JPH0634484B2 - Tone detection method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of use] The present invention outputs to a calling side when a called side when a called side is automatically called when performing communication using a switched line. The present invention relates to an answerback tone detection method, and more particularly, to a tone detection method improved so that erroneous detection does not occur due to line noise.

Generally, in order to perform data transmission using a switched line, a modem for modulating / demodulating data into a voice band signal is used as shown in FIG. The connection procedure is the same as that of the telephone, and a dial signal is given from the calling modem A to call the called modem B. The called modem B has an automatic incoming call function, and when an automatic incoming call is made, an answer back tone (ABT) is returned to the calling side.

The modem A on the calling side detects this answer back tone, recognizes the execution of the automatic incoming call on the modem B on the called side, and prepares for data transmission.

This answerback tone is given a specific frequency, 2025Hz in the US and 2100Hz in Europe.
Is.

[Conventional technology]

For this reason, each modem is provided with an answer back tone detector, and the method shown in FIG. 6 has been conventionally used. That is, the signal from the line is demodulated by the carrier wave, the waveform is shaped by the roll-off filter, and then the output AGCout whose gain is controlled by the AGC (automatic gain control) circuit is 20 times.
Filter 1 having pass band characteristics of 25 Hz and 2100 Hz
The signal of 2025Hz or 2100Hz component was extracted through 0 and 11 and input to the slicer / integrator circuit 13 via the adder circuit 12, and after the integration, the answerback tone was detected by slicing at a predetermined slice level. . Therefore, the filters 10 and 11 convert the frequency component into a DC component, determine the level of the DC component, and detect the answerback tone.

[Problems to be solved by the invention]

However, the conventional method has a problem that a detection error occurs when noise is placed on the line. That is, in the conventional method, if there is noise passing through the filters 10 and 11, this will cause erroneous detection. Especially in a switched line, bell noise, busy tone, voice, etc. are mixed with normal noise due to crosstalk, and noise that easily passes through the filters 10 and 11 is likely to occur. It was

[Means for solving problems]

An object of the present invention is to provide a tone detection method capable of stable tone detection with little false detection even in a noisy line.

Therefore, the present invention is a tone detection method for detecting a tone signal of a predetermined frequency from a received signal received, in which the in-phase component X and the quadrature component Y of the received signal demodulated by the carrier wave are
With respect to the complex number X + jY, the difference (X + j) by the complex conjugate X′-jY ′ of the in-phase component X ′ and the quadrature component Y ′ before t time.
Y). (X'-jY ') to obtain the in-phase component X and the quadrature component Y of the phase difference signal, and the in-phase component X of the DC component of each of the in-phase component X and the quadrature component Y of the phase difference signal. And a quadrature component Y are extracted, and a vector indicated by the in-phase component X and the quadrature component Y of the DC component is predetermined on the frequency and amplitude determination plane which is the vector plane of the in-phase component X and the quadrature component Y. The method is characterized by including a step of determining whether the tone signal is within the frequency range, and detecting the tone signal by the determination.

[Action]

In the present invention, since a tone signal such as a regular answer-back tone is a signal having a single frequency, it becomes constant when the phase difference is taken on the time axis. Therefore, by obtaining the phase difference on the vector plane by the difference due to the complex conjugate, setting the frequency correlation to the DC correlation, and further extracting the DC component, only a single frequency component having the phase difference correlation is extracted. Tones can be detected by determining on the frequency plane whether or not the frequency of the tone for this purpose. That is, the frequency correlation is set to a DC correlation, a single frequency component having a correlation is extracted from the DC correlation, and a tone of a predetermined frequency is detected from this.

〔Example〕

 Hereinafter, the present invention will be described in detail with reference to an embodiment.

FIG. 1 is a block diagram of an embodiment of the present invention, in which 1 is an AGC.
An (automatic gain control) unit that demodulates a received signal from a line with a carrier wave and equalizes the level of a signal that has been waveform-shaped by a roll-off filter, 2 is a complex conjugate difference unit,
As will be described later, the complex conjugate difference of the signals before and after the sampling time of the input signal is obtained to obtain the phase difference component of these signals. Reference numeral 3 is a low-pass filter unit, which is a DC component of the output of the complex conjugate difference unit 2. What is extracted, 4 is a determination unit, which determines the output of the low-pass filter unit 3 in a predetermined frequency plane and detects an answerback tone. Note that each of these units can also be realized by the processing executed by the signal processor.

The operation of the embodiment shown in FIG. 1 will be described with reference to FIGS. 2 and 3. The output of the roll-off filter is the AGC unit 1.
The gain is controlled by, the level change due to the line fluctuation is compensated, and the result is input to the complex conjugate difference unit 2. The complex conjugate difference unit 2 obtains the complex conjugate difference of two input signals that are temporally preceding and following, and obtains the phase difference thereof. That is, the AGC output is the in-phase component x (Real) and the quadrature component y (Imaginary), and A before t time
GC output is in-phase component x '(Real), quadrature component y' (Imagina
ry), the complex conjugate difference output (X + jY) is expressed by the following equation.

X + jY = (x + jy). (X'-jy ') (1) = (xx' + yy ') + j (yx'-xy') (2) ∴X = xx '+ yy' (3) Y = yx'-xy ′ (4) If it is a single frequency component, the phase difference of the input signals at the time point t apart is constant, so the complex conjugate difference between them is constant, that is, the DC component. On the other hand, random frequency components that are not single frequency components do not become direct current because the phase difference is not constant. Therefore, the low-pass filter unit 3 extracts only the DC component. Therefore, the low-pass filter unit 3 is provided with the pass band characteristic as shown in FIG. 2B, and the direct-current component, which is a single frequency component having a correlation with the phase difference, is extracted. This removes most of the noise. still,
FIG. 2a shows the pass band characteristic of the roll-off filter. Further, in order to extract only a single frequency component of the answerback tone frequency, the determination unit 4 makes a determination on the determination surface of FIG. Here, the detection frequency range is 2025Hz
For 2025Hz ± 7.0% (1883 to 2167H)
z) For 2100 Hz, 2100 Hz ± 7.0% (19
53 to 2247 Hz) and its center frequency is 2062.
Set to 5 Hz. When the demodulated carrier wave is 1800 Hz, the output of the low-pass filter unit 3 described above is a vector centered on the origin O in the frequency plane of FIG. 3, and this is designated as a point P ₁ . On the other hand, the ABT (answer back tone) region in the frequency plane is 206 from the above-mentioned detected frequency range.
Centering on a point P ₀ (z ₀ + jy ₀ ) on 2.5 Hz,
The inner radius _{r 1} of the circle C1 intersecting the 00Hz + 7.0% and 2025 Hz-7.0% line. That is, if the point P ₁ that is the output of the low-pass filter 3 is inside the circle C1, it is an answerback tone ABT. On the other hand, the circle C2 has a radius r ₂
Centered on the point P ₀ at 2342.5Hz, 1782.5Hz
It intersects with the line of and has a relationship of circle C1 and circle center circle. This circle C
2 is set in order to provide a hysteresis characteristic to the detection characteristic. Initially, the circle C1 is set as a boundary, but once it is detected that it is within the circle C1, the boundary is circled C2.
By providing a hysteresis, the waveform of the answerback tone detection output against noise (chattering) is prevented.

Furthermore, to perform the determination with respect to noise, the radius _r 3, _r 2 a circle N1, N2 of ₄ centered at the origin O are provided. That is, since the noise tends to converge on the origin O, a circle N1 that is a noise region centering on the origin O and a circle N2 for giving hysteresis to this are provided.

Therefore, the ABT detection range in the initial state is the range excluding the shaded area A2 inside the circle C1, and when the input signal enters the noise circle N1, the ABT detection range is the shaded area A1 inside the circle C1. , A2 is excluded, and when the input signal enters the ABT detection range, the ABT detection range is a range obtained by subtracting the inside of the noise circle N2 from the inside of the circle C2. In this way, erroneous detection at a frequency close to noise is prevented.

The present invention will be described in more detail.

FIG. 4 is a detailed configuration diagram of a main part in the configuration of FIG. 1, which can also be realized by processing executed by the processor. In the figure,
The same components as those shown in FIG. 1 are represented by the same symbols, and 21 and 22 are taps (delay means), each of which is A
A component that delays the x and y components of the GC output by t time (for example, T / 3), and 23 and 24, which are multipliers, respectively.
The output x and the output x ′ of the tap 21, the AGC output y and the tap 22 and the output y ′ are multiplied, 26 and 27 are multipliers, and the AGC output y and the output x ′ and A of the tap 21, respectively.
Multiplying the GC output x and the output y'of the tap 22,
25 and 28 are adders, and the adder 25 is a multiplier 2
The outputs of 3 and 24 are added to execute the above-mentioned expression (3), and the adder 28 adds the outputs of the multipliers 26 and 27 to execute the above-mentioned expression (4). The complex conjugate difference unit 2 is configured by these, the expressions (3) and (4) are executed, and X and Y of the phase difference components are obtained. Reference numerals 30 and 31 denote multipliers, which respectively output the outputs X and Y of the complex conjugate difference unit 2 with a coefficient α.
R (resistance) of the active RC filter
, 32 and 35 are adders, 33 and 36 are taps, and 34 and 37 are multipliers, which perform integration operation.
It corresponds to C (capacitor) of the RC filter, and constitutes a digital filter by these, and sets the pass characteristic to a desired one by the value of α.
In the present invention, α is set so as to pass the DC component as shown in FIG.
Is set. Reference numeral 40 denotes an error unit, which is an adder 40a, 4
0b for subtracting the coordinates x ₀ and y ₀ of the center point P ₀ of the detection region C1 from the outputs X and Y of the low-pass filter unit 3 to obtain the coordinates around the center point P ₀ , 41 Is a power calculation unit, and the output (X−x ₀ ) of the error unit 40,
Multipliers 41a and 41b for squaring each of (Y−y ₀ ).
And an adder 41c that sums these outputs, and a multiplier 41 that squares each of the outputs X and Y of the low-pass filter unit 3.
d and 41e, and an adder 41f that takes the sum of these outputs, the vector length P ₀ P ₁ from the center point P ₀ of the point P ₁ from the adder 41c, and the point O from the adder 41f. To obtain the vector length OP ₁ . 42 is a determination unit, for the vector length _P 0 _{P 1} from the center point _{P 0} is to determine the radius _r 1, _{r 2} within the ABT determination area, an adder 42a for subtracting the radius _{r 1,} or _{r 2} And the adder 42
The polarity determination unit 42b that determines the polarity of the output of a and the origin O
From the vector length OP ₁ to the noise region radii r ₃ and r ₄
In order to determine whether it is within the range, an adder 42c that subtracts the radii r ₃ and r ₄ and a polarity determining unit 42d that determines the polarity of the output of the adder 42c are provided, and 43 is a negative logic OR unit. , The outputs of the polarity determination units 42b and 42d are
Only when it is "0", "0" output (ABT detection output) is issued.

Next, the operation of the configuration shown in FIG. 4 will be described.

The output AGCout from the AGC unit 1 in FIG. 1 is input to the complex conjugate difference unit 2, and each of the outputs x and y is tap 21,
22 delayed by delayed outputs x ′, y ′ and output x,
y is multiplied by the multipliers 23, 24, 26 and 27 and further added by the adders 25 and 28 to obtain the phase difference components X and Y of the expressions (3) and (4).

The phase difference components X and Y are input to the low-pass filter unit 3 and subjected to correlation integration to extract only the DC component. These X and Y represent a point P ₁ centered on the origin O of the vector plane (frequency plane) of FIG. Next, in order to obtain the distance (vector length) from the center point P ₀ of the detection areas C1 and C2, the error section 40 subtracts the coordinates (x ₀ , y ₀ ) of the center point P ₀ ,
The power calculation unit 41 obtains the vector length P ₀ P ₁ . At the same time, the power calculation unit 41 outputs the outputs X and Y of the low-pass filter unit 3.
Is used to calculate the vector length OP ₁ from the origin O. These vector lengths P ₀ P ₁ and OP ₁ are input to the determination unit 42. When the output of the adder 42a is negative, the polarity determination unit 42b outputs “1” as an outside of the ABT region and also outputs the radius r ₁ to the converter 42a, and AB when it is positive.
The "0" output is issued as in the T region and the adder 42
The radius r ₂ is output to a. That is, the negative inverted vector length P ₀ P ₁ arrives at the adder 42a, and the radius r
₁ , r ₂ are added together, and P ₀ P ₁ is output from the adder 42a.
> R ₁ (r ₂ ) is negative, P ₀ P ₁ ≦ r ₁ (r ₂ )
Then a positive output is emitted and r is initially set.
_The range set to ₁ is a positive output (P ₀ P ₁ ≦ r ₁ )
By this, r ₂ is changed, and hysteresis can be provided.

On the other hand, when the output of the adder 42c is negative (that is, OP ₁ <r ₃ , r ₄ ), the polarity determining unit 42d outputs “1” as the noise area and outputs the radius r to the adder 42c.
_{When 4} is positive (that is, OP ₁ ≧ r ₃ , r ₄ ), “0” is output outside the noise region and the radius r ₃ is output to the adder 42c. Similarly, the range set to r ₃ in the initial state is changed to r ₄ by the minus (OP ₁ <r ₃ ) output, so that hysteresis can be provided.

Therefore, from the negative logic OR unit 43, the output of the polarity determination unit 42b is “O” (P ₀ P ₁ ≦ r ₁ or r ₂ ) and the output of the polarity determination unit 42d is “0” (OP ₁ ≧ r ₃ or r ₄ )
Only in this case, the ABT detection signal of "0" which is within the ABT area is output.

In this way, the determination is performed on the frequency plane.

In the above description, the answerback tone frequency is 20.
The range is set as shown in FIG. 3 so that both 25 Hz and 2100 Hz can be detected, but in the case of either one, the range is set similarly. Also, the frequency of the answerback tone is not limited to this.

Although the present invention has been described with reference to the embodiments, the present invention can be variously modified according to the gist of the present invention, and these modifications are not excluded from the present invention.

〔The invention's effect〕

As described above, the present invention has the following effects.

Noise that does not have phase correlation is removed in the extraction step, and noise that has correlation is also removed by the determination on the frequency determination surface. Therefore, it is possible to accurately detect the tone signal even in a line with large noise. The tone signal can be detected even on a public line with a lot of noise.

Since the determination in the frequency plane is performed using vector information,
Since the amount of information is large, it can be detected accurately compared to the case where it is determined by the scalar information.

Since the final determination is performed using vector information, the complex complex conjugate difference needs only one step, and can be realized easily and inexpensively.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an embodiment of the present invention, FIG. 2 is a characteristic diagram of a low-pass filter unit having the configuration of FIG. 1, FIG. 3 is an operation explanatory diagram of a determination unit of the configuration of FIG. 1, and FIG. FIG. 1 is a detailed configuration diagram of a main part of the configuration, FIG. 5 is an explanatory diagram of an answerback tone, and FIG. 6 is an explanatory diagram of a conventional detection method. In the figure, 2 ... Complex conjugate difference part, 3 ... Low-pass filter part,
4 ... Judgment unit.

Claims (1)

[Claims] 1. A tone detection method for detecting a tone signal of a predetermined frequency from a received signal received, wherein an in-phase component X and a quadrature component Y of a received signal demodulated by a carrier wave are provided.
Of the in-phase component X ′ before time t with respect to the complex number X + jY,
Difference (X +) due to complex conjugate X′-jY ′ of orthogonal component Y ′
jY). (X'-jY ') to obtain the in-phase component X and the quadrature component Y of the phase difference signal, and the in-phase component X of the DC component of each of the in-phase component X and the quadrature component Y of the phase difference signal. And an extraction step of extracting a quadrature component Y, and a vector indicated by the in-phase component X and the quadrature component Y of the direct-current component is predetermined on the frequency and amplitude determination plane that is the vector plane of the in-phase component X and the quadrature component Y. A tone detecting method comprising a step of determining whether the tone signal is within a frequency range, and detecting the tone signal by the determination.