1020-1 generates a tone with frequency f0, sine wave generator 1020-2 generates a tone with frequency fp+f0+Δf, ..., sine wave generator 1020-k generates a tone with frequency L0fp+f0+L0Δf. In this case, k is L0+1. Here, each tone is a sine wave signal having a single frequency.

[0069] In this embodiment, each sine wave generator 1020 receives an instruction of the initial phase at which a tone should be output from the phase and amplitude setting unit 1050, and outputs a tone having a specified initial phase relative to a trigger. Each sine wave generator 1020 may receive an instruction of a delay time for the start timing of the tone relative to the trigger as the instruction of the initial phase at which the tone should be output.

[0070] A plurality of attenuators 1030-1 to k (also referred to as "attenuators 1030") are connected to a plurality of sine wave generators 1020, and each amplifies or attenuates the amplitude of the tone output by the corresponding sine wave generator 1020 according to a weight specified by phase and amplitude setting section 1050. As a result, each attenuator 1030 outputs a tone obtained by multiplying the tone from the corresponding sine wave generator 1020 by the weight specified by phase and amplitude setting section 1050.

[0071] The multiplexer 1040 is connected to the multiple attenuators 1030. The multiplexer 1040 combines (i.e., adds) the tones output by the multiple attenuators 1030, with at least one of the amplitudes or phases adjusted, and outputs the combined signal as a calibration input signal xcal(t).

[0072] The phase and amplitude setting section 1050 instructs each sine wave generator 1020 of the initial phase of the tone to be output by that sine wave generator 1020. The phase and amplitude setting section 1050 also instructs each attenuator 1030 of the weight by which that attenuator 1030 should multiply the tone. In this way, the phase and amplitude setting section 1050 performs at least one of shifting the initial phases of the multiple tones and adjusting the amplitudes of the multiple tones, thereby limiting the maximum amplitude of the calibration input signal xcal(t) within the rated range of the converter.

[0073] More specifically, to ensure proper operation of a converter including multiple mixers 130, multiple band-pass filters 140, and multiple AD converters 150, the calibration signal supplying unit 1000 according to this embodiment limits the amplitude of the calibration input signal to within the rated range of the converter. Here, the calibration signal supplying unit 1000 supplies a composite wave of multiple tones to the converter as the calibration input signal. If multiple tone peaks occur simultaneously, the maximum amplitude of the composite wave becomes large. Therefore, the phase and amplitude setting unit 1050 adjusts the initial phases to be instructed to the multiple sine wave generators 1020 so that the peaks of the multiple tones are offset. Here, the phase and amplitude setting unit 1050 may store a set of initial phases designated in advance by the manufacturer or user of the converter 100, and set the stored initial phases for each sine wave generator 1020. Alternatively, phase and amplitude setting section 1050 may calculate the maximum amplitude of the composite wave corresponding to the initial phases of the multiple tones by numerical calculation, simulation, or the like, and adjust the initial phase of each tone so that the amplitude of the composite wave is minimized. Also, phase and amplitude setting section 1050 may limit the maximum amplitude of the calibration input signal xcal(t) within the rated range of the converter by making the amplitudes of the multiple tones substantially identical and setting the phases of the multiple tones to the Neuman phase described in Non-Patent Document 9.

[0074] The multiplexer 1040 may also limit the amplitude of the calibration input signal within the rated range of the converter by specifying weights for the tones output by the corresponding sine wave generator 1020 for each of the multiple attenuators 1030. The phase and amplitude setting unit 1050 may uniformly change the amplitudes of the multiple tones, or may adjust the amplitudes of at least two tones individually so that the amplitudes of these tones are different. The phase and amplitude setting unit 1050 may store a set of weights specified in advance by the manufacturer or user of the converter 100 and set the stored weights for each attenuator 1030. Alternatively, the phase and amplitude setting unit 1050 may calculate the maximum amplitude of the composite wave corresponding to the amplitudes of the multiple tones by numerical calculation, simulation, or the like, and adjust the amplitude of the composite wave to minimize it. For example, the phase and amplitude setting unit 1050 may reduce the amplitude of a tone that contributes more to increasing the amplitude of the composite wave at the timing when the composite wave reaches its maximum amplitude. The phase and amplitude setting section 1050 may adjust the initial phase of each tone and the amplitude of each tone together to keep the maximum amplitude of the composite wave within the rated range of the converter.

[0075] In the above, if the conversion device 100 does not measure any of the multiple frequency bands, the calibration signal supplying section 1000 may be configured without a set of sine wave generator 1020 and attenuator 1030 corresponding to that frequency band. Note that instead of some or all of the multiple sine wave generators 1020 and multiple attenuators 1030, a method may be used in which a computer generates a composite wave of some or all of the corresponding tones using calculations based on program processing, and a DA converter or an arbitrary waveform generator (AWG) converts this into analog and outputs it.

[0076] Fig. 11 shows the configuration of a calibration processing unit 1100 according to this embodiment. The calibration processing unit 1100 may be used as the calibration processing unit 190 in Fig. 1, or may have a different configuration from the calibration processing unit 190 in Fig. 1.

[0077] The calibration processing unit 1100 includes a signal component detection unit 1110 and a calibration parameter calculation unit 1120. The signal component detection unit 1110 is connected to a calibration bandpass signal acquisition unit such as the calibration bandpass signal acquisition unit 180, and receives a plurality of bandpass signals y i From each of [n], the corresponding signal pattern p i Each frequency component c contained in (t) i,l The signal component detector 1110 according to this embodiment detects a plurality of band-pass signals y i [n] into the frequency domain bandpass signal Y i (f) (see equation (3)). Then, the signal component detector 1110 converts the frequency domain bandpass signal Y i From (f), signal pattern p i Each frequency component c contained in (t) i,l Here, since the calibration input signal xcal(t) is known, the signal component detector 1110 extracts the signal pattern p i Each frequency component c contained in (t) i,l The amplitude and initial phase of can be calculated inversely.

[0078] The calibration parameter calculation unit 1120 is connected to the signal component detection unit 1110 and detects each frequency component c i,l Using the above, multiple bandpass signals Y i(f) (i.e., the vector on the left side of equation (4)), the reconstruction parameters for calculating the signals of each frequency band of the frequency domain output signal X(f) (i.e., the vector X(f-lfp) on the right side of equation (4)) are calculated. Here, since the frequency domain output signal X(f) is sufficiently sparse, if the frequency band in which there is no signal in the output signal X(f) is known, X(f-lfp) corresponding to that frequency band and the column of matrix A corresponding to that X(f-lfp) can be deleted from equation (4). Then, the calibration parameter calculation unit 1120 calculates the generalized inverse matrix (denoted as B) of the matrix A thus degenerated, and sets it as the reconstruction parameters. The reconstruction unit 160 uses the reconstruction parameters thus obtained to calculate a plurality of band-pass signals y(t) corresponding to the input signal x(t). i [n] is converted into multiple bandpass signals Y in the frequency domain by discrete Fourier transform. i (f) and convert it into multiple bandpass signals Y i By multiplying the vector of (f) by matrix B, multiple signals X(f-lfp) included in the output signal X(f) can be obtained.

[0079] Alternatively, the calibration parameter calculation unit 1120 may calculate the frequency component c i,l may be supplied to the reconstruction unit 160 as a reconstruction parameter. In this case, the reconstruction unit 160 may supply each frequency component c i,l The generalized inverse matrix B may be calculated from the values ​​of (f) and then the output signal X(f) may be calculated.

[0080] Here, if the frequency band in which there is no signal in the output signal X(f) is unknown, the reconstruction unit 160 may perform the above processing after detecting the frequency band in which there is a signal in the output signal X(f). As an example, the reconstruction unit 160 may reconstruct the output signal X(f) using the method described in Non-Patent Document 1.

[0081] The calibration processing unit 1100 according to this embodiment calculates the reconstruction parameters using a calculation in the frequency domain. Alternatively, the calibration processing unit 1100 may calculate the reconstruction parameters using a calculation in the time domain that corresponds to the above-described calculation in the frequency domain.

[0082] FIG. 12 shows an example of the ideal and actual frequency characteristics of a low-pass filter used as the band-pass filter 140. In this graph, the horizontal axis represents frequency and the vertical axis represents frequency characteristic H(f). Ideally, a low-pass filter has a cutoff frequency (f bpf ), and the gain is preferably 0 at frequencies equal to or higher than the cutoff frequency. However, since the low-pass filter used as band-pass filter 140 is an analog circuit, in reality, the gain varies depending on the frequency below the cutoff frequency as shown by the thick solid line, and the gain decreases gradually even at frequencies equal to or higher than the cutoff frequency.

[0083] Ideally, a low-pass filter should have a constant phase characteristic that is independent of frequency. However, in reality, the low-pass filter used as band-pass filter 140 has a phase characteristic in which the phase changes depending on the frequency due to the influence of group delay, etc., as shown by the thin solid line.

[0084] Figure 13 shows an example of an unwanted image that appears in an output signal obtained by reconstructing a multi-tone signal. In this figure, the horizontal axis represents frequency and the vertical axis represents signal strength (dBFS). The graph in this figure shows the output signal obtained by reconstructing a multi-tone signal having tones at regular frequency intervals in the frequency band from 90 to 90.5 MHz centered at 90.25 MHz using WMC.

[0085] When using a real low-pass filter, the reconstructed output signal contains spurious signals generated by the low-pass filter that are superimposed as unwanted images on the signal corresponding to the original multi-tone signal. In the example shown in this figure, these unwanted images are particularly noticeable in the frequency range of 89.5 to 90 MHz.

[0086] 14 shows a model of the front-end portion of the conversion device 100. The model in this figure is a schematic representation of a set of mixer 130, band-pass filter 140, and AD converter 150 in the conversion device 100.

[0087] The mixer 130 can be represented by a model in which the frequency characteristics of a signal input characteristic 1410, a PSF input characteristic 1420, and a mixer output characteristic 1430 are superimposed on an ideal mixer 1400. The ideal mixer 1400 is a mixer having ideal frequency characteristics (frequency response). In an actual mixer 130, the input signal x(t) is affected by a signal input characteristic 1410 that is specific to the mixer 130. This signal input characteristic 1410 is expressed as H RF This indicates:

[0088] In the actual mixer 130, the ideal signal pattern p i (t) is affected by a PSF input characteristic 1420 specific to the mixer 130. This PSF input characteristic 1420 is expressed as H LO In the actual mixer 130, the signal output from the mixer 130 is affected by a mixer output characteristic 1430 that is specific to the mixer 130. This mixer output characteristic 1430 is expressed as H IF This indicates:

[0089] An actual low-pass filter used as the band-pass filter 140 has a frequency characteristic H that is different from the ideal frequency characteristic. LPF This frequency characteristic H LPF is the frequency characteristic shown by the solid line in FIG.

[0090] The AD converter 150 can be represented by a model in which the influence of an AD conversion characteristic 1450 is superimposed on an ideal AD converter 1440 having ideal frequency characteristics. ADC This indicates:

[0091] In the above, each actual frequency component c of the signal pattern input to the ideal mixer 1400 ~ i,l is expressed as the ideal frequency component c i,l PSF input characteristics 1420 (H LO ) is multiplied.

number

[0092] As shown with respect to FIGS. 1 to 11, the calibration processor 190 and the calibration processor 1100 calculate the actual frequency component c ~ i,l can be calculated to calibrate the reconstruction unit 160. Here, as shown in the following equation (6), the mixer output characteristic 1430 (H IF ), the frequency characteristics of the bandpass filter 140 (H LPF ), and AD conversion characteristics 1450 (H ADC ) is the equivalent frequency characteristic of H eq Let's say.

number

[0093] When the sampling frequency fs of the AD converter 150 is equal to the frequency fp of the signal pattern (when fs=fp), the actual bandpass signal Y output by the i-th front-end AD converter 150 is ~ i (e jωTs ) is expressed by the following equation (7).

number

[0094] where R l(f) is H RF (f) is the spectrum of one frequency slot (a frequency band of frequency width fp centered at frequency l fp), and R l (f)=H RF (f-lf p ) In addition, the spectrum of the signal in the frequency slot with the center frequency lfp and the frequency width fp in the input signal X(f) is Z l Marked as (f).

[0095] When the sampling frequency fs of the AD converter 150 is q times the frequency fp of the signal pattern (when fs=qfp and q is a positive integer), the actual bandpass signal Y' output by the i-th front-end AD converter 150 is i (e jωTs ) is expressed by the following equation (8).

number

[0096] Note that the influence of the signal input characteristic 1410 is small and can be ignored. Alternatively, a conversion device such as the conversion device 100 may use, as an example, an equivalent frequency characteristic H eq After correcting for and calibrating the reconstruction unit 160 as shown in FIGS. 1 to 11, a reference input signal may be input to perform calibration for the signal input characteristic 1410.

[0097] Fig. 15 shows the configuration of a conversion device 1500 according to a modified example of this embodiment. The conversion device 1500 shown in this figure is a modified example of the conversion device 100 shown in Fig. 1. Of the components shown in this figure, components with the same reference numerals as in Fig. 1 have the same functions and configurations as the corresponding components in Fig. 1. Therefore, explanations will be omitted below except for the differences.

[0098] The conversion device 1500 is a conversion device 100 shown in FIG. 1, which is a conversion device having an equivalent frequency characteristic H eqThe conversion device 1500 has a configuration in which a plurality of correction filter units 1555-1 to 1555-m (also referred to as "correction filter units 1555") are added to the rear stage of the plurality of AD converters 150 to 150m in the conversion device 100. The plurality of correction filter units 1555 output a plurality of corrected band-pass signals obtained by correcting the plurality of band-pass signals y1[n] to ym[n] from the plurality of AD converters 150. As a result, each correction filter unit 1555 corrects the band-pass signal y i Equivalent frequency characteristic H superimposed on [n] eq The reconstructor 160 reconstructs the output signal from the plurality of corrected bandpass signals.

[0099] 1 , the conversion device 1500 includes a calibration signal supplying unit 1570 and a calibration processing unit 1590 instead of the calibration signal supplying unit 170 and the calibration processing unit 190. The calibration signal supplying unit 1570 is connected to the calibration processing unit 1590. In response to an instruction from the calibration processing unit 1590, the calibration signal supplying unit 1570 generates a calibration input signal xcal(t) for calibrating each correction filter unit 1555, and supplies the signal via the selector 110 to a converter having a plurality of periodic signal generators 120, a plurality of mixers 130, a plurality of band-pass filters 140, a plurality of AD converters 150, a plurality of correction filter units 1555, and a reconstruction unit 160.

[0100] The calibration processing unit 1590 is connected to the calibration bandpass signal acquiring unit 180. The calibration processing unit 1590 acquires a plurality of calibration bandpass signals y obtained by the converter in response to the calibration input signal xcal(t) generated by the calibration signal supplying unit 1570. i1 to 11 , the calibration processing unit 1590 receives a plurality of calibration band-pass signals [n] from the calibration band-pass signal acquisition unit 180. In the example shown in the figure, the calibration processing unit 1590 receives a plurality of calibration band-pass signals corrected by a plurality of correction filter units 1555 from the calibration band-pass signal acquisition unit 180. The calibration processing unit 1590 calibrates the correction filters in the correction filter units 1555 for correcting each of the plurality of band-pass signals, based on each of the plurality of calibration band-pass signals. Note that the calibration signal supplying unit 1570 and the calibration processing unit 1590 may include the functions and configurations of the calibration signal supplying unit 170 and the calibration processing unit 190 for calibrating the reconstruction parameters used by the reconstruction unit 160, as shown in FIGS.

[0101] The calibration processing unit 1590 described above receives a plurality of calibration band-pass signals corrected by a plurality of correction filter units 1555 from the calibration band-pass signal acquisition unit 180. Therefore, the equivalent frequency characteristic H eq is the frequency characteristic of a part of the path from the calibration input signal to obtaining the calibration band-pass signal. Calibration processing unit 1590 calibrates the correction filter using the multiple calibration band-pass signals corrected by multiple correction filter units 1555, and can further repeat the calibration process on the calibrated correction filter, thereby improving the accuracy of the correction filter.

[0102] Alternatively, the calibration processing unit 1590 may receive, from the calibration band-pass signal acquisition unit 180, a plurality of calibration band-pass signals output by the plurality of AD converters 150 before being corrected by the plurality of correction filter units 1555, and calibrate the plurality of correction filters. In this case, the equivalent frequency characteristic H eq is the frequency characteristic of the entire path from the calibration input signal to obtaining the calibration bandpass signal.

[0103] 16 shows the configuration of a correction filter unit 1555 according to a modification of this embodiment. The correction filter unit 1555 includes an upsampler 1600, a low-pass filter 1610, a correction filter 1620, and a downsampler 1630.

[0104] The upsampler 1600 upsamples the bandpass signal input to the correction filter unit 1555 out of the multiple bandpass signals from the multiple AD converters 150. The lowpass filter 1610 is connected to the upsampler 1600. The lowpass filter 1610 low-pass filters the bandpass signal upsampled by the upsampler 1600 to reduce or remove high-frequency components equal to or higher than the cutoff frequency. The cutoff frequency of the lowpass filter 1610 may be the same as the ideal cutoff frequency that the bandpass filter 140 should have (the cutoff frequency shown in FIG. 12). Alternatively, the lowpass filter 1610 may have a cutoff frequency higher than the cutoff frequency of the bandpass filter 140. The frequency characteristics of the lowpass filter 1610 can be expressed as H S (e jω ) is indicated.

[0105] The correction filter 1620 is connected to the low-pass filter 1610. The correction filter 1620 corrects the band-pass signal that has passed through the low-pass filter 1610. The frequency characteristic of the correction filter 1620 is expressed as H C (e jω ) is indicated.

[0106] The downsampler 1630 downsamples the bandpass signal corrected by the downsampler 1630. The downsampler 1630 may be a decimation filter, or may be a thinning filter that thins out the bandpass signal and outputs it.

[0107] The correction filter unit 1555 described above is configured to correct the cutoff frequency (f bpf) (see FIG. 12 ) can be low-pass filtered by a digital low-pass filter 1610. As the low-pass filter 1610, for example, a digital filter whose order or the like is adjusted in advance so that the error of the reconstructed output signal output by the reconstruction unit 160 in accordance with the frequency components in the band-pass signal that are equal to or higher than the cutoff frequency falls within an allowable range can be used.

[0108] Furthermore, the correction filter unit 1555 can correct frequency components below the cutoff frequency in the digital bandpass signal using the correction filter 1620. As an example, the correction filter 1620 can correct the equivalent frequency characteristic H eq The compensation filter 1620 is adjusted to have the inverse characteristic of the above. This allows the compensation filter 1620 to compensate the band-pass signal so that the front end from the mixer 130 to the compensation filter unit 1555 has a constant gain below the cut-off frequency. The compensation filter 1620 may also compensate the band-pass signal so that the front end has a constant phase delay below the cut-off frequency.

[0109] In this modification, the upsampler 1600 and the downsampler 1630 temporarily increase the frequency of the bandpass signal in order to improve the accuracy of the low-pass filtering process by the low-pass filter 1610 and the correction process by the correction filter 1620. Alternatively, the correction filter unit 1555 may be configured without at least one of the upsampler 1600 and the downsampler 1630.

[0110] Fig. 17 shows an example of a calibration flow of a correction filter according to a modified example of this embodiment. The calibration flow according to this modified example is a modified example of the configuration flow shown in Fig. 8, so a description will be omitted below except for the differences.

[0111] In S1700, the calibration signal supplying unit 1570 generates a calibration input signal xcal(t), which is a multi-tone signal having tones at multiple frequencies. The calibration input signal xcal(t) used to calibrate the correction filter 1620 will be described later with reference to Figures 20 and 21. In the calibration operation, the selector 110 supplies the calibration input signal xcal(t) to multiple mixers 130.

[0112] S810 to S830 are the same as S810 to S830 in Fig. 8. In S1740, the plurality of AD converters 150 generate the plurality of calibration bandpass signals y i (t) is sampled and converted to analog signals to generate multiple digital calibration bandpass signals y i In this modification, the plurality of correction filter units 1555 outputs the plurality of digital calibration bandpass signals y [n] output by the plurality of AD converters 150. i The corrected signal [n] is then converted into multiple digital calibration bandpass signals y i The calibration bandpass signal acquisition unit 180 outputs the plurality of calibration bandpass signals y [n] output by the plurality of correction filter units 1555. i In other respects, S1740 is the same as S840, so explanations will be omitted hereinafter except for the differences.

[0113] In S1750, the calibration processing unit 190 generates a plurality of calibration bandpass signals y i [n], the conversion device 1500 calibrates the correction filters 1620 of the plurality of correction filter units 1555. Note that the conversion device 1500 may calibrate the correction filters 1620 of the plurality of front ends based on the results of simultaneously supplying the calibration input signal xcal(t) to the plurality of front ends (S1700 to S1740). Alternatively, the conversion device 1500 may perform the process of FIG. 17 individually for each of the plurality of front ends. Below, the configuration of the calibration processing unit 1590 and the calibration operation of the correction filter 1620 by the conversion device 1500 will be described in more detail with reference to FIGS. 18 to 26.

[0114] 18 shows the configuration of a calibration processing unit 1590 according to a modification of this embodiment. The calibration processing unit 1590 includes a frequency characteristic calculation unit 1810, an interpolation processing unit 1820, a correction filter determination unit 1830, and a correction filter setting unit 1840.

[0115] The frequency characteristic calculation unit 1810 calculates a plurality of corrected calibration bandpass signals y i Here, the frequency characteristic calculation unit 1810 receives a plurality of calibration bandpass signals y i [n] is the bandpass signal Y' in equation (8). i (e jωTs ) is corrected by the correction filter unit 1555, each tone of the calibration input signal xcal(t) corresponds to the signal pattern P i Each frequency component c of (f) i,l Each calibration bandpass signal y i Each frequency tone in [n] has an equivalent frequency characteristic H eq The frequency characteristics are affected by the

[0116] The frequency characteristic calculation unit 1810 calculates the calibration bandpass signal y i Based on each tone in [n], the equivalent frequency characteristic H^ at the frequency of each tone is calculated. eq The equivalent frequency characteristic H^ calculated by the frequency characteristic calculation unit 1810 is calculated. eq becomes a discrete equivalent frequency characteristic having characteristics corresponding only to the frequency position of each tone.

[0117] The interpolation processing unit 1820 is connected to the frequency characteristic calculation unit 1810. The interpolation processing unit 1820 calculates a discrete equivalent frequency characteristic Ĥ for each front end i. eq By interpolating the eq Estimate.

[0118] The correction filter determination unit 1830 is connected to the interpolation processing unit 1820. The correction filter determination unit 1830 determines the estimated equivalent frequency characteristic H eq Here, the correction filter determination unit 1830 determines the frequency characteristic of the correction filter 1620 based on the estimated equivalent frequency characteristic H eq or equivalent frequency characteristic H eq The frequency characteristics of the correction filter 1620 are determined so as to reduce the influence of

[0119] The correction filter setting unit 1840 is connected to the correction filter determination unit 1830. The correction filter setting unit 1840 calculates parameters for the correction filter 1620 so that the correction filter 1620 has the determined frequency characteristics. The correction filter setting unit 1840 sets the calculated parameters in the correction filter 1620.

[0120] 19 shows an example of the operational flow of the calibration processing unit 1590 according to the modified example of this embodiment. The calibration processing unit 1590 performs the process shown in this drawing in S1750 of FIG.

[0121] In S1910, the frequency characteristic calculation unit 1810 calculates the calibration bandpass signal y i Based on each tone in [n], the discrete equivalent frequency characteristic H^ at the frequency of each tone is calculated. eq In S1920, the interpolation processing unit 1820 calculates a discrete equivalent frequency characteristic Ĥ for each front end i. eq By interpolating the eq Estimate.

[0122] In S1930, the correction filter determination unit 1830 determines the estimated equivalent frequency characteristic H eqBased on this, the correction filter setting unit 1840 determines the frequency characteristics of the correction filter 1620. In S1940, the correction filter setting unit 1840 calculates parameters for the correction filter 1620 so that the correction filter 1620 has the determined frequency characteristics. The correction filter setting unit 1840 sets the calculated parameters in the correction filter 1620.

[0123] 20 shows an example of the baseband spectrum of the calibration input signal X(f) according to a modification of this embodiment. ^ k As shown on the right side of the figure, (f) has a frequency slot with a center frequency of 0 and a repetition frequency width fp, and has tones at each of multiple frequencies arranged asymmetrically around the center frequency.

[0124] Calibration input signal Z^ k At least one tone of (f) is a calibration input signal Z^ at a frequency symmetrical to the frequency of the tone around the center frequency. k (f) has a frequency at which no other tone exists. For example, the calibration input signal Z^ k (f) shows a tone having a frequency 3 / 4Δf in this figure, and does not have another tone at a frequency -3 / 4Δf symmetrical to the frequency of this tone with respect to the center frequency 0. In this modification, the calibration input signal Z^ k All the tones in (f) are calibrated with the calibration input signal Z^ at frequencies symmetrical to the frequency of the tone in question around the center frequency. k The other tones in (f) are defined to have frequencies that are not present.

[0125] Calibration input signal Z^ k Each tone of (f) may be arranged at a predetermined frequency interval within the frequency slot. k The tones in (f) may be arranged at frequency intervals of Δf in a frequency slot having a frequency width fp, and one tone among them may be arranged at a position −¼Δf from the center frequency.

[0126] The left side of the figure shows the baseband calibration input signal Z^ k (f) occurs, and Z^ is centered on the frequency axis. k (f) The conjugate image Z^ of the baseband, where the positive and negative of the frequency of each tone are inverted. -k (f) shows that the tones of the calibration input signal Z^k(f) are asymmetrically positioned with respect to the center frequency, resulting in the conjugate image Z^ -k Each tone of (f) will be located at a different frequency than each tone of the calibration input signal Z^k(f).

[0127] 21 shows an example of the spectrum of the calibration input signal X(f) according to a modification of this embodiment. In S1700 of FIG. 17, the calibration signal supplying unit 1570 supplies the baseband calibration input signal Ẑ shown in FIG. k The calibration input signal X(f) is generated by generating (f) within a frequency slot whose center frequency is an integer multiple (k times in the figure) of the repetition frequency fp of the signal pattern, where k may be an integer greater than or equal to 1, or may be greater than or equal to 2.

[0128] The k-th frequency slot (frequency slot with center frequency kfp) of the calibration input signal X(f) contains the baseband calibration input signal Z^ shown on the right side of Figure 20. k In this modification, the calibration input signal X(f) may be expressed in the time domain by the following equation (9):

number

[0129] where φ ξ is the calibration input signal x LPF is the initial phase of (t), and G is the number of tones in the frequency slot. For example, G is an odd number. fc is the baseband calibration input signal Z^ k is the center frequency of the frequency slot in which (f) occurs, which may be fc=kfp. Δf is the frequency spacing between tones, which may be Δf=fp / G.

[0130] The calibration input signal x shown in equation (9) LPF Each tone of (t) has the same signal strength. Alternatively, at least one tone of the calibration input signal may have a different signal strength than the other tones.

[0131] The calibration signal supplying unit 1570 supplies the calibration input signal x LPF The initial phase φ of each tone is chosen so that the amplitude of (t) is minimized. ξ For example, the calibration signal providing unit 1570 may adjust the initial phase φ of each tone as shown in the following equation (10): ξ is set to the Neuman phase described in Non-Patent Document 9, the calibration input signal x LPF The maximum amplitude of (t) may be limited to within the rated range of the transducer.

number

[0132] The baseband calibration input signal Z^ is input to the kth frequency slot of the calibration input signal X(f). k When (f) is shifted in frequency and placed, as shown in this figure, the -kth frequency slot (frequency slot with center frequency -kfp) of the calibration input signal X(f) contains Z^ as shown on the right side of Figure 20. k Conjugate image Z^ of (f) -k The calibration input signal X(f) does not need to have a tone in any frequency slot other than the k-th and -k-th slots. The calibration input signal X(f) actually input to the mixer 130 has, for example, the signal input characteristic 1410(H RF ) is affected.

[0133] FIG. 22 shows a signal pattern P i 17, the periodic signal generator 120 generates the signal pattern P i (f) occurs.

[0134] As shown with respect to FIG. 4, the signal pattern P i (f) is the frequency component c of frequency l·fp in the range of l=-L0 to L0. i,l Including the signal pattern P i Negative frequency component c in (f) i,-l1 (l1>0) is the positive frequency component c symmetrical with respect to the frequency axis (vertical axis positioned at frequency 0) i,l1 is the conjugate image of the positive frequency component c i,l1 The signal pattern P actually input to the mixer 130 has the same magnitude as i (f) is a PSF input characteristic 1420 (H LO ) is affected.

[0135] 23 shows an example of the spectrum of the signal output by the AD converter 150 according to a modification of this embodiment. In S820 of FIG. 17, the mixer 130 applies the signal pattern P shown in FIG. 22 to the calibration input signal X(f) shown in FIG. i (f). This multiplies Z^ in the frequency slot with kfp as the center frequency in the calibration input signal X(f). k (f―kfp) is P i Each frequency component c contained in (f) i,l and frequency-translated into each frequency slot with a center frequency of (k+l)fp, and i,l Signal c multiplied by i,l Z^ k (f-(k+l)fp).

[0136] As shown in this figure, the calibration input signal Z^ in the frequency slot with kfp as the center frequency k (f-kfp) is the frequency component c i,-(k+1) and frequency-translated into a frequency slot with a center frequency of -fp, resulting in signal c i,-(k+1) Z^ k (f+fp) Calibration input signal Z^ k (f-kfp) is the frequency component c i,-kand frequency-translated into a frequency slot with a center frequency of 0, resulting in signal c i,-k Z^ k (f) Also, the calibration input signal Z^ k (f-kfp) is the frequency component c i,-(k-1) and frequency-translated into a frequency slot with a center frequency of fp, resulting in a signal c i,-(k-1) Z^ k The result is (f-fp).

[0137] Similarly, the calibration input signal Z^ in the frequency slot with -kfp as the center frequency -k (f+kfp) is the frequency component c i,k-1 and frequency-translated into a frequency slot with a center frequency of -fp, resulting in signal c i,k-1 Z^ -k (f+fp) Calibration input signal Z^ -k (f+kfp) is the frequency component c i,k and frequency-translated into a frequency slot with a center frequency of 0, resulting in signal c i,k Z^ -k (f) Also, the calibration input signal Z^ k (f-kfp) is the frequency component c i,k+1 and frequency-translated into a frequency slot with a center frequency of fp, resulting in a signal c i,k+1 Z^ -k The result is (f-fp).

[0138] For the sake of convenience, only frequency slots with center frequencies of 0 and ±1·fp are shown in this figure. In reality, the calibration input signal X(f) is a signal pattern P i Frequency component c of (f) i,-L0 From c i,L0 and the center frequency is converted to each frequency slot from -(L0+k)·fp to (L0+k)·fp.

[0139] In this way, the calibration input signal X(f) and the signal pattern P i The signal multiplied by (f) is the baseband calibration input signal Z^ in each angular frequency slot.k (f) is frequency-converted into multiple tones, and Z^ is centered on the center frequency. k (f) is the inverted conjugate image Z^ -k (f) and multiple tones obtained by frequency-translating the baseband calibration input signal Z^ k (f) has multiple tones asymmetrically positioned with respect to the conjugate image inverted around the center frequency, so that the calibration input signal Ẑ in each frequency slot k Each tone from (f) is the conjugate image Z^ -k (f) has a different frequency from each tone.

[0140] For example, in the frequency slot with a center frequency of 0 in this figure, the calibration input signal Z^ k The tones in (f) are located at frequencies , -Δf / 4, 3Δf / 4, and the conjugate image Z^ -k The tones in (f) are placed at frequencies of , -Δ3f / 4, Δf / 4, . Therefore, the calibration input signal Z^ k (f) Each tone and its conjugate image Z^ -k There is a frequency difference of Δf / 2 between adjacent tones from (f).

[0141] In S830 of FIG. 17, the bandpass filter 140 converts the signal yp output by the mixer 130 into i In this modification, the band-pass filter 140 is, for example, a low-pass filter, and performs low-pass filtering on the signal output from the mixer 130. In S1740 of FIG. 17, the AD converter 150 performs low-pass filtering on the calibration band-pass signal y (t) band-limited by the band-pass filter 140. i (t) is sampled to generate a digital calibration bandpass signal y i [n]. Digital calibration bandpass signal y i The frequency domain calibration bandpass signal Y corresponding to [n] i (f) is the equivalent frequency characteristic H eq The actual calibration bandpass signal Y i(f) is Y shown in the following equation (11) ~ i It is represented by (f).

number

[0142] In S1740 of FIG. 17, the correction filter unit 1555 generates the digital calibration bandpass signal y i [n] and outputs it. For convenience of explanation, it is assumed that the low-pass filter 1610 in the correction filter unit 1555 is an ideal low-pass filter. It is also assumed that the correction filter 1620 has a gain of 1 and a phase delay of 0 regardless of frequency, and is initialized to output the input signal as is. In this case, the calibration processing unit 1900 receives the signal c via the calibration band-pass signal acquisition unit 180 as shown in the figure. i,l Z^ k (f-(k+l)fp) and signal c i,l ·Z^- k Equivalent frequency characteristics H, including (f-(-k+l)fp) eq Calibration bandpass signal Y affected by ~ i (f) and the calibration bandpass signal Y ~ i (f) is low-pass filtered by low-pass filter 1610.

[0143] In S1750 of FIG. 17, the calibration processing unit 1590 calculates the calibration bandpass signal Y ~ i First, in step S1910 of FIG. 19, the frequency characteristic calculation unit 1810 in the calibration processing unit 1590 calculates the calibration bandpass signal y i Based on each tone in [n], the discrete equivalent frequency characteristic H^ at the frequency f^ of each tone is calculated. eq Calculate [f^].

[0144] (1) When the influence of PSF input characteristic 1420 is not considered If the influence of the PSF input characteristic 1420 is small and can be ignored, the signal pattern P i Actual frequency component c of (f) ~ i,l = ideal frequency component c i,l By substituting the above, the following equation (12) can be obtained: Note that in transforming equation (11) into equation (12), the influence of signal input characteristics 1410 is also ignored.

number

[0145] f ^ is the calibration bandpass signal Y ~ i (f) is the frequency at which the tone exists, and in Figure 23, ^ =-Δf / 4+n Δf / 2 (where n is Y ~ i (f) is an integer within the range in which a tone exists in the signal pattern P i (f) Ideal frequency component c i,l is the signal pattern P i (f) can be calculated. Also, the ideal calibration input signal Z k (f) is also known. Therefore, the frequency characteristic calculation unit 1810 calculates the calibration bandpass signal Y ~ i (f) for each frequency f ^ H in eq (f) Discrete equivalent frequency characteristic H eq [f ^ ] can be calculated.

[0146] As shown in FIG. 23, the calibration bandpass signal Y ~ i In each frequency slot of (f), the baseband calibration input signal Z^ k (f) Each tone corresponds to the conjugate image Z^ -k (f) has a different frequency from each tone corresponding to the calibration bandpass signal Y ~i In (f), the baseband calibration input signal Z^ k (f) Each tone corresponds to the conjugate image Z^ -k (f) can be distinguished from each corresponding tone.

[0147] For example, substituting f^=3·Δf / 4 into equation (12) gives the calibration input signal Z as shown in Figure 20. k Conjugate image Z of (f^-lp) (where l=0) -k (f^-lp) is 0 because there is no tone at this frequency f^ in the frequency slot. Therefore, when f^=3·Δf / 4, equation (12) becomes H eq [3 Δf / 4]=Y ~ i (3·Δf / 4) / {c i,-k Z k (3·Δf / 4)}. Also, f^ is the center frequency lf p Z included in the frequency slot k If it corresponds to the tone of (f), H eq [f^]=Y ~ i (f^) / {c i,-(k-l) Z k (f^-lf p )}.

[0148] Therefore, the frequency characteristic calculation unit 1810 calculates the calibration bandpass signal Y ~ i The signal strength of a tone To in (f^) is calculated by dividing the signal strength of the original tone Ti in the calibration input signal X(f) that has been frequency-converted to tone To by the signal pattern P i In (f), the frequency component c obtained by frequency converting tone Ti to tone To i,α The signal strength is divided by the equivalent frequency characteristic H eq The frequency characteristic calculation unit 1810 can calculate the gain of the calibration bandpass signal Y ~ i From the phase of a certain tone To in (f^), the phase of the original tone Ti in the calibration input signal X(f) that has been frequency converted to tone To and the signal pattern P iIn (f), the frequency component c obtained by frequency converting tone Ti to tone To i,α By subtracting the phase of and the equivalent frequency characteristic H eq In this way, the frequency characteristic calculation unit 1810 calculates the phase of the calibration bandpass signal Y ~ i Depending on the signal strength and phase between the multiple tones included in (f), the discrete equivalent frequency characteristic H eq The gain and phase of [f^] can be calculated.

[0149] (2) When considering the influence of PSF input characteristics 1420 When the influence of the PSF input characteristic 1420 is taken into consideration, the signal pattern P i Actual frequency component c of (f) ~ i,l is unknown. And, c in Eq. (12) i,-(k-l) and c i,-(k+l) is the actual frequency component c ~ i,-(k-l) and c ~ i,-(k+l) Therefore, the direct equivalent frequency characteristic H eq [f^] cannot be calculated.

[0150] However, even in this case, as shown in FIG. 23, within the same frequency slot, the positive side spectrum Z k The frequency component c multiplied by each tone in (f) ~ i,-(k-l) are the same, and the negative spectrum Z -k The frequency component c multiplied by each tone in (f) ~ i,k-l is also the same.

[0151] Therefore, the frequency characteristic calculation unit 1810 calculates the calibration bandpass signal Y ~ iAccording to the signal strength and phase changes between the multiple tones included in each frequency slot in (f), the discrete equivalent frequency characteristic H eq Calculate the relative change in gain and phase of [f^]. For example, in Figure 23, H^ eq [3 Δf / 4]=Y ~ i (3·Δf / 4) / {c i,-k Z k (3·Δf / 4)} has a gain of 1.0, H^ eq [7·Δf / 4]=Y ~ i (7·Δf / 4) / {c i,-k Z k (7·Δf / 4)} is 0.9. k If the signal strength of each tone in (f) is constant regardless of frequency, the frequency characteristic calculation unit 1810 calculates H^ eq The gain of [7·Δf / 4] is H^ eq The frequency characteristic calculation unit 1810 can calculate that the gain is 0.9 times the gain of [3·Δf / 4]. Similarly, the frequency characteristic calculation unit 1810 calculates the phase characteristic as follows: ~ i From the phase of (3·Δf / 4) to Z k H^ with a phase subtraction of (3·Δf / 4) eq The phase of [3·Δf / 4] and Y ~ i From the phase of (7·Δf / 4) to Z k H^ with a phase subtraction of (7·Δf / 4) eq The phase difference between the phases of [7·Δf / 4] can be calculated.

[0152] In both of the above (1) and (2), the frequency characteristic calculation unit 1810 calculates the original spectrum Z of the calibration input signal X(f) in the calibration bandpass signal Y~i(f). k (f) and the conjugate image Z -k (f) Using both tones, the discrete equivalent frequency characteristic H^ eq Alternatively, the discrete equivalent frequency characteristic H^ can be calculated using tones from only one of the two. eq may be calculated.

[0153] In S1920 of FIG. 19, the interpolation processing unit 1820 calculates the discrete equivalent frequency characteristic H^ eq By interpolating the eq When the influence of the PSF input characteristic 1420 is not taken into consideration (the case of (1) above), the interpolation processing unit 1820 estimates the discrete equivalent frequency characteristic Ĥ calculated at intervals of frequency Δf or Δf / 2 in this modification. eq The gain and phase characteristics are interpolated to obtain a continuous equivalent frequency characteristic H eq The interpolation unit 1820 may use linear interpolation, polynomial interpolation, spline interpolation, or any other method as the interpolation method.

[0154] Even when the influence of the PSF input characteristic 1420 is taken into consideration (the case (2) above), the interpolation processing unit 1820 calculates the calibration bandpass signal Y ~ i In (f), the equivalent frequency characteristic H eq As a result, the interpolation processing unit 1820 estimates changes in the gain and phase characteristics of the equivalent frequency characteristic H eq can be estimated.

[0155] The interpolation processing unit 1820 also calculates the calibration bandpass signal Y ~ i At the boundary between an adjacent frequency slot (referred to as the "first frequency slot") and the adjacent frequency slot (referred to as the "second frequency slot") in (f), the equivalent frequency characteristic H eq and the equivalent frequency characteristic H eq For example, in FIG. 23, the interpolation processing unit 1820 determines a frequency slot whose center frequency is 0 as the first frequency slot, and performs an adjustment to bring the first equivalent frequency characteristic H eqand a second equivalent frequency characteristic H that is continuous within the second frequency slot with the center frequency fp. eq At this stage, the first equivalent frequency characteristic H eq and the second equivalent frequency characteristic H eq Therefore, for example, the interpolation processing unit 1820 calculates the second equivalent frequency characteristic H eq The gain and phase characteristics of the first equivalent frequency characteristic H eq The gain and phase characteristics are adjusted to match those of the second equivalent frequency characteristic H eq The interpolation processing unit 1820 regards a frequency slot whose center frequency is other than 0 as a first frequency slot, and calculates an equivalent frequency characteristic H eq may be adjusted to be closer to

[0156] As a result, the interpolation processing unit 1820 sequentially generates the equivalent frequency characteristic H eq By adjusting the eq It should be noted that the interpolation processing unit 1820 may use a frequency slot whose center frequency is other than 0 as the starting point of this processing.

[0157] In S1930 of FIG. 19, the correction filter determination unit 1830 determines the equivalent frequency characteristic H eq The frequency characteristic H of the correction filter 1620 is set to C The correction filter determination unit 1830 determines the frequency characteristic H C is the cutoff frequency f of the band limit in the band-pass filter 140 and the low-pass filter 1610. bpf For each frequency slot up to eq Here, the correction filter determination unit 1830 determines the frequency characteristic HC may be determined by the following equation (13):

number

[0158] In this way, the calibration processing unit 1590 calculates the frequency characteristic Hc(f) by the estimated H eq (f) and the known frequency characteristic H of the low-pass filter 1610 S In this modification, the calibration bandpass signal Y ~ i (f) is the signal that has passed through the correction filter unit 1555. Therefore, the frequency characteristic H C If the gain and phase are set to values ​​other than 1 and 0, the compensation filter 1620 uses the set frequency characteristic H C Then, the estimated equivalent frequency characteristic H eq Multiplying the inverse characteristic of (f) produces a new frequency characteristic H C may be determined.

[0159] In S1940 of FIG. 19, the correction filter setting unit 1840 sets the correction filter 1620 to the determined frequency characteristic H C For example, if the correction filter 1620 is a FIR (Finite Impulse Response) filter, the correction filter setting unit 1840 calculates the parameters of the correction filter 1620 such that the determined frequency characteristic H C The correction filter setting unit 1840 sets the calculated filter coefficients in the correction filter 1620.

[0160] According to the conversion device 1500 described above, the actual frequency characteristics of the band-pass filters 140 and the like at the front ends can be compensated for by the correction filter unit 1555. Furthermore, according to the conversion device 1500, by supplying the calibration input signal shown in FIG. 21 to the converter, the equivalent frequency characteristics H eqcan be estimated collectively, and the equivalent frequency characteristic H eq The frequency characteristics of the correction filter 1620 in the correction filter unit 1555 can be set based on the inverse characteristics of the above.

[0161] 24 shows an example of the spectrum of the output signal before and after correction according to a modification of this embodiment. This figure shows the spectrum of the calibration input signal x LPF 15 shows the results of a simulation of the spectrum of the output signal xo(t) reconstructed from (t) by the conversion device 1500. Note that the number m of front ends is 4, and the sampling frequency fs of the AD converter 150 is 16 times (q=16) the frequency fp of the signal pattern.

[0162] When correction by the correction filter unit 1555 is not performed, the spectrum of the output signal xo(t) contains conjugate images (spurious signals) of up to about −45 dB depending on the actual frequency characteristics of the band-pass filter 140, etc. In contrast, after the correction filter 1620 is calibrated by the calibration processing unit 1590, the conversion device 1500 is able to suppress the spurious signals to about −76 dB.

[0163] FIG. 25 shows the calibration input signal X before and after correction according to a modification of this embodiment. LPF 15 shows an example of the spectrum of the output signal output by the AD converter 150 and the correction filter unit 1555 in response to the input of (f). This figure shows the results of measurements using an actual circuit in which the circuit of the converter part of the conversion device 1500 is implemented on-board. Here, the signal pattern p i The repetition frequency fp of (t) is 1 MHz, the number of frequency slots L0 is 160, the frequency bandwidth to be detected for the input signal x(t) is 160 MHz, and the signal pattern p i The number of symbols in one period of (t) is 500, the number m of front ends is 4, the sampling frequency fs of the AD converter 150 is 100 MHz, and the cutoff frequency of the bandpass filter 140 is 30 MHz.

[0164] When no correction is made by the correction filter unit 1555, as shown on the left side of the figure, in the frequency slot of 0 to 0.5 MHz, a decrease in the signal strength of the output signal Xo(f) is observed at low frequencies due to a decrease in gain of the bandpass filter 140, etc. Also, in each of the frequency slots of 0.5 to 1.5 MHz, 1.5 to 2.5 MHz, and 2.5 MHz and above, the gain gradually decreases as the frequency increases.

[0165] In contrast, the output signal Xo(f) after the correction filter 1620 is calibrated by the calibration processing unit 1590 has a substantially constant signal strength for each frequency slot, regardless of the frequency, as shown on the right side of the figure. LPF (f) Baseband spectrum Z k (f) and Z -k For (f), signal pattern P i (f) Different frequency components c for each frequency slot i,l Since the output signal Xo(f) is obtained by multiplying the frequency slots, the boundaries of the frequency slots in the output signal Xo(f) are discontinuous.

[0166] 26 shows an example of the frequency characteristics of a low-pass filter (LPF) as an example of the band-pass filter 140 and the correction filter 1620 according to a modification of this embodiment. This diagram, with frequency on the horizontal axis and gain and phase on the vertical axis, shows the frequency characteristics of the actual band-pass filter 140 ("LPF" in the diagram), the frequency characteristics of the ideal correction filter 1620 ("desired correction filter" in the diagram), and the frequency characteristics of the correction filter 1620 determined by the actual conversion device 1500 ("actual correction filter" in the diagram).

[0167] 25, the gain of the actual bandpass filter 140 decreases at low frequencies below approximately 0.3 MHz. The gain of the actual bandpass filter 140 also decreases gradually at frequencies above approximately 0.3 MHz. Furthermore, the actual bandpass filter 140 has phase characteristics that vary with frequency, as shown in the lower graph of the figure.

[0168] As shown in the figure, the conversion device 1500 according to this modification can calibrate the correction filter 1620 so that it has frequency characteristics that are approximately the same as those of an ideal correction filter that offsets the frequency characteristics inherent to the actual band-pass filter 140. In this way, the conversion device 1500 according to this modification can calibrate the correction filter 1620 so that it has frequency characteristics that are approximately the same as those of an ideal correction filter that offsets the frequency characteristics inherent to the actual band-pass filter 140. i Even if the repetition frequency of (t) is higher than that of (t), that is, for example, fs=q·fp (q is an integer greater than 1), the equivalent frequency characteristic H eq The correction filter 1620 can be adjusted to cancel out

[0169] Note that the calibration bandpass signal acquisition unit 180, the calibration signal supply unit 1570, and the calibration processing unit 1590 of the conversion device 1500 may calibrate the correction filter 1620 as shown in relation to Figures 12 to 26, and then calibrate the reconstruction unit 160 as shown in relation to Figures 1 to 11. This allows the conversion device 1500 to adjust the reconstruction parameters in the reconstruction unit 160 using multiple calibration bandpass signals from each front end that have been made closer to ideal by the correction filter unit 1555.

[0170] Conversely, the calibration bandpass signal acquisition unit 180, the calibration signal supply unit 1570, and the calibration processing unit 1590 of the conversion device 1500 may calibrate the correction filter 1620 as shown in relation to Figures 12 to 26 after calibrating the reconstruction unit 160 as shown in relation to Figures 1 to 11. As described in relation to Figure 9, the calibration processing unit 1590 may calibrate the correction filter 1620 using the signal pattern P iEach frequency component c of (f) i,l The calibration processing unit 1590 can detect the actual values ​​(signal strength and phase) of each frequency component c detected during calibration of the reconstruction unit 160. i,l By using the actual value of , the equivalent frequency characteristic H can be calculated by taking into account the influence of the PSF input characteristic 1420 while using the method described in "(1) When the influence of the PSF input characteristic 1420 is not considered" in relation to FIG. 23. eq The calibration processing unit 1590 may further improve the accuracy of the correction filter 1620 and the reconstruction unit 160 by repeatedly calibrating the correction filter 1620 and the reconstruction unit 160.

[0171] The calibration processing unit 1590 described above adjusts both the gain characteristic and the phase characteristic of the correction filter 1620. Alternatively, the calibration processing unit 1590 may adjust only one of the gain characteristic and the phase characteristic of the correction filter 1620.

[0172] Various embodiments of the present invention may be described with reference to flowcharts and block diagrams, where the blocks may represent (1) stages of a process in which operations are performed or (2) sections of an apparatus responsible for performing the operations. Particular stages and sections may be implemented by dedicated circuitry, programmable circuitry provided with computer-readable instructions stored on a computer-readable medium, and / or a processor provided with computer-readable instructions stored on a computer-readable medium. Dedicated circuitry may include digital and / or analog hardware circuitry, and may include integrated circuits (ICs) and / or discrete circuits. Programmable circuitry may include reconfigurable hardware circuitry, including logical AND, OR, XOR, NAND, NOR, and other logical operations, flip-flops, registers, memory elements such as field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), and the like.

[0173] A computer-readable medium may include any tangible device capable of storing instructions that are executed by an appropriate device, such that the computer-readable medium having instructions stored thereon comprises an article of manufacture containing instructions that can be executed to create means for performing the operations specified in the flowcharts or block diagrams. Examples of computer-readable media may include electronic, magnetic, optical, electromagnetic, and semiconductor storage media. More specific examples of computer-readable media may include floppy disks, diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), electrically erasable programmable read-only memory (EEPROM), static random access memory (SRAM), compact disc read-only memory (CD-ROM), digital versatile disc (DVD), Blu-ray disc, memory stick, integrated circuit card, and the like.

[0174] The computer readable instructions may include either assembler instructions, Instruction Set Architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state-setting data, or source or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk®, JAVA®, C++, etc., and conventional procedural programming languages ​​such as the “C” programming language or similar programming languages.

[0175] The computer-readable instructions may be provided to a processor or programmable circuitry of a programmable data processing apparatus, such as a general-purpose computer, special-purpose computer, or other computer, either locally or over a wide-area network (WAN) such as a local area network (LAN), the Internet, etc., which executes the computer-readable instructions to create means for performing the operations specified in the flowcharts or block diagrams. Examples of processors include computer processors, processing units, microprocessors, digital signal processors, controllers, microcontrollers, etc.

[0176] 27 illustrates an example of a computer 2200 in which aspects of the present invention may be embodied, in whole or in part. Programs installed on the computer 2200 may cause the computer 2200 to function as or perform operations associated with an apparatus or one or more sections of the apparatus according to embodiments of the present invention, and / or to perform a process or steps of a process according to embodiments of the present invention. Such programs may be executed by the CPU 2212 to cause the computer 2200 to perform specific operations associated with some or all of the blocks of the flowcharts and block diagrams described herein.

[0177] A computer 2200 according to this embodiment includes a CPU 2212, a RAM 2214, a graphics controller 2216, and a display device 2218, which are interconnected by a host controller 2210. The computer 2200 also includes input / output units such as a communication interface 2222, a hard disk drive 2224, a DVD-ROM drive 2226, and an IC card drive, which are connected to the host controller 2210 via an input / output controller 2220. The computer also includes legacy input / output units such as a ROM 2230 and a keyboard 2242, which are connected to the input / output controller 2220 via an input / output chip 2240.

[0178] The CPU 2212 operates according to programs stored in the ROM 2230 and RAM 2214, thereby controlling each unit. The graphics controller 2216 acquires image data generated by the CPU 2212 into a frame buffer or the like provided in the RAM 2214 or into the graphics controller 2216 itself, and causes the image data to be displayed on the display device 2218.

[0179] The communications interface 2222 communicates with other electronic devices via a network. The hard disk drive 2224 stores programs and data used by the CPU 2212 in the computer 2200. The DVD-ROM drive 2226 reads programs or data from the DVD-ROM 2201 and provides the programs or data to the hard disk drive 2224 via the RAM 2214. The IC card drive reads programs and data from an IC card and / or writes programs and data to an IC card.

[0180] The ROM 2230 stores therein a boot program or the like that is executed by the computer 2200 upon activation, and / or programs that depend on the hardware of the computer 2200. The input / output chip 2240 may also connect various input / output units to the input / output controller 2220 via a parallel port, a serial port, a keyboard port, a mouse port, etc.

[0181] The programs are provided by a computer-readable medium such as a DVD-ROM 2201 or an IC card. The programs are read from the computer-readable medium, installed in the hard disk drive 2224, RAM 2214, or ROM 2230, which are also examples of computer-readable media, and executed by the CPU 2212. Information processing described in these programs is read by the computer 2200, and brings about cooperation between the programs and the various types of hardware resources described above. An apparatus or method may be configured by realizing information manipulation or processing in accordance with the use of the computer 2200.

[0182] For example, when communication is performed between the computer 2200 and an external device, the CPU 2212 may execute a communication program loaded into the RAM 2214 and instruct the communication interface 2222 to perform communication processing based on the processing described in the communication program. Under the control of the CPU 2212, the communication interface 2222 reads transmission data stored in a transmission buffer processing area provided in the RAM 2214, the hard disk drive 2224, the DVD-ROM 2201, or a recording medium such as an IC card, and transmits the read transmission data to the network, or writes reception data received from the network to a reception buffer processing area or the like provided on the recording medium.

[0183] The CPU 2212 may also cause all or a necessary portion of a file or database stored on an external recording medium such as the hard disk drive 2224, the DVD-ROM drive 2226 (DVD-ROM 2201), an IC card, etc. to be read into the RAM 2214, and perform various types of processing on the data on the RAM 2214. The CPU 2212 then writes back the processed data to the external recording medium.

[0184] Various types of information, such as various types of programs, data, tables, and databases, may be stored on the recording medium and may undergo information processing. The CPU 2212 may perform various types of processing on data read from the RAM 2214, including various types of operations, information processing, conditional judgment, conditional branching, unconditional branching, information search / replacement, etc., as described throughout this disclosure and specified by the instruction sequences of the programs, and write the results back to the RAM 2214. The CPU 2212 may also search for information in a file, database, etc. on the recording medium. For example, if multiple entries each having an attribute value of a first attribute associated with an attribute value of a second attribute are stored on the recording medium, the CPU 2212 may search for an entry that matches a condition specified by the attribute value of the first attribute from among the multiple entries, read the attribute value of the second attribute stored in the entry, and thereby obtain the attribute value of the second attribute associated with the first attribute that satisfies a predetermined condition.

[0185] The above-described programs or software modules may be stored in a computer-readable medium on or near the computer 2200. A recording medium such as a hard disk or RAM provided in a server system connected to a dedicated communication network or the Internet can also be used as a computer-readable medium, thereby providing the programs to the computer 2200 via the network.

[0186] Although the present invention has been described above using embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments. It will be apparent to those skilled in the art that various modifications and improvements can be made to the above embodiments. It is clear from the claims that such modifications and improvements can also be included within the technical scope of the present invention.

[0187] It should be noted that the execution order of each process, such as operations, procedures, steps, and stages, in the devices, systems, programs, and methods shown in the claims, specifications, and drawings is not specifically stated as "before," "prior to," etc., and that the processes can be performed in any order unless the output of a previous process is used in a subsequent process. Even if the operational flow in the claims, specifications, and drawings is described using "first," "next," etc. for convenience, this does not mean that the processes must be performed in this order. [Explanation of symbols]

[0188] 100 conversion device, 110 selector, 120 periodic signal generator, 130 mixer, 140 band-pass filter, 150 AD converter, 160 reconstruction unit, 170 calibration signal supply unit, 180 calibration band-pass signal acquisition unit, 190 calibration processing unit, 310 signal, 320 signal, 1000 calibration signal supply unit, 1010 trigger generator, 1020 sine wave generator, 1030 attenuator, 1040 multiplexer, 1050 phase amplitude setting unit, 1100 calibration processing unit, 1110 signal component detection unit, 1120 calibration parameter calculation unit, 1400 ideal mixer, 1410 signal input characteristics, 1420 PSF input characteristics, 1430 mixer output characteristics, 1440 ideal AD converter, 1450 AD conversion characteristics, 1500 conversion device, 1555 correction filter section, 1570 calibration signal supply section, 1590 calibration processing section, 1600 upsampler, 1610 low-pass filter, 1620 correction filter, 1630 downsampler, 1810 frequency characteristic calculation section, 1820 interpolation processing section, 1830 correction filter determination section, 1840 correction filter setting section, 2200 computer, 2201 DVD-ROM, 2210 host controller, 2212 CPU, 2214 RAM, 2216 graphics controller, 2218 display device, 2220 input / output controller, 2222 communication interface, 2224 hard disk drive, 2226 DVD-ROM drive, 2230 ROM, 2240 input / output chip, 2242 keyboard

Claims

1. a calibration signal supplying unit that supplies a multi-tone signal having tones at a plurality of frequencies as a calibration input signal to a converter that obtains a plurality of band-pass signals by multiplying an input signal by each of a plurality of signal patterns to band-limit the input signal, and reconstructs an output signal corresponding to the input signal from the plurality of band-pass signals; a calibration bandpass signal acquisition unit that acquires a plurality of calibration bandpass signals obtained by the converter in response to the calibration input signal; a calibration processing unit that calibrates a correction filter that corrects each of the plurality of band-pass signals based on each of the plurality of calibration band-pass signals; A calibration device comprising:

2. 2. The calibration device according to claim 1, wherein the calibration input signal has a center frequency that is an integer multiple of a repetition frequency of the plurality of signal patterns, and has tones at each of the plurality of frequencies arranged asymmetrically with respect to the center frequency within a frequency slot having a width of the repetition frequency.

3. 3. The calibration device according to claim 2, wherein each tone of the calibration input signal has a frequency that is symmetrical to the frequency of the tone about the center frequency and at which no other tone of the calibration input signal exists.

4. 4. The calibration device according to claim 2, wherein the tones of the calibration input signal are arranged at predetermined frequency intervals within the frequency slot.

5. 5. A calibration device according to claim 2, wherein each tone of the calibration input signal has the same signal strength.

6. The calibration processing unit estimating, in accordance with a change in signal strength among a plurality of tones included in one frequency slot of a first calibration band-pass signal among the plurality of calibration band-pass signals, a frequency characteristic of at least a part of a path from the calibration input signal to obtaining the first calibration band-pass signal within the one frequency slot; The correction filter that corrects the first calibration bandpass signal is calibrated to a characteristic corresponding to an inverse characteristic of the frequency characteristic within the one frequency slot. Calibration device according to any one of claims 2 to 5.

7. The calibration device according to claim 6, wherein the calibration processing unit calibrates the correction filter that corrects the first calibration bandpass signal to have characteristics corresponding to the inverse characteristics of the frequency characteristics in each frequency slot up to the cutoff frequency of the band limit.

8. 8. The calibration device according to claim 6, wherein the calibration processing unit adjusts the frequency characteristics of the path on the first frequency slot side and the frequency characteristics of the path on the second frequency slot side at the boundary between adjacent first and second frequency slots in the first calibration bandpass signal to bring them closer together.

9. a converter that obtains a plurality of band-pass signals by multiplying an input signal by each of a plurality of signal patterns to band-limit the signal, and reconstructs an output signal corresponding to the input signal from the plurality of band-pass signals; A calibration device according to any one of claims 1 to 8. A conversion device comprising:

10. The converter comprises: a plurality of mixers that multiply the input signal by a plurality of signal patterns, respectively; a plurality of bandpass filters that band-limit each of the plurality of signals output from the plurality of mixers; a plurality of AD converters that output the plurality of band-pass signals obtained by sampling signals that have passed through the plurality of band-pass filters; a plurality of correction filter units that correct the plurality of band-pass signals and output a plurality of corrected band-pass signals; a reconstruction unit that reconstructs the output signal from the plurality of corrected bandpass signals; 10. The conversion device according to claim 9, comprising:

11. Each of the plurality of correction filter units a low-pass filter that low-pass filters a band-pass signal input to the correction filter unit out of the plurality of band-pass signals; a correction filter that corrects the band-pass signal that has passed through the low-pass filter; 11. The conversion device according to claim 10, comprising:

12. a converter that obtains a plurality of band-pass signals by multiplying an input signal by each of a plurality of signal patterns to band-limit the input signal, and reconstructs an output signal corresponding to the input signal from the plurality of band-pass signals, and supplies a multi-tone signal having tones at a plurality of frequencies as a calibration input signal; obtaining a plurality of calibration bandpass signals obtained by the converter in response to the calibration input signals; calibrating a correction filter that corrects each of the plurality of band-pass signals based on each of the plurality of calibration band-pass signals; Calibration methods.

13. The method is executed by a computer, causing the computer to: a calibration signal supplying unit that supplies a multi-tone signal having tones at a plurality of frequencies as a calibration input signal to a converter that obtains a plurality of band-pass signals by multiplying an input signal by each of a plurality of signal patterns to band-limit the input signal, and reconstructs an output signal corresponding to the input signal from the plurality of band-pass signals; a calibration bandpass signal acquisition unit that acquires a plurality of calibration bandpass signals obtained by the converter in response to the calibration input signal; a calibration processing unit that calibrates a correction filter that corrects each of the plurality of band-pass signals based on each of the plurality of calibration band-pass signals; A calibration program that functions as a

Images