JPS6145909B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a digital signal receiver that detects a digital signal of a predetermined frequency by discrete Fourier transform.

N sample number sequence f(nt), 0nN-1
The discrete Fourier transform F(kΩ) of is defined as k=0, 1, 2..., N-1, where T is the sampling interval in the time domain and w(nt) is the window function Ω=2/NT (1) against Defined by Since Fw (kΩ) is given by the convolution integral of the Fourier transform W (ω) of the window function W (nT) and the Fourier transform F (ω) of the input signal f (nT), the Fourier transform W (ω) of the window function If the minimum positive frequency that provides a zero point is ω ₀ , then two frequencies separated by ω ₀ can be detected independently. Therefore, in a signal receiver whose frequency interval to be detected, that is, the frequency resolution is Δω, it is sufficient to perform discrete Fourier transform using a window function with an integration time NT such that ω ₀ =Δω.

It took this integration time NT to perform one discrete Fourier transform. For this reason, it takes a long time to determine the presence or absence of a signal by looking at the results of several discrete Fourier transforms. Furthermore, since the integration time is long, there are cases in which it is not possible to perform discrete Fourier transform a sufficient number of times to determine the presence or absence of a signal while the signal continues.

An object of the present invention is to provide a digital signal receiver that can accurately detect signals in a short period of time by shortening the time required for one dispersive Fourier transform.

Since the window function is small near the beginning and end of the integration time NT of the window function, it is thought that omitting this part will not have much effect. A discrete Fourier transform is performed on the product of the window function and the input digital signal. However, the integration time NT of the window function is determined by the frequency resolution of this digital signal receiver, that is, the interval between the two frequencies to be separated.
That is, in the present invention, the discrete Fourier transform calculation circuit is controlled so as to omit calculations in at least one portion near the beginning and end of the integration time NT.

Next, an embodiment of the digital receiver according to the present invention will be described with reference to the drawings. Input terminal 11
The digital signal f(n _T ) is then supplied to the arithmetic circuits 1 ₁ to 1 _P that perform discrete Fourier transform, and in these arithmetic circuits 1 ₁ to 1 _P a window function w is applied, respectively.
(n _T ) is multiplied to calculate a discrete Fourier transform. The minimum frequency ω that gives the zero point of the Fourier transform W(ω) of the time window function W(n _T ), where the frequency interval to be detected in the arithmetic circuits 1 ₁ to 1 _P is Δω.
The integration time NT is chosen so that ₀ is equal to Δω. The arithmetic circuits 1 ₁ to 1 _P are selected, for example, to detect signals of each frequency component in a multi-frequency dial signal. The calculation times of these calculation circuits 1 ₁ to 1 _P are controlled by the timing of the control circuit 12 .

The outputs of these arithmetic circuits 1 ₁ to 1 _P are binary values, for example, "1" or "0", for each arithmetic period, and these outputs are supplied to the output logic circuit 13. The output logic circuit 13 uses, for example, majority logic or logic to determine how many times the same thing is detected, and outputs it to a predetermined output terminal corresponding to the state of its input.

FIG. 2 shows a specific example of discrete Fourier transform characteristics when calculation is performed for the entire integration time NT determined by the frequency interval to be separated. The window function w(n _T ) is a Hamming window function w(n _T ) = 0.54−0.46 cos (2n/N) (3) The sampling interval used is T = 125 μs, the number of samples N = 80, and the angular frequency is 2f _C = kΩ. Discrete Fourier transform F′(kΩ) for This is the frequency characteristic when

In this invention, as described above, calculations near at least one of the beginning and end of the integration time NT are omitted. For example, in the specific example shown in Figure 2, the discrete Fourier transform F'' _W (kΩ) is obtained by omitting 10 calculations at the beginning and end of the integration time, that is, shortening the integration time by three-quarters. The characteristics are shown in Figure 3.

Generally, the window function w(n _T ) is even symmetric around n=N/2 or n=(N-1)/2, and n=
The value of the window function w(n T ) near n=0 and n=N is significantly smaller than the value w(N/ _2T ) of the window function near N/2. Therefore, even if the terms near 0 and near N of n are omitted in the integration of equation (4), the calculation result will not change significantly.

This is clear even when compared with FIGS. 2 and 3. In other words, the effect of omitting the above operation is F′ _W (k
Ω) appears where the value is small, and the side rope increases due to omission, but the outline of the characteristics does not change and the center frequency f _C also matches.

Therefore, it is understood that the integration time may be shortened. As mentioned above, since the level of the side rope increases, the dynamic range of the input signal that can be detected becomes smaller compared to the case where the integration time is not short. However, if you use the so-called variable threshold method, which changes the threshold value for determining whether the calculation output is "0" or "1" depending on the magnitude of the input signal, it is possible to handle input signals with a wide dynamic range. It can also be detected. Although it depends on the detection conditions, it is also possible to omit the first 1/4 and the last 1/4 of the integration time and shorten the overall integration time to about 1/2 of the time required for frequency resolution. . Therefore, according to the present invention, a signal can be detected in a shorter time, and detection can be performed more times than before in a limited period of time, making detection more accurate. In the above, the Hamming window function was used as the window function, but other window functions such as the Hanning window function, glass window function, and cos window function were similarly calculated by omitting the integral time near the beginning and end. A relationship similar to that shown in FIGS. 2 and 3 was obtained.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of a digital signal receiver according to the present invention, and FIG. 2 is a curve diagram showing characteristics of discrete Fourier transform by a conventional receiver.
FIG. 3 is a curve diagram showing an example of characteristics of discrete Fourier transform by the receiver of the present invention. 1 ₁ ~ 1 _P : Discrete Fourier transform circuit, 11:
Input terminal, 12: control circuit, 13: output logic circuit.

Claims (1)

[Claims] 1 The input digital signal is multiplied by an indifferent function and subjected to discrete Fourier transform in an arithmetic circuit, and depending on whether the magnitude of the transform output is greater than or equal to a predetermined value, the input digital signal is determined to be a signal with a predetermined frequency. In a digital signal receiver that detects the presence or absence of
As the above-mentioned window function, a window function is used in which the minimum positive frequency that gives the zero point of the Fourier transform matches the target frequency resolution, and at least one of the areas near the beginning and end of the integration time of the window function is excluded. A digital signal receiver comprising a control circuit configured to perform the above-mentioned discrete Fourier transform on a portion.