JPH0340874B2 - - Google Patents

Description

[Detailed Description of the Invention] [Technical Field of the Invention] The present invention is a correlation method for detecting a signal having a spectrum equivalent to a reference power spectrum from an input signal by calculating a cross-correlation coefficient of a power spectrum. It relates to a processing device.

[Technical background of the invention and its problems]

FIG. 1 shows a conventional correlation processing device. Input signal _i is analog/digital (A/
D) is supplied to the conversion circuit 11, and this conversion circuit 11
At the sampling pulse signal _s of a predetermined period, the signal is converted into a digital time series signal. This sample data is supplied to the first storage circuit 12 and stored therein. The number of samples stored is 2N. The sample data read out from the storage circuit 12 is sequentially supplied to the power spectrum calculation circuit 13, and is processed using, for example, FFT (Fast Fourier Transform).
N power spectrum data x _i (i
=0, 1,...,N-1) are obtained. This power spectrum data x _i is supplied to the second storage circuit 14 and stored therein.

On the other hand, the third storage circuit 15 stores reference power spectrum data _{y i} corresponding to the signal to be detected in advance.
(i=0, 1, . . . , M-1; M<N) are stored. The data stored in storage circuits 15 and 14 are read out in sequence and supplied to cross-correlation calculation circuit 16. This calculation circuit 16 performs the following calculation.

However, k=0, 1,...N-1 = 1/M _k+M-1 〓 ^i=K x _i = 1/M _M-1 〓 ⁱ⁼ⁿ _y iThe correlation obtained by this calculation The number r(k) is supplied to a determination circuit 17, which determines the maximum value r(k) _nax of the correlation coefficient r(k). Then, from this value of k, the center frequency of the signal whose power spectrum pattern matches the reference spectrum pattern is determined. Figures 2a, b, and c show examples of x _i , y _i , and r(k), respectively.

By the way, since the power spectrum x _i of the input signal is discrete data for each frequency resolution Δ (Δ = _s / 2N), it is determined by the value of the center frequency ₀ of the signal as shown in Figure 3 a and b. Therefore, what should originally be the same spectrum has the possibility of becoming quite different spectrum data. The dotted line _between a and b in Figure 3 shows the _continuous spectral envelope. (a positive integer), and a value of ₀ makes these two extremes and a spectral pattern in between.

This means that if the reference spectral pattern is stored, for example, in the form shown in Figure 3a,
In the case of a pattern between a and b shown in FIG. 3, the maximum value r(k) _nax of the correlation coefficient varies accordingly. As described above, the conventional apparatus has the drawback that the correlation coefficient varies greatly depending on the relative relationship between the center frequency of the signal and the frequency resolution.

[Purpose of the invention]

The present invention has been made based on the above circumstances, and its purpose is to provide a correlation processing device that can obtain a predetermined correlation coefficient no matter what the center frequency of the input signal. It is something.

[Summary of the invention]

This invention changes the input signal by a frequency Δ/n smaller than the frequency resolution Δ determined by the power spectrum calculation circuit, calculates the correlation coefficient with the reference spectrum for each of the changed signals, and calculates the maximum value from among these. By making this determination, the correlation coefficient that most closely approximates the reference spectrum pattern is determined.

[Embodiments of the invention]

An embodiment of the present invention will be described below with reference to the drawings. Note that the same parts in FIG. 4 as in FIG. 1 are given the same reference numerals.

In FIG. 4, input signal _i is supplied to frequency conversion circuit 41. In FIG. An output signal from a local oscillation circuit 42 is supplied to this conversion circuit 41, and the frequency of the input signal _i is changed by this signal. That is,
The output frequency of the local oscillation circuit 42 can be varied within a predetermined range, and this signal causes the frequency of the input signal _i to be shifted within the range of (N-M+1)Δ. This frequency-converted signal is sampled in the A/D conversion circuit 11 in exactly the same manner as described above, and this sample data is supplied to the first storage circuit 12 and stored therein. The sample data read from the storage circuit 12 is sequentially supplied to the power spectrum calculation circuit 13, and N pieces of power spectrum data x _i are determined. This power spectrum data x _i is supplied to the second storage circuit 14 and stored therein. Then, from this memory circuit 14, a fifth
As shown in the figure, only M pieces of data are read out, and this data is the reference power spectrum data y _i (i=0, 1, . . . , M−) read out from the third storage circuit 15.
1; M<N) and is supplied to the cross-correlation calculation circuit 16. This calculation circuit 16 performs the calculation shown in equation (2) to determine the correlation coefficient r.

However, = 1/M _M-1 〓 ⁱ⁼⁰ x _i = 1/M _M-1 〓 ⁱ⁼⁰ y _i Now, in this device, first, the output frequency of the frequency conversion circuit 41 is equal to that of the input signal _i. The frequency of the local oscillation circuit 42 is selected to be equal to the frequency, and the maximum value r ₁ of the correlation coefficient at this time is determined. Next, for the second time, the frequency of the local oscillation circuit 42 is changed by Δ/n, and as shown in _{FIG .} The correlation coefficient r ₂ is determined so that it is output as a _-1 spectrum. Thereafter, the frequency of the local oscillation circuit 42 is similarly changed by Δ/n and calculations are performed n(N-M+1) times, and the correlation coefficients r ₁ , r ₂ , ..., r _{o(N-M+1 )} is required. As a result,
Since the spectrum of the desired signal passes through the shaded area shown in FIG. 5 in steps of Δ/n, an intermediate pattern between FIG. 3a and FIG. 3b will necessarily appear. Therefore, the determination circuit 17 selects the largest correlation coefficient from r ₁ to _{r o (N-M+1)} to obtain the correlation coefficient closest to the reference frequency spectrum pattern. I can do it. That is, a correlation coefficient r _nax with little variation can always be obtained no matter where the frequency ₀ of the signal is located.

〔Effect of the invention〕

As described in detail above, according to the present invention, it is possible to provide a correlation processing device that can obtain a predetermined correlation coefficient regardless of the center frequency of the input signal.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an example of a conventional correlation processing device, FIG. 5 is a block diagram showing one embodiment of the correlation processing device according to the present invention, and FIG. 5 is a diagram shown for explaining the operation of the present invention. 11...A/D conversion circuit, 13...Power spectrum calculation circuit, 16...Cross correlation calculation circuit, 17
... Judgment circuit, 41 ... Frequency conversion circuit, 42 ...
...Local oscillation circuit.

Claims (1)

[Claims] 1. An oscillation circuit that can change the oscillation frequency, a frequency conversion circuit that changes the frequency of an input signal using this oscillation signal, a circuit that A/D converts this conversion output signal with a predetermined sampling pulse signal, and A circuit that obtains a power spectrum from a sampled signal is supplied with a power spectrum and a reference power spectrum that are obtained by changing the oscillation frequency by a predetermined frequency that is smaller than the frequency resolution of the circuit that obtains the power spectrum, and a cross-correlation coefficient is obtained. What is claimed is: 1. A correlation processing device comprising: a circuit for determining the correlation coefficient; and a determination circuit for determining the maximum value from the determined correlation coefficient.