JP2747562B2 - Modem communication system with primary and secondary channels - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a communication system, and more particularly, to a modem used for transmitting data via a transmission medium.

[Conventional technology]

Data modem communication systems are known which use a relatively high data rate main channel and a relatively low data rate secondary channel, wherein both channels share a common transmission medium. For example, U.S. Pat.
No. 955 has a main communication channel of 2400 bits / sec and 110 bits / sec.
A data communication system for telemeter information using a second auxiliary channel is disclosed. On the receiving side, high-pass and low-pass filters are used to separate the signals of the respective transmission channels.

[Problems to be solved by the present invention]

Applying a secondary channel transmission of a data modem communication system having a main channel data communication rate significantly higher than the above 2400 bit / sec rate, such as 14400 bit / sec, applies a low main channel such as 2400 bit / sec.・
It is considerably more complicated than that of the data communication rate. This is because the transmission media used, such as telephone lines, typically have limited bandwidth and are highly susceptible to jamming at high data rates.

Accordingly, it is an object of the present invention to provide a data modem communication system having primary and secondary channels suitable for high data transmission rates on the primary channel with a minimum of circuitry used for digital signal processing.

[Means for solving the problem]

To solve the above problems, the present invention provides a primary data channel transmitting data at a relatively high bit rate and a secondary data channel transmitting data at a relatively low bit rate. A transmission medium connected to the transmission medium, a modem transmission means for transmitting data through the transmission medium, a modem reception means having a main channel receiver and a secondary channel receiver, and connected to the transmission medium. An analog-to-digital converter; and first filter means connected to the output of the analog-to-digital converter and the input of the main channel receiver for suppressing the signal of the secondary data channel. A next channel receiver having second filter means connected to the output of the analog-to-digital converter; Processes the signal samples received over a plurality of times for each sample rate decreasing stepwise suppresses the signal of the main data channel lowpass
A modem communication system having primary and secondary channels characterized by including a digital filter is provided. According to the data modem communication system of the present invention, a single analog-channel is used for both the primary and secondary channels.
By using a digital converter and using a single low-pass filter multiple times at different sample rates, the main channel signal could be removed with high efficiency despite minimal circuitry.

〔Example〕

FIG. 1 shows a multipoint modem network 10 having a control modem 12 in communication with three tributary modems 14, 16, 18. actually,
The number of modems can be more or less. Control modem
12 is a 4-wire telephone line 20 including a 2-wire transmission line 22 and a 2-wire reception line 24
Connected to. The 4-wire telephone line 20 is connected to the respective tributary modems 14, 16,
It is connected to branch lines 20A, 20B, 20C connected to 18.
Therefore, the two-wire transmission line 22 is connected to the two-wire transmission lines 22A, 22B, 22C connected to the respective modems 14, 16, 18 via the branch point 26. The two-line receiving line 24 is for each modem 14,16,18
A two-wire receiving line 24A connected to receive signals from
24B and 24C are connected via a branch point 28.

Communication within multipoint network 10 occurs over a main channel at a relatively high bit rate, such as 14400 bits / second, and over a secondary channel at a relatively low bit rate, such as 75 bits / second. The main channel uses quadrature amplitude modulation (QAM), but other modulation schemes may be used. Secondary channels can transmit status, diagnostic and circuit management information. Therefore, four lines 2
0, 20A, 20B, and 20C can transmit and receive information on both the primary and secondary channels. In some applications, the configuration can be changed. One alternative configuration is a tributary modem 14,1
Although all 6,18 can receive information from the controlling modem over the primary and secondary channels, in some configurations the modem 14 only transmits over the secondary channel and the modems 16,1
8 can be able to transmit over both the primary and secondary channels.

FIG. 2 shows a plot of power spectral density versus frequency (Hz). The shape of the power spectral density of the main channel is indicated by a solid line 30, and the shape of the power spectral density of the secondary channel is indicated by a dotted line 32. The vertical scale of both plots is 24
It is shown in dB corresponding to 00 Hz converted to 0 dBm (0 dB milliwatt). From plots 30 and 32, the main channel (6
00-3,000Hz band) has a very wide spectrum,
The secondary channel has a very narrow spectrum (30-350 Hz band).

FIG. 3 is a block diagram of a transceiver of the control modem 12, and the configuration of the corresponding branch modems 14, 16, 18 is the same as that of FIG. The main channel transmitter 40 and the secondary channel transmitter 42 of the transmission unit transmit signals on respective output lines 44 and 46, the signals are added by an adder 48, and output on a line 50,
It is connected to the input of a digital-to-analog converter 52 whose output is connected to the transmission line 22.

Receive line 24 is an analog to digital converter 54 whose output is connected from line 56 to a digital notch filter 58.
And its output is connected to the main channel receiver 60. Notch filter 58 is a secondary channel signal (300
~ 350Hz band) and remove the main channel signal (600 ~ 30
(00 Hz band) through a main channel receiver 60. Therefore, the notch filter 58 acts as a high-pass filter,
Use a notch filter. Notch filters are known.

The output line 56 of the analog-to-digital converter 54 is also connected to a secondary channel receiver 62 to achieve high main channel signal suppression in the manner described below.

FIG. 4 is a block diagram of the secondary channel transmitter 42 (FIG. 3). A signal representing a mark or space symbol (eg, a high or low level signal) is provided to composite signal generator 72 via input line 70. Composite signal generator 72 is a binary continuous transfer FSK modulator. That is, it is 2
Frequency shift keying with a phase rotation corresponding to one frequency and a continuous phase change between successive symbol periods. The use of continuous phase FSK modulation can reduce the bandwidth and limit the effects of noise and distortion. Composite signal generator
72 generates a composite value sample at 9600 Hz on output line 74. In the figure, double connecting lines are used for complex values and single connecting lines are used for actual value quantities. Each composite value signal sample is -2π20 / 9600 radians corresponding to a negative frequency of −20 Hz and a positive frequency of +30 Hz, respectively, depending on whether the signal on input line 7 represents a mark symbol or a space symbol for the previous sample. It has a phase rotation of + 2π30 / 9600 radians. The mark symbol corresponds to 128 phase rotation of -2π20 / 9600 during the symbol period, and the space symbol corresponds to + 2π30 / 9600 during the symbol period.
Corresponds to a 128 phase rotation. After each symbol period (1/75 second, 128 samples), the same symbol and corresponding phase rotation can be generated, and other symbols and corresponding phase rotation can be generated.

The composite output from the composite signal generator is applied via line 74 to an IIR (Infinite Impulse Response) digital low-pass
Filter 76 removes signal components from the spectral side protrusions that interfere with the main channel signal. The IIR filter 76 shown in FIG. 5 includes a delay unit 100 and a multiplier 9 connected in a manner shown in FIG.
Including 4,96,98 and adders 90,92. The coefficients of the preferred embodiment are
C ₁₁ = 0.02395, C ₁₂ = 0.95209, and C ₁₃ = 0.5.

First, select an analog low-pass filter that has a minimum effect on the secondary channel in-band components (-37.5 Hz to +37.5 Hz) and a high suppression effect on out-band components above 150 Hz. Is determined by Then, the analog low-pass filter is converted into a digital low-pass filter by a known method.

The composite value output of the IIR filter 76 in FIG. 4 is supplied via a line 78 to a frequency converter 80 in the form of a composite multiplexer, which multiplexer
A signal exp (+ j2π (320n / 9600)) representing a 320 Hz composite value carrier signal at a 9600 Hz sample rate is received over line 82. Frequency conversion multiplier 80 generates real value signal samples on output line 84. Frequency converter 80 operates to achieve a 320 Hz spectral shift of the signal spectrum. FIG. 6A shows the signal spectrum at the input of the IIR filter 76. The center of this signal spectrum is about 0 Hz. FIG. 6B shows the signal spectrum output on the output line 84 of the frequency converter 80 after the removal of the side protrusions and the spectral shift, and is shown in more detail than the dotted line 32 in FIG. 2 showing the overall shape of the secondary channel signal spectrum. . Output line 84 is line
The actual value signal sample of the secondary channel is connected to the digital-analog converter by an adder 48 (FIG. 3).
It is added to the actual samples at the output line 44 of the main channel transmitter 40 for feeding to 52. Main channel transmitter
The 40 also works at 9600Hz sample rate.

The filtering operation of the secondary channel receiver 62 (FIG. 3) will be described with particular reference to FIGS. FIG. 7 shows the principle of the filter operation, and FIG. 9 shows its actual implementation. First, in FIG. 7, the signal on line 56 from the analog-to-digital converter 54 (FIG. 3) is processed at a 9600 Hz sample rate and the low pass IIR digital filter 110A is designed to remove frequencies above 800 Hz. Supplied to At the output of filter 110A every fourth sample, it is selected to provide a real valued signal at a sample rate of 2400 Hz appearing on output line 112 of filter 110A. Line 112
This reduced sample rate signal of the input signal exp (−
j.2π (320.n.4./9600)) is provided to a frequency converter 114 in the form of a multiplier which is provided on line 116. The resulting composite value output signal is passed via line 118 to filter 110
Lowpass I, identical to A but processed at 2400Hz rate
The signal is supplied to the IR digital filter 110B. Filter 11
0B is 250 Hz before frequency conversion occurs in frequency converter 114
The frequencies above 200 Hz corresponding to the above main channel components are removed.

Every other sample (even) at the output of filter 110B is chosen to provide a 1200 Hz sample rate on line 120. The signal on line 120 is provided to a low-pass IIR digital filter 110C that is identical to filters 110A and 110B but is processed at a 1200 Hz rate. The filter 110C removes a frequency of 100 Hz or more corresponding to a main channel component of 420 Hz or more before performing frequency conversion by the frequency converter 114. Every other sample (even number) at the output of the filter 110C will be described later,
The detector 124, which produces an output signal representing the bits, is selected to provide a 600 Hz sample rate signal provided over line 122. The above filters 110A, 110B,
From the 110C description, it was found that each filter rejects frequencies greater than 0.08 times the sample rate at which it is processed. Thus, all frequency components that are more than half of the new sample rate are removed, thereby avoiding aliasing degradation (frequency holdover). In addition, the filters 110A, 110B, 110C are 9600 Hz, 240
The filter is processed at a sample rate of 0 Hz and 1200 Hz, so the low-pass processing of the filter is 500-800 Hz, 125-200 Hz, and 62-
It has a transition band at 100 Hz. For such a filter, the action on the sample rate is identical. Such filtering allows the use of simple filters with low cut-off frequencies and small transition bands.

The above-mentioned secondary channel receiver filter operation is performed in 8A, 8B,
It is well understood in FIG. 8C. FIG. 8A shows the filter 11
7 is a plot showing signal spectrum level versus frequency at the input of 0A. Solid segments 130 and 132 show the main channel signal spectrum. Dotted lines 134 and 136 represent the secondary channel signal spectrum and dotted line segment 13
8 represents a filter characteristic.

FIG. 8B is a plot showing signal spectrum level versus frequency at the input of filter 110B. Solid line segments 140 and 142 represent the main channel signal spectrum, dotted line segments 144 and 146 represent the secondary channel signal spectrum, and dotted line 148 represents the filter characteristics.

FIG. 8C is a plot showing signal spectrum level versus frequency at the input of filter 110C. Dotted line 15
0 indicates the secondary channel signal spectrum, and dotted line 152 indicates the filter characteristic. Filter during 1 symbol period of 1/75 second
The processing performed in 110A, 110B, and 110C is shown in Table A below.

The filters 110A, 110B, 11C have the same structure and the same coefficient. This allows the use of one filter in the implementation of the filter operation of FIG. Its implementation is shown in FIG. 9 and uses a single IIR digital filter 110 in the preferred embodiment of the present invention which implements the principles of FIG. 7 in the manner shown.

The IIR filter 110 that performs the functions of the filters 110A, 110B, 110C (FIG. 7) of FIG. 9 includes an input line 160 connected to a three-position input switch 162 and an output line 164 connected to a three-position output switch 166. Connected between The three terminals of the third three-position switch 168 are connected to respective storage devices 170, 172, 174. The storage devices 170, 172, 174 contain filters corresponding to periods during which the filter 110 is processing at different processing rates.
Used to store internal samples from the 110 delay elements. Switches 162, 166, 168 operate synchronously so that the corresponding terminals having positions labeled 1, 2, 3 can execute simultaneously.

The filters 110A, 110B, and 110C in FIG. 7 correspond to the filter 110 by switches 162, 166, and 168 at positions 1, 2, and 3, respectively.
For example, 1/9600 at position 1 of switches 162, 166, 168
For a sample period of seconds, the stored value of the delay element is read from storage 170 and provided to the delay element of filter 110. There, the filter is processed and the new contents of the delay element are stored in the storage device 170. For each fourth sample period when switch 166 is in position 1, the output of filter 110 is used as an input to frequency converter 114. The switch is in position 2 and the input to filter 110 is drawn from position 2 of input switch 162 to store
The contents of 72 are read into the delay element of filter 110. Thereupon, the filter 110 is processed and the contents of the delay element are read out and stored in the storage device 172.

Because of the odd output samples of the filter 110 when the switch is in position 2, the process returns to processing at switch position 1, but for each even output sample of the filter 110, the output sample of the filter at switch position 2 is As input to the filter 110 at position 3 (once every eight periods of 1/9600 seconds). Thus, the contents of storage 174 at switch position 3 are read into the delay element of filter 110, the filter is processed, and the contents of the delay element are stored in storage 173.

Due to the odd output samples of filter 110 at switch position 3, filter 110 at switch 1 position
But for each even output sample of the filter 110 at position 3, the output sample of the filter is provided as input to the detector 124 (1 for 16 periods of 1/9600 seconds).
Times) used.

FIG. 10 illustrates an implementation of the IIR low pass filter 110 of FIG. The filter 110 is connected between the input line 160 and the output line 164, and includes adders 180 to 210, multipliers 212, 214,
216, 218, 220 and (using the respective coefficients _{C 21, C 22, C 24} , C 25), using a delay element 222-230. Each element of the filter 110 is connected as shown in FIG. The IIR low-pass filter 110 is an elliptic filter whose design is
E International Symposium on Circuits and System, April 1981, pages 550-553, R. Ansari and B. Liu described "A Class of Low Noise Computationally Efficie
It is preferable to use the principle of "nt Recursive digital filters" and the filter structure. The coefficient values used in the preferred embodiment of the present invention are as follows.

Path _{C 21 = -0.79235 C 22 = 0.91922} C 23 = -0.94024 C 24 = 0.74005 C 25 = -0.95751 These factors are described in the article above Ansari and Liu ^- and stop ^- request for the band acts And elliptic filter design technology and IEEE Transactions on Acoustics, Speec
h and Signal Processing, vol ASSP-33, No. 1, 1985 2
"A Class of Low-Noise Computati"
onally Efficient Recursive Digital filters with Ap
plications Sampling Rate Alterations ".

11A and 11B show plots illustrating the transmission action of the total secondary channel transmission path including radio wave distortion (amplitude distortion in FIG. 11A and delay distortion in FIG. 11B). In the amplitude distortion of FIG. 11A, a solid line 240 indicates no distortion, and a dotted line 242 indicates the worst distortion. In FIG. 11B (delay distortion), a solid line 250 indicates no distortion, and a dotted line 252 indicates the worst state. 350Hz with different radio lines
The amplitude at 300 Hz and the delay distortion for that distortion in vary as follows.

Amplitude distortion: 0 to 1.5 dB Delay distortion: 0 to 1.3 msec. The total amplitude delay distortion of the filter operation in the secondary channel transmitter and receiver is concentrated around about 320 Hz. This makes the difference in amplitude and delay distortion at 300 Hz relative to those at 350 Hz as described below.

Amplitude distortion: -0.5 dB Delay distortion: -0.3 msec. Similarly, the sum of the effects resulting from transmitter filter operation, telephone line distortion, and receiver filter operation for different telephone lines at 300 Hz versus 350 Hz is: Changes to

Amplitude distortion: -0.5 to 1.0 dB Delay distortion: -0.3 to 1.0 msec. Thus, the filter operation provided in this preferred embodiment achieves pre-compensation for the distortion difference between 300 Hz and 350 Hz,
Therefore, it allows for more reliable detection on telephone lines even when high distortion levels occur.

FIG. 12 is a block diagram of the detector 124 of FIGS.
The detector 124 has an input line 122 connected to a -20 Hz correlator 260 and a +30 Hz correlator 262. The -20 Hz correlator 260 includes a multiplier 264 that receives the input signal on line 122 at a 600 Hz sample rate over line 266 and has the following multiplication factor:

exp (−j.2π (−20.1.16) / 9600) where 1 = 0,1,..., 7 The output of the multiplier 264 is 8 according to the following equation.
It is connected to an adder 268 which performs continuous addition for successive samples.

Once every 1/75 second, with 8 input samples, s _n = s _k , s _{k + 1} ,... S _{k + 7} +30 Hz correlator 262 via line 272 at 600 Hz
A multiplier 270 receives the input signal from line 122 and includes the following multiplication factors.

exp (−j.2π (30.1.16) / 9600) where 1 = 0,1,..., 7 The output of the multiplier 270 is 8 according to the following equation:
It is connected to an adder 274 which performs continuous addition for successive samples.

Once every 1/75 second, 8 input samples, s _n = s _k , s _{k + 1} , ... s _{k + 7} The combined output 276,278 of each adder 168,174 carries the signal at 75Hz rate, nom comparator (Comparator) 28
Connected to 0. The square vector length of the correlator output in the comparator 280 is settled and compared. So the comparator
280 determines if a -20 Hz signal is likely to occur from the +30 Hz signal as an input to the detector 124 during the last 1/75 second. Thus, the correlator 260 output = x ₁ + jy ₁ correlator 262 output = x ₂ + jy _2, square vector length calculates ^{_{v 1 2 = x 1 2 +}} y 1 2 v 2 2 = x 2 2 + y 2 2.

If v ₁ ² > v ₂₂ , -20 corresponding to the mark symbol
Hz signal is detected. Further, if v _{1 ^2>} v ² a ^2, + 30 Hz signal corresponding to the space character is detected.

The above description is the normal operation of detector 124 during transmission. However, in order to perform an effective initial detection when transmitting an initial pattern consisting of 14 mark symbols followed by 2 space symbols, both correlators are initialized to -20 Hz with a timing difference of a half symbol shift (1/150 second). Rotate. Thus, the adder 268 of the correlator 260 performs the following addition, The adder 274 of the correlator 262 performs the following addition.

When the correlation for -20 Hz drops in response to the presence of the space symbol, correlator 262 is changed to correlate at +30 Hz by changing the input signal on line 272.
Thus, the best first sample to begin normal operation of the detector 124 is drawn based on such a drop in correlation to -20 Hz. Hence, two correlation measurements during subsequent normal operation of detector 124 are made at the appropriate time.

In summary, the preferred embodiment of the present invention achieves high performance and can be formed with minimal circuitry,
It has been found that it includes a secondary channel receiver that can be implemented with a digital composite signal processed by the circuit. Thus, the modem of the receiver only requires one analog-to-digital converter, since the same signal samples as for the main channel receiver can be used for the secondary channel receiver. Moreover,
Accurate filtering and high suppression of the main channel signal frequency can be achieved without the frequency holdover caused by using a single filter that is processed many times at different sample rates. Further, the signals and timing of the secondary channels are independent of the main channel,
The secondary channel is not interfered by the main channel and can operate without degrading the main channel during transmission.

[Brief description of the drawings]

FIG. 1 is a block diagram of a data modem communication system including primary and secondary channels, FIG. 2 is a plot of power spectral density for primary and secondary channels, and FIG. 3 is a primary and secondary channel transmission of each modem. FIG. 4 is a block diagram of a secondary channel transmitter, FIG. 5 is a diagram illustrating an IIR digital filter included in the secondary channel transmitter, FIG. 6A. And FIG. 6B is a diagram showing the signal spectrum at different locations of the secondary channel transmitter, FIG. 7 is a block diagram showing the principle of the filter of the secondary channel receiver, FIG. 8A, FIG. 8B and FIG. Secondary channel receiver 3
FIG. 9 is a graph showing the signal spectrum of a place; FIG. 9 is a view showing the use of a signal IIR digital filter in a secondary channel receiver; FIG. 10 is a block diagram of the IIR digital filter of FIG. 9; FIG. 11B is a diagram showing amplitude distortion and delay distortion of the secondary channel, and FIG. 12 is a diagram showing a detector of the secondary channel receiver. In the figure, 10: Multipoint modem network, 12: Control modem, 1
4,16,18 ... tributary modem, 20 ... 4-wire telephone line, 22 ... 2
Line transmission line, 24 ... 2-line reception line, 26, 28 ... Branch point, 30 ...
... power spectral density of the main channel, 32 ... power spectral density of the secondary channel.

Continuation of front page (56) References JP-A-62-217740 (JP, A) JP-A-62-51821 (JP, A) JP-A-63-51713 (JP, A)

Claims (1)

(57) [Claims] A main data channel for transmitting data at a relatively high bit rate and a relatively low bit rate.
A transmission medium having a secondary data channel for transmitting data at a rate; a modem transmission means connected to the transmission medium for transmitting data through the transmission medium; a main channel receiver and a secondary channel receiver. A modem receiving means, an analog-to-digital converter connected to the transmission medium, an output of the analog-to-digital converter and an input of the main channel receiver for suppressing signals on the secondary data channel. The secondary channel receiver comprises second filter means connected to the output of the analog-to-digital converter, wherein the second filter means comprises: Processing the signal samples received multiple times at each decreasing sample rate to suppress the signal on the main data channel; Characterized in that it comprises a pass digital filter,
Modem communication system with primary and secondary channels.