JPH0462618B2 - - Google Patents

Description

[Detailed Description of the Invention] [Summary] In order to detect whether a signal input to a receiving device is a training signal, the frequency information of the input signal is converted into a DC vector component, and then the DC component is converted into a DC vector component. Extraction processing was performed to detect whether the signal was a training signal or not.

[Industrial application field]

The present invention relates to a training signal detection device for a data transmission system.

[Conventional technology]

When transmitting and receiving data between a terminal device and a host computer, or between a terminal device, the data transmitting modem transmits a training signal before transmitting the actual data, and the receiving modem transmits a training signal. The training process is performed based on
Determination of initial values of correction coefficients for line equalization, input of data timing, etc. are performed. Examples of such training signals include, for example,
The one shown in No. 278219 is used. In this case, conventional methods for detecting the training signal include the methods shown in FIGS. 5 and 6, for example.

This is a method of starting training by detecting energy, and when the modem of the receiving device receives the input of the data signal _D1 as shown in a in Figure 5, it detects the energy and as shown in c in the figure. CDi
A signal is generated, and the operation of the receiving section is started (ie, set up) at the rising edge of this CDi signal. As shown in b in FIG. 5, the data signal D ₁ includes a training signal D ₂ at the beginning and a data signal for arithmetic processing (hereinafter referred to as processed data signal) at the rear.
_D3 . Then, when the receiving section starts operating, the training signal _D2 is taken into the receiving modem and a training operation is performed.

[Problem that the invention seeks to solve]

However, in such a conventional training signal detection method, if any increase in energy occurs at the input terminal, the modem activates and starts setting up the receiving section. This may cause some inconvenience. This is because although the transmitting device initially sent out a data signal _D1 similar to that shown in Figure 5, noise N was added to the line in the middle, and a signal as shown in a in Figure 6 was received. Indicates the case where the data is input to the side device. If such noise N exceeds a certain level, the modem of the receiving device will normally
As shown in b and c in the figure, at time t ₁ to receive the training signal D ₂ and the processed data signal D ₃ .
Although the CDi signal should be generated, the CDi signal is generated at time _t0 , which is the rise of the noise N. For this reason, in the receiving side device, setup of the receiving section is performed at time _t0 , causing a problem that training operations cannot be performed.

Additionally, if a momentary interruption occurs during data transmission, the modem may mistakenly recognize the input of the second half of the data signal D ₁ (or processed data signal D ₃ ) at the point where the momentary interruption occurs as a training signal input. This may result in an error signal indicating that training cannot be performed.

The present invention has been made by focusing on such conventional problems, and its purpose is to determine whether a signal input to a receiving section is a training signal, and to determine whether a signal input to a receiving section is a training signal or not. It is an object of the present invention to provide a training signal detection device that causes a receiving section to start operating.

[Means for solving problems]

FIG. 1 is a diagram showing the principle of the present invention. In the present invention, a training signal detection section 8 is provided within the modem, which is branched from the data transmission path 7 that generally performs demodulation processing of data signals. The training signal detection unit 8 incorporates means for converting the carrier frequency component of the input signal into a DC vector component, and a low-pass filter for extracting the DC component from this vector component. Then, it is determined whether the input signal is a training signal or not.

[Function] Analog signals input to the receiving device (regardless of whether they are training signals, processed data signals, or even noise signals) are sent to the A/D installed in the modem.
It is converted into a digital signal by a converter. The digitally converted signal is sent to a training signal detector 8 which is branched from the data transmission line in the modem.
The training signal is demodulated in the training signal detection section 8, and the frequency band is extracted.AGC control and energy control are then performed, and the phase difference component is extracted, as well as the DC component is extracted by a low-pass filter. is performed, and it is determined whether the signal is a training signal or not based on DC correlation.

〔Example〕

FIGS. 2 to 6 are diagrams showing one embodiment of the present invention. Of these, FIG. 2 is a block diagram showing the configuration of a modem incorporating the training signal detection device of the present invention. This modem includes an A/D converter 1 that converts a data signal input from a line from analog to digital, a data transmission line 7 that performs demodulation and equalization processing of the data signal, and a data transmission line 7 that performs demodulation and equalization processing of the data signal.
and a training signal detection section 8 which is provided in a branched manner.

The data transmission path 7 includes a carrier signal and a carrier signal for the digital signal from the A/D converter 1.
A demodulator 2 which multiplies the values of signals having a phase difference of 90° to obtain x-axis and y-axis components on the complex plane (that is, obtains a baseband signal), and a demodulator 2 which obtains the Nyquist frequency of each demodulated data signal. A roll-off filter 3 that cuts and extracts only a stable band in a predetermined band (in this case, the central portion), an AGC circuit 4 that performs automatic gain control on the output signal from the roll-off filter 3, and an AGC circuit that performs automatic gain control on the output signal from the roll-off filter 3. It consists of an equalizer 5 that performs equalization processing on four outputs, and a code conversion section 6 that performs processing such as code conversion on the demodulated and equalized signal.

The training signal detector also detects the digital signal (whether it is a training signal or a processed data signal) converted and output by the A/D converter 1.
A second demodulator 9 is used to demodulate the signal (or a noise signal, it is unknown at this point) with a carrier of a certain predetermined frequency. A second roll-off filter, which is a filter means for extracting the
AGC is a means for regulating the level of the output signal from the filter 10 and the roll-off filter 10 to a size of radius 1 on a complex plane by gain control.
a circuit 11, and a limiter 1 which is a means for limiting the energy of a signal having a radius greater than 1 to within a range of 1.
2, a differentiator 13 which is a means for extracting a phase difference component by the complex conjugate of a vector, a low-pass filter 14 which is a means for extracting a DC component from a vector component and creating a DC correlation, and a low-pass filter 1.
4, and a determination section 15 that checks whether the signal inputted to the training signal detection section 8 is a training signal based on the DC correlation created by step 4.

Figure 3 shows the AGC inside the training detection unit 8.
A specific example of the configuration of the circuit 11, limiter 12, difference device 13, low-pass filter 14, and determination section 15 is shown. The "" mark in the figure is a multiplier, "T"
is a tap register, “+” is an adder, and α ₁ ~
α ₁₆ represents a constant coefficient determined as appropriate. The AGC circuit 11 includes a plurality of multipliers 16, and squares the power of the vector signal (indicated by DRR and DRI) output from the roll-off filter 10, and subtracts it from a certain reference value to calculate the error. and accumulate them. A tap 17 is incorporated for automatic adjustment of the reference value. Limiter 12 incorporates multipliers 18 and 19. Multiplier 18 doubles the power of the input signal and multiplier 19
has the function of multiplying the signal power by 1/2, thereby forming a simple limiter that causes overflow at the limit of calculation.

Further, the differentiator 13 includes the vector signals DRR,
Taps 20 and 21 corresponding to the DRI, multipliers 22, 23, 24, and 25 that take vector differences between these tap outputs and limiter outputs, and an adder 2.
6,27. The low-pass filter 14 comprises multipliers 28, 29, taps 30, 31, and adders 32, 33 for adding the outputs of the multipliers 28, 29 and the outputs of taps 30, 31. The determination unit 15 includes a multiplier and an adder that multiply or add a predetermined coefficient value to the output from the low-pass filter 14, and the multiplied and added DRR component and DRI.
It has a determiner 34 that determines the polarity between the components.

The operation of the training signal detection section 8 when a data signal is input to the modem having such a configuration will be explained.

The training signal shown in Japanese Patent Application Laid-Open No. 61-278219 or the training signal of CCITT Recommendations V.29, V.33, and V.32 is one in which two signal points on a complex plane are repeatedly sent alternately. Therefore, the training signal consists of a single signal of a certain frequency, which is the alternating repetition frequency of the signal points, as shown in FIG. 10 cuts the Nyquist frequency, resulting in a waveform as shown in b in FIG. 4, which is input to the second AGC circuit as DRR and DRI.

On the other hand, the processed data signal or noise signal has a waveform over a certain frequency range as shown in g in FIG. This results in an aggregation pattern of dispersed points as shown in h in Figure 4.

In the second AGC circuit 11, gain control is performed on the input signal, and the training signal is
As shown in FIG. 4, the processed data signal is
As shown in , the level of the output signal is defined on a plurality of planes at a radius of 1. Then the second AGC
The output from the circuit 11 is input to the limiter 12,
Here, the energy of the signal whose signal level is larger than radius 1 is limited to within the range of 1, and the training signal and processed data signal are respectively d and j in Fig. 4.
It is output as the pattern data shown in . Furthermore, the output signal from this limiter 12 is sent to a differentiator 13.
The phase difference component is extracted here by the complex conjugate of the vector. By extracting this phase difference component, the training signal is transformed as shown by e in FIG. 4, while the processed data signal is transformed as shown in FIG.
The pattern signal is as shown in . The signal output from the subtractor 13 is then passed through the low-pass filter 1.
4, where the DC component is extracted from the vector component to create a DC correlation. For the training signal, a pattern indicated by f in Fig. 4 appears, indicating that there is a DC correlation, while for the processed data signal, a pattern indicated by l in Fig. 4 indicates that there is no DC correlation.
The pattern shown appears. As a result, the determination unit that receives the output signal from the low-pass filter 14 displays whether there is a DC correlation in the input signal (in the case of a training signal) or not (in the case of a processed data signal), and based on this display result. It is determined whether the receiving section starts operating.

〔Effect of the invention〕

As explained above, according to the present invention, the frequency information of the received signal is converted into a DC vector component in the modem, the DC component is extracted by filter processing, and the training signal is detected. After detecting whether the received signal is a training signal or not, the receiving section can be set up, resulting in the effect that malfunctions due to noise signals or malfunctions due to instantaneous interruptions can be prevented.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the principle of the present invention, FIG. 2 is a block diagram showing the configuration of a modem incorporating the training signal detection device of the present invention, and FIG. 3 is a diagram showing the main parts of the training detection section in the modem. FIG. 4 is a diagram showing a signal processing state in the training detection section in the modem, FIG. 5 is a diagram showing a training detection method using energy detection in a conventional modem, and FIG. 6 is a diagram showing a detailed example. 5 is a diagram showing the operation when a noise signal is input in the same training detection as in FIG. 5. FIG. 1... A/D converter, 2, 9... Demodulator, 3, 10
...roll-off filter, 4,11...AGC circuit,
8... Training signal detection section, 12... Limiter,
13...Differentiator, 14...Low pass filter, 15
...Judgment Department.

Claims (1)

[Claims] 1. Means 1 for converting a received analog signal into a digital signal by being connected to an input terminal connected to a receiving line, and demodulating the A/D converted signal with a carrier of a predetermined frequency to generate a baseband signal. means 9 for obtaining a signal; means 10 for cutting the Nyquist frequency of the demodulated baseband signal and extracting the baseband signal in the frequency band of the carrier signal; means 11 for regulating the level; means 13 for extracting a phase difference component by performing a complex conjugate operation of a vector indicated by the baseband signal on the level-regulated baseband signal; and a vector after the complex conjugation. A training signal detection device in a data transmission system, comprising: a filter means 14 for extracting a DC component from the components; and a means 15 for determining whether an input signal is a training signal based on the output from the filter means 14.