JPS5897928A - Training detection system - Google Patents

Abstract

PURPOSE:To improve the instantaneous breaking resistance of a transmission and reception circuit by providing a means of discriminating a training signal. CONSTITUTION:A training circuit 10 consists of sum circuits 11 and 12, absolute value circuits 13 and 14, low-pass filter (LPF)15, difference circuit 16, and code deciding circuit 18. Each sum circuit has a delay circuit T in its shunt circuit and sums the real part and imaginary part of the output of a roll-off filter (ROF)4 with a one-signal shift; the absolute value circuit 13 squares and sums up the outputs of the circuits 11 and 12, and the absolute value circuit 14 and LPF15 branch the real part and imaginary part of the output of the ROF4 and square and sum up those outputs. The difference circuit 16 calculates the difference between the output A of the circuit 13 and the output B of the LPF15; and the output (c) of the difference circuit 16 is sent to a tap delay line 17 and the output of delay timing of each signal is supplied to the code deciding circuit 18 to decide on training, tone, or data. Then, reception is led to by a signal discriminated as a training signal.

Description

DETAILED DESCRIPTION OF THE INVENTION (1) Technology O minute hand The present invention relates to a training detection method for identifying whether a received signal is a training signal or data in order to improve the resistance to momentary interruptions of a transmission/reception line.

ψ) Conventional technology Conventionally, in a data transmission system using phase modulation data,
A method is often used in which a training signal including a π continuous transmission pattern at the beginning is transmitted, and the receiving side performs p equalization using an automatic equalization amount to perform reception pull-in.

In this conventional data transmission system, when there is an instantaneous wR on the line, all the next received good signals are regarded as training signals. This rounding and reception (the automatic equalizer in Gill's modem goes into a divergent state, so I guess it takes a while to get into a state where it can send).

According to the recommendation of CCITTOV27 regarding the training signal, as shown in FIG. 1, the training signal consists of five segments (BNG)], 5IG1 is π continuous transmission, and KO as shown by ■ on the phase plane.

The stomach is sent out alternately (π continuous transmission) and used for carrier detection in the modem of the receiving section, pulling in the mu GC, pulling in the /ing PLL, and pulling in the carrier PC, 8EG2.
is 0. as shown by ■. It is sent out randomly and used for pulling in the automatic equalizer and the carrier PC. 8
EG5 sends out an 8-value statin pull as shown by ■,
Pull in the descrambler. It is indicated by 8ffiG4a(4) and serves as a data receipt.

Since the above-mentioned 0IIb operation is performed even when the line O@ is disconnected, if the next signal is data, the automatic equalization amount will be skewed.

(1) 尭＠O Purpose The present invention distinguishes whether a received signal is a training signal or data, and performs the optimum processing of the transmission/reception signal.
It is an object of the present invention to provide a nine-training detection method that improves resistance to instantaneous interruptions in XA rays.

(In the four-shot @O configuration, the train-ender detection method of the present invention connects the transmitting device and the receiving device to the rounding that connects the above-mentioned rounds.) Receives the train-ender signal, performs reception pull-in for the data signal, and transmits the data. The data transmission system is characterized in that a means for identifying a training signal is provided, and reception pull-in is performed using a signal identified as a training signal.

(@Embodiment of the Invention j1211! El/Ii This is a flowchart explaining the present invention in detail.

When a failure occurs in a data transmission system using a training signal as described above, there is an L requirement for identifying whether the next signal to come is a training signal or a data signal. Furthermore, in addition to this, the case of tone can also be considered. In the present invention, the signals that come next after the training signal (segment continuous transmission pattern) are separated into training, data, tone 0 order (Kg) in the order of 0 after the training signal, and if the training signal is distorted, the training sequence is used. 0 Execute the pull-in sequence of the receiving side modem, and if the data signal is disturbed, the self-excitation equalization amount O
In order to prevent the w4 difference from jeopardizing Osaka University, a bi-cutting sequence is executed in which a predetermined input phase, tag coefficient, etc. are given.

A similar tone sequence is executed when patterning the tone signal. ? -o The instantaneous interruption resistance can be increased.

Figure 5) and -) are 63
1. The same figure (-) is the overall configuration diagram of the receiving device, and the same figure (b) is l.
I! FIG. 3 is a detailed explanatory diagram of the training detection circuit of the section.

In the same figure (1), the received signal consisting of the Letore ensign signal and the data p is input to the automatic gain controller (AGC) 1 and the gain is 1llll! 11, converted to p digital signal by A/D converter 2, and converted to 11111 (DICM)! to demodulate. This demodulated output is input to a roll-off filter (RoF) 4 for **00 point detection.ROjF4 is a low-pass filter having a low cosine characteristic. This output is input to the self-excitation equalization (to) (mu IQ) 5 and the training detection circuit 10 which is the main part of the present invention. The training detection circuit has means for identifying the continuous transmission pattern, and identifies whether the next signal at the time of an instantaneous interruption is training, data, or tone, and applies the corresponding resistance.

FIG. 2B is a detailed Ia explanatory diagram of the training detection circuit 10. An absolute value circuit that squares each output and adds the real part and imaginary part of the ROj'44D output through a summation circuit 11 and 12 that has a delay circuit in each branch and shifts the number 1111. 15. On the other hand, an absolute value circuit 14 and a low-pass filter (LP
F) 15, the output (2) is input into the difference circuit 16, and the difference between the output (2) and the output (2) of the absolute value circuit 15 is determined. in this case,
The LPF 15 multiplies the continuous signal by a constant α1 and adds the resultant signal.The LPF 15 supplies a nine-integral set value to the difference path 14 based on the output of the absolute value circuit 14. Next, the differential circuit 160
Output OTP: Send to M line 17, delay tie (T) for each signal, output to sign judgment line 18.

ζζ determines whether it is training, tone, or data using the method shown in Figure 4 (a) and (9), and outputs it as an identification signal.

FIG. 4 shows the waveforms of the output (■) and 0 in the case of training, tone, and data O, respectively.

(-) in the same figure is 6 for training O], absolute value circuit 15
The output [present] is the sum of the absolute values of the summation circuits 11, 12 and 40, so it becomes 0.

Also, the output (2) is a constant positive value], and the output 0 is the difference between the output (2), the output (2), and O, so a constant O negative value e as shown in the figure is shown for each tie.

1iiQig (b) is a tone, and the p1 output (2) has a positive value in the summing circuit 11 and 0 in the summing circuit 12, so it shows a positive value as the absolute value O sum. The output ■ indicates a positive 1iTL that is the same as the output ■, and therefore the output 0 indicates a constant positive value e for each tie as shown in the figure.

1ji1 Figure (・) is for data (To), and the output ■ is shown as a waveform where each point has a positive value,
Depending on the relationship, the output may be large or small. Therefore, since the output 0 is the difference between the output and the output ■, the positive value e and the negative value e are sent to each timing.As described above, if the code arrangement of the signal at each timing is determined, training Able to identify tones and data.

No. 511A is a configuration i1 of another embodiment of the present invention! This is a clear diagram.

In the same figure, the input data is 7 slow #! The circuit (T) generates one signal flIA, the upper plural conjugate (to) is multiplied by 1, and the output of circuit 21 is multiplied by the next signal and sent to tap delay aj117 p1
The output of oia'tg timing (T) for each signal is input to a sign determination circuit 18 and an identification signal is extracted.

Now, the input data series is Muo・J'S, Mu, eJ'
! , Mu3.1'.

..., then the circuit 210 output Ohiro-1#in J#a+tmu,・■mu,+1・-A4・Am+16
'c′""-"' (1).In ζζ, (:) In the case of K continuous training, 'tH-'m-
Since it becomes 7±, 0-mu, ・mu m+1/"" (
2) (■) Tone O pigeonhole, #, 4 @ -#, = Q, so O-AM ``m+1@@ts (Ill) For data, % 'm+1- #@ is random, so 0-M m ” Anal @ ' ”””-”) (4 Take (23, (1, (4) is shown in the diagram #14 diagram (&),
O)), 61)), the code array input to the sign determination circuit 18 is the same as in the OjI example in Fig. 3. In the training, a series of negative values e, a series of positive values e in the tone, and a series of positive values e in the data. Since the positive value e and the negative value 1 are randomly detected, it is difficult to clearly identify them.

(Q 尭@O EFFECT) As explained above, according to the present invention, it is possible to clearly determine whether the signal that follows an instantaneous interruption is training, tone, or data by using continuous transmission 1-discriminating means. As a result, it is possible to apply each corresponding sequence.
)#It helps to increase the shear strength.

[Brief explanation of the drawing]

Fig. 1 is a general diagram of a tray base YG reconnaissance, Fig. 2 is a functional diagram of the present invention, and Fig. 3 (a) and (b) are configuration 11 of an embodiment of the present invention. Clear diagram, FIG.
1 in the figure IL diagram is an automatic gain*m device, 2 is an A/D converter,
5 is 611111111.4 halo filter, 5
is the equalization amount, 10 is the training detection circuit 1.11.
12 is a summation circuit, 14 is an absolute value circuit, 15 is a low-pass filter, 16 is a differencer, 17 is a Fi tap delay line, 18 is a sign judgment circuit, and 21 is a multiplication circuit.

Claims (2)

[Claims] (1) In a data transmission system that includes a transmitting device and a receiving device, receives a training signal, performs reception pull-in for the data signal, and transmits data, a means for identifying the training signal is provided, and a good signal is identified as the training signal. A training detection method characterized by performing reception pull-in by. (2) The training signal includes segments of a continuous transmission pattern; The training detection method according to item (alpha)), characterized in that the training detection method is characterized in that it identifies: