JP3243664B2 - Sample circuit and correlation operation device using the same - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a correlation calculating device, and more particularly to a correlation calculating device used for a radio interferometer or the like.

[0002]

2. Description of the Related Art Generally, in the field of radio astronomy, a radio interferometer is widely used as a technique for obtaining an image having a high spatial resolution. This radio interferometer performs image composition by arranging a large number of element antennas one-dimensionally or two-dimensionally and obtaining a radio wave radiation intensity distribution at each position of an observation target (for example, the sun).

[0003] A typical example of such an image combining method is a Fourier combining method. This is a method that utilizes the fact that a Fourier transform relationship is established between the distribution of the radio wave intensity of the observation target and the correlation function of the received signal between two points (between antennas) on the observation surface. Is a method in which the correlation value (integral value) of the received signal is obtained by one measurement at the same time, and the radio wave image can be synthesized each time.

Therefore, in order to effectively perform image synthesis, it is necessary to obtain the correlation value with sufficiently high accuracy.
Usually, the correlation value is obtained by converting a received signal into a digital value and integrating the correlation result at each sampling point.

Conventionally, when a signal in a video band is converted into a binary signal, a high-speed operational amplifier is used as a comparator to apply a reference voltage to an inverting terminal. A reference voltage was applied to the comparison point for use.

[0006]

However, in the above-mentioned conventional method, when the received signal is a noisy signal, sampling is performed, so that the obtained correlation value varies. This variation is determined by the frequency band of the noisy signal and the integration time, and when viewed temporally, it occurs randomly for each combination of correlations, so the difference from the true correlation value is the cause of degradation during imaging. Becomes

Here, if the number of digitized bits for each sampling is limited to 1, the integration time must be sufficiently long in order to obtain a correlation value between noise signals with high accuracy by a conventional method, and a high time resolution (1 However, there is a drawback that it is not possible to obtain the time required to capture a radio wave image.
In particular, when the observation target changes rapidly in a short time (for example, an impulse-like burst generated in the sun), a sufficiently high time resolution is required for the change of the phenomenon in order to elucidate the phenomenon in detail. Nevertheless, there was a drawback that it could not be raised.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned drawbacks, and an object of the present invention is to suppress variations in data caused by sampling when performing 1-bit A / D conversion of a noise signal in a video band. To provide a sample circuit capable of accurately obtaining a correlation value between signals with high time resolution and a correlation operation device using the same.

[0009]

A sample circuit according to the present invention includes an integrating circuit for integrating a noisy input signal, and a transfer circuit for matching the phase of the integrated output with the phase of the input signal.
A phase circuit and the integrated output after the phase shift as a reference value.
A sampler for sampling a force signal, and a sampler for the sampler
Means for counting the pull output in a predetermined cycle;
Compensation circuit for compensating the reference voltage of the integration circuit according to the reset value.
And a road.

[0010]

[0011]

Next, the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of a first embodiment of a sample circuit according to the present invention. In the figure, a sample circuit according to a first embodiment of the present invention includes an amplifier 1 for amplifying an input noise signal 100, a capacitor 2 for removing a DC component, and A / D conversion of an input signal. And a sampler 3 for performing the operation.

The sample circuit of this embodiment includes an integrating circuit 40 for integrating the input signal after removing the DC component, and a phase shifting circuit 70 for matching the phase of the integrated output with the phase of the input signal. It is composed of

The integrating circuit 40 is composed of a low offset operational amplifier 4, a time constant resistor 5 and a capacitor 6, and is a circuit for extracting a low frequency component of the input signal.

The phase shift circuit 70 comprises a low offset operational amplifier 7 and resistors 8 and 9 for determining the gain of the inverting amplifier.
And a variable capacitor 10 for phase adjustment.

In such a configuration, it is assumed that the reference voltage for binary conversion in the sampler 3 is originally a ground point. Then, the reference voltage point in the sampler 3 is changed by the integration circuit 40 and the phase shift circuit 70.

The noise signal 1 input to the sample circuit
After being amplified by the amplifier 1, the DC component is removed by the capacitor 2 and enters the sampler 3. Further, the input signal is integrated by the integration circuit 40, and as a result, the influence of the low frequency component appears as a undulating waveform.

Among these low-frequency waveform components, those whose period is longer than the integration time for obtaining a correlation value cause variation during sampling. Therefore, the difference between the zero point (the point at which the integrated value becomes zero) of the waveform of the input noise signal and the zero point of the low-frequency component is calculated by the integration circuit 4.
Compensate based on 0 output waveform. At this time, the phase shift circuit 70
Thus, the phase of the low-frequency component of the signal entering the sampler 3 is matched with the phase of the integrated output waveform applied to the comparator point, and the relationship between the two is brought close to the best state. For example, the integral output waveform may be delayed by one cycle.

FIG. 2 is a block diagram showing the configuration of a second embodiment of the sample circuit according to the present invention, and the same parts as those in FIG. 1 are denoted by the same reference numerals.

In the figure, the sample circuit of this embodiment is
A counter circuit 11 and a latch circuit 1 are added to the circuit configuration of FIG.
2, a PROM 13, a D / A converter 14, a low offset operational amplifier 15, resistors 16 to 19, and a capacitor 20 are added. These are added to adjust the reference voltage of the operational amplifier 4 in the integration circuit 40.

That is, in the integrating circuit 40 shown in FIG. 1, the reference voltage is at the ground point. In this embodiment, the reference voltage is adjusted.

In such a configuration, the value "1" of the output data of the sampler 3 is counted by the counter circuit 11 at every period sufficiently shorter than an integration time for obtaining a correlation value by a correlator described later. The count result is stored in the latch circuit 1
2 and the stored data is input as an address of the PROM 13. A value proportional to the difference between the output data of the sampler 3 and the ideal count value (1/2 of the number of samples in one sampling cycle) is written in the PROM 13 in advance. Is extracted as a compensation voltage by the current output D / A converter 14 and the operational amplifier 15.

This compensation voltage compensates for the difference between the zero point of the waveform of the noise signal and the zero point of the high frequency component which cannot be removed by the circuit configuration of FIG. 1 when the band of the noise signal is wide. . Then, the waveform is smoothed by the resistors 18 and 19 and the capacitor 20, and then applied as a reference voltage of the operational amplifier 4 for integrating the noise signal.

That is, each time the sampler 3 performs N samplings (1 bit A / D conversion), the latch circuit 12 stores the count value of the data value “1” by the counter circuit 11. If the time required for N times of sampling is T, it is possible to compensate for the difference between the zero point of a component higher than 1 / T among the frequency components of the noise signal and the variation due to sampling.

Here, assuming that there is no variation due to sampling in the sampler 3, the values "1" and "0" appear evenly in the output data of the sampler 3.

However, if the number of times of sampling changes to a value of “1”, the analog signal becomes negative by the sampling operation.
The same effect as having an offset appears. In this case, a negative voltage may be applied to compensate for the DC offset equivalent to the comparison point of the sampler 3 only during the next sampling period.

On the other hand, if the value fluctuates in the direction of increasing the number of times of sampling to “1”, a plus voltage may be applied.

At this time, the output of the latch circuit 12 is set to 0.5
The relationship between the excess and deficiency from N (the state where “1” and “0” are equalized) and the compensation voltage value are proportional in the range of 0.4 to 0.6 N, and are not more than 0.4 or FIG. 3 shows an example in which a constant value is set at 0.6 or more. In FIG. 3, the reason why the constant value is set when the value is 0.4 or less or 0.6 or more is to prevent the system from becoming unstable if it is always in a proportional relationship.

In FIG. 3, data "00" to "F" are applied to input addresses (count values) 0 to N.
F ”(hexadecimal number) is read from the PROM 13, and the data“ 00 ”causes -3 [V] and data“ 8
0 [V] by "0", +3 by data "FF"
In addition to the voltage of [V], a voltage as shown is generated.

This voltage is, during the time T, the solid line 1 in FIG.
As shown at 50, the value is constant. However, it is necessary to further suppress statistical variations within an integration time for obtaining a correlation value described later, and for this purpose, a time constant circuit (the resistor 19 and the capacitor 20 in FIG. 2) is provided. Therefore, a compensation voltage having a smoothed waveform as shown by a broken line portion 400 in FIG. 4 is supplied to the integration circuit 40 in FIG. 2 as a reference voltage.

In FIGS. 1 and 2, the integrating circuit 4 is used.
The output of 0 is phase-adjusted by the phase shift circuit.
If necessary, it may be added on the input signal line side, that is, between the capacitor 2 and the sampler 3, or may be provided only on the input signal line side without being provided on the output of the integration circuit. Thus, the phase of the integrated output waveform and the phase of the low-frequency component of the signal input to the sampler 3 can be brought closer to a better state. The amount of phase shift may be adjusted by a variable capacitor.

Next, FIG. 5 shows the configuration of an embodiment of the correlation operation device according to the present invention. This apparatus uses a plurality of sample circuits shown in FIG. 1 or 2 to sample a noise signal from a celestial body or the like, and synthesizes the image of the celestial body by image processing.

For this reason, the correlation calculating apparatus according to the present embodiment is provided with antennas 21 to 2N for receiving a noise signal from an observation target, and is provided for each of these antennas. Frequency conversion into a band, and quadrature detection to obtain an I video signal and Q
Receivers 31 to 3N for transmitting as video signals, sample circuits 41 to 4N for sampling I video signals from the respective receivers, sample circuits 51 to 5N for similarly sampling Q video signals, and sample circuits And a correlator 60 for calculating a correlation value based on the data from.

In such a configuration, signals received by the antennas 21 to 2N from the observation target are transmitted to the receivers 31 to 3N.
The frequency is converted to a video band by N, and I and Q video signals are formed by quadrature detection to be sampled circuits 41 to 4.
N, 51 to 5N. The output data in which the variation caused by sampling the noise signal by each sampling circuit is input to the correlator 60, and the correlation values (amplitude and phase) between all the received signals are calculated.

The correlation value is a complex time function having intensity (amplitude) and phase information. Therefore, when the receiver converts the frequency to the video band, the I component and Q
The components are extracted separately, and a correlation operation is performed on these components.

The correlation operation is basically performed by A /
If the result is integrated after D-conversion, information comparable to that of the correlation calculation using an analog amount can be obtained at the time of imaging, but here, 1-bit A / D conversion is performed, and the obtained 1-bit correlation is obtained. A method of converting the value to a multi-bit correlation value later is adopted.

Here, the two antennas are E1 and E2, and the 1-bit A / D converted values of the I video signal component and the Q video signal component are respectively (E1 COS, E1 SIN),
(E2 COS, E2 SIN), the correlation value is Σ [E1
E2] (E2 is a conjugate complex number). Then, in this case, the real part is Σ (E1 COS · E2 COS)
+ Σ (E1 SIN · E2 SIN), and the imaginary part is Σ (E1 SIN
IN · E2 COS) −Σ (E1 COS · E2 SIN). These operations are performed by the Exclu
It is replaced by a live OR circuit and integrated for a predetermined time.

[0038]

As described above, according to the present invention, low-frequency components and high-frequency components of the components included in the noise signal to be sampled are separately compensated for the position of the zero point, thereby accommodating the sampling. Data variation can be suppressed, the correlation value between signals can be determined with high accuracy, and there is an effect that an image is not deteriorated.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a first embodiment of a sample circuit according to the present invention.

FIG. 2 is a block diagram showing a configuration of a second embodiment of the sample circuit according to the present invention.

FIG. 3 is a diagram showing an example of data written in advance to a PROM in FIG. 2;

FIG. 4 is a waveform diagram of a compensation voltage input to the integration circuit in FIG. 2;

FIG. 5 is a block diagram showing a configuration of a correlation operation device using the sample circuit of FIG. 1 or 2;

[Explanation of symbols]

 Reference Signs List 3 Sampler 11 Counter circuit 12 Latch circuit 13 PROM 14 D / A converter 21-2N Antenna 40 Integrator circuit 60 Correlator 70 Phase shift circuit

Claims (4)

(57) [Claims] An integration circuit for integrating a noise input signal;
The input signal is sampled using the integrated output as a reference value.
Sampler and the sampler output of the sampler
A compensation circuit for compensating a reference voltage of the integration circuit.
A sample circuit that specializes in this. 2. An integrating circuit for integrating a noisy input signal,
A phase shift circuit allowed to match the phase of said input signal from the integrator output, a sampler for sampling the input signal as a reference value an integration output after this phase, the sampler
Means for counting the sample output of
Compensating the reference voltage of the integration circuit according to the count value
And a compensating circuit. 3. A correlation value of a plurality of noisy input signals is obtained.
A correlation operation device, corresponding to each of the input devices
An integrating circuit provided for integrating a corresponding input signal;
Match the phase of each integrated output of the circuit with the phase of the input signal
And the integrated output after this transfer to the reference value
A sampler that samples each of the input signals as
An arithmetic circuit for determining a correlation value of the sample results;
Each reference of the integrating circuit according to the sample output of the sampler
A compensating circuit for compensating voltage.
Related arithmetic unit. 4. A correlation operation device for obtaining a correlation value of a plurality of noise-like input signals, comprising: an integration circuit provided corresponding to each of the input devices for integrating a corresponding input signal; A phase shift circuit for matching the phase of the integrated output with the phase of the input signal, a sampler for sampling the input signal with each integrated output after the transfer as a reference value, and an operation for obtaining a correlation value of the sampled result Circuit and the said
A means for counting the sampler sample output at a predetermined cycle
And each reference voltage of the integrating circuit according to the count value.
And a compensating circuit for compensating the noise.
Arithmetic device.