System and method high-data rate transmission
The invention relates to a method for performing high-data rate transmission and to a system using said method.
The invention refers to the area of the telephone and modem communication. The prior art comprises a number of methods for high-data rate transmission, e.g. the known ITU standards as for instance the specification V.34 or V.32 for analogous modems. These techniques try to guarantee a transmission rate of 33,6 kBit/s on an analogous telephone line. For this data rate a ratio of the power of the signal to the power of the noise S/N of about 40 dB is required. This is only achieved in modern communication systems but hardly in all telecom systems throughout the world.
The specification V.90 of the ITU even tries to expand the possibilities of modem connections up to 56 kBit/s. This goal is very seldom reached, especially in common telephone networks, i.e. with a lesser S/N ratio. The main limitation of said known modems is the low transmission rate for the noised channels, i.e. at a S/N ratio up to (15-20) dB the transmission rate does not exceed (10-15) kbit/s. Therefore the usable transmission rate often drops to 14,4 kBit/s or even below.
In the technical field of the information technology there exists a never-ending interest to increase the transmission rate at least until the limit of 64 kBit/s due to the restriction of the world-wide used PCM-equipment (pulse code modulation) .
It is therefore an object of the invention to improve the transmission rate, e.g. of a modem connection in the telephone
channel, up to 64 kBit/s, even in the case of worst S/N ratios down to 10 dB.
The object of the invention is therefore to provide a method of the above-mentioned type having higher transmission rate for any (uncompressed) signal.
The method according to the invention is characterised by the features of claim 1 and 2 for the realisation in the time domain and frequency domain, respectively.
The method is based on the use of the bandwidth of the isolated transmission line, the so called last mile, and allows increasing the data transmission rate.
Devices performing the above-mentioned methods are disclosed in claims 3 and 4, respectively.
These and other objects, features and advantages of the invention will become more apparent in light of the following detailed description of embodiments thereof, as illustrated in the accompanying drawings, in which:
Fig. 1 a schematic view of the transmitting part of a first system according to the invention, Fig. 2 a schematic view of the receiving part of a first system according to the invention, Fig. 3 a schematic view of the transmitting part of a second system according to the invention, Fig. 4 a schematic view of the receiving part of a second system according to the invention,
Fig. 1 and Fig. 2 show the transmitting and receiving parts, respectively, of the system device implementing a first method
of the data transmission disclosed in the course of this description.
Fig. 1 shows the device to implement the method of the formation of the signal for the transmission into the line 20 and its reception in the telephone exchange for reception 21.
An incoming digital information signal 22 enters the extractor 1 provided to separate the incoming information signal 22 into 8 bits 23 at a time. For convenience only two of the means 2 and 3 are shown. The eight outputs of the extractor 1 are connected to devices 2 forming the envelope. These envelope generating means 2 receive the tact (or clock) frequency signal 24. The outputs of the envelope generating means 2 are connected with the inputs of shifters 3 performing an SSB shift upwards. Every shifter 3 is connected to lines 25 being fed with increasing frequencies with the values fl to f8. Then the eight signals corresponding to eight information bits are added together in an adder 4 in order to be transmitted on the transmission line 20, the so called last mile, on the way to the service provider. Therefore each channel represents one bit of the message.
At the offices of the service provider there is the telephone exchange for reception 21. The signal is divided and enters eight stripe filters or bandpass filters 5, the outputs of which are connected with shifters 6 in order to perform a SSB shift downwards. Therefore a shift frequency 25 with the values fl to f8 as used in the shifters 3 is fed to the shifter 6. The output signal of each of the eight shifters 6 is fed to a demodulation means 7 which is fed with the known tact (clock) frequency 24. At the output of the demodulation means 7 the 1-bit signal is transformed and means 8 are used to form an 8-bit signal which is to be transmitted to the PC -equipment on line 26.
Fig. 2 shows a schematic view of the device implementing the reception of the signal from the telephone exchange station towards the reception end of the subscriber.
An incoming digital information is fed from PCM-equipment on line 27 to the extractor 1 provided to separate the incoming information signal 27 into 8 bits 23 at a time. For convenience only two of the means 2 and 3 are shown. The eight outputs of the extractor 1 are connected to devices 2 forming the envelope. The clock (tact) frequency is fed through the extractor 1 to these envelope forming means 2. The outputs of the envelope generating means 2 are connected with the inputs of shifters 3 performing a SSB-shift. Every shifter 3 is connected to lines 25 being fed with increasing frequencies with the values fl to f8. Then the eight signals corresponding to eight information bits are added together in an adder 4 in order to be transmitted on the reception line 30, the so called last mile on the other side of the transmission line 61, on the way to the end user.
At the location of the end user there is the modem for reception with an incoming line 61. The signal is divided and enters the input of eight bandpass or stripe filters 5, the outputs of which are connected with shifters 6 in order to perform a SSB shift. Therefore each shifter 6 is fed with a shift frequency 25 with the values fl to f8 as used in the shifters 3. The output signal of each of the eight shifters 6 is fed to a demodulation means 7 which is fed with the known tact frequency 24. At the output of the demodulation means 7 the 1-bit signal is transformed and means 8 are used to form an 8-bit signal which is to be transmitted as the received digital information signal on line 62.
It is now explained, in which way the system described in connection Fig. 1 and 2 is performing, starting with the bit separation means 1 receiving the digital information 22. For the transmission of the digital information signal 22 this information signal 22 is divided with the help of the per bit separation means 1 in such a way so that bits 1, 9, 17 of the information message form the first channel of the transmitted message, 2, 10, 18 and so on the second one, 3, 11, 19 and so on the third, 4, 12, 20 and so on the fourth, 5, 13, 21 and so on the fifth, 6, 14, 22 and so on the sixth, 7, 15, 23 and so on the seventh, and finally 8, 16, 24 and so on the eighth.
In each channel in the means 2 to form the envelope the building of the envelope occupying the range from 0 to 4 kHz takes place, wherein the tact frequency 24 of the count down points is equal to 8 kHz. Each of the eight envelopes is then transferred with the help of the SSB-means 3 to its respective frequency range, so that the first channel is shifted for fl, the second channel for f2,..., until the eighth channel for f8.
The signals of all these channels are summed in the adder 4 and are transmitted to the transmission line 20.
In the reception end of the telephone exchange station the filtering of the received signal 21 takes place with the help of the eight stripe filters 5. Then the signal of each of the eight channels is subjected to the reverse frequency shift with the help of the SSB-means 6 being fed with the relative frequency fl to f8, respectively. In the demodulation means 7 the extraction of the tact frequency 24, the reading of the count down points of the envelope and the restoration of one out of eight bits take place, i.e. in the first channel the extraction of the first bit takes place, in the second channel the extraction of
the second bit, ... , in the eighth channel the extraction of the eighth bit. Then in the byte forming means 8 the transformation of the received 8 bits into the 256 level signal with the tact frequency of 8 kHz is performed, i.e. in one countdown there are 256 graduations, which corresponds to the range of the message from 0 to 4 kHz, and then the signal comes to the PCM-equipment 26 with the processing rate of 64 kbit/s.
As in each of eight channels the two level signal is transmitted, the error probability of such a channel is known. The limit of the system at the convolution coding for the rates is , % is 7 dB.
Thus, at the S/N ratio of about 10 dB no errors are observed, and at S/N = 13 to 14 dB the error free reception of information takes place, i.e. the PCM equipment perceives the rate of 64 kbit/s at very low signal to noise ratio, i.e. 256 level countdowns at the given ratio will be formed without errors.
The 256 level signal from the PCM-equipment comes to the divider means 1 to separate the bits, where with the tact frequency of 8 kHz this signal is divided into 8 channels. In the first channel the most senior bit is identified, and in the eighth channel the most junior bit is identified. In each channel in the envelope forming means 2 the envelope carrying the information on the respective bit is formed with the tact frequency of 8 kHz, wherein the signal spectrum is located from 0 to 4 kHz. In the shifter 3 the signal of each channel is shifted upward the frequency for fl, f2, ... to f8 respectively. Then the signals are summed in the adder 4 and transmitted into the line 30 for the reception by the subscriber.
The signal which has passed the line 61 is filtered at the subscribers equipment with the help of band pass filters 5 with such an 4 kHz large filtering curve. Then it is shifted downward the frequency by the means of the SSB-shifter 6. In the means of the identification of the envelope 7 the signals are read by the count down points of each of the eight receiving channels and in the device of the formation of the information signal 8 the digital sequence with the data transmission rate of 64 kbit/s is formed.
The frequencies fl to f8 are preferably chosen in the following way: f1=104 kHz, fn=f1+ (n-1) *8 kHz. This gives rise to evenly spaced protection intervals of 4 kHz width which are in addition equal to the 4 kHz width signal. It is clear for someone skilled in the art that other frequency distribution may be chosen. The protection interval can be shorter (e.g. 2 kHz) or longer, it may not be equal in all cases. The starting frequency fl may be chosen totally different although the mentioned example provides the easiest approach.
Fig. 3 and Fig. 4 show the receiving and transmitting parts of the system implementing the second method of data transmission.
In Fig. 3 the system to implement the second method of the formation of the signal for the transmission into the line 20 and its reception 21 in the telephone exchange station is presented: There are provided thirty-two generators 31, 32, 33 of the support frequencies, generated by the formula ω/2π+nΔω/2π. The output of each generator 31, 32, 33 is fed to a phase rotator 34 for π/2. The signal of the generators 31, 32, 33 and the phase shifted signals from the phase rotators 34 are fed to sixty-four phase manipulators 35 with the function 0-π. The output signal of a converter 37 is equally fed to the
manipulators 35. The outputs of the manipulators 35 are presented to an adder 36. The converter 37 transfers the sequential code into the parallel one in the interval of T=2π/Δω. The converter 37 receives the code combination. This is the same code as in the first example.
On the incoming line 21 in the telephone exchange for reception the signal is fed to an A/D-converter 38 and to means 39 to extract the interval frequency. Both informations are fed to a FFT-means 40 performing a Fast Fourier transformation in the interval T=2π/Δω. The outputs of the FFT-means 40 are fed to sixty-four means 41 of per bit decision making, the output of which are connected with the conversion means 42 of the parallel 2n bit code into 256 level countdown.
For the transmission of the digital information the information signal 10 of the sequence of zeros and ones is converted into the 2n bit parallel code existing in the interval of T=2π/Δω with the help of means 37 in Fig. 3. Each pair out of the 2n bits of the parallel code carries out the phase manipulation 0-π in the phase manipulator 35 of the support frequencies, generated by generators 31, 32, 33 and their quadrature components, received at the output of the phase rotator 34 for π/2.
Signals from the inputs of the phase manipulators 35 are summed up and converted into the analogue signal in device 36 and are transmitted into the transmission line 20.
In the reception end 21 of telephone exchange system the conversion of the analogue signal into the digital one with the help of A/D-converter 38 takes place and the filter 39 extracts the signal with period T (the interval frequency) . Then FFT- means 40 accomplishes the direct Fourier transformation in
interval T. In the sixty-four means 41 the decision is made on the values of the phase of each frequency component by the rule: 0-π; π-0; π/2-0; 3π/2-0.
In conversion means 42 the conversion of each of the 8 bits out of 2n bit parallel code into the 256 level countdown is carried out.
Fig. 4 shows the system to implement the reception the signal 27 from the PCM-equipment for the transmission into the line 30 and its reception by the subscriber at 62. The A/D-converter 51 converts the analogue signal 27 into the 8-bit code with the tact frequency 44 of 8 kHz. The former 52 builds the parallel 2n bit code existing in the interval of T=2π/Δω. The thirty-two generators 53 and 54 generate support frequencies according to the formula ω/2π+nΔω/2π. The phase rotators 55 rotate the phase of the signal by π/2 and the phase manipulators 56 manipulate the phase by 0-π. Therefore, the sixty-four outputs of the phase manipulators 56 are connected with the inputs of the adder 57 before being transmitted to the line 30.
On the reception side 61 the signal is passed through the A/D- converter 38 and means 39 to extract the interval frequency. The resulting signal is transformed in the FFT-means 40 according to a Fast Fourier Transformation in the interval. Finally sixty- four means 41 make the per bit decision and the converter 42 converts the parallel 2n bit code into 256 level count down.
As for each frequency component the signal of the FM 4 is formed, the error probability of such channel is known. The limit of the system at the convolution coding for the rate of 1/2 is 4 - 4,5 dB, and for the rate of 3/4 is 6,6 dB, without coding 9 dB.
Thus, even at the S/N ratio of about 10 dB no errors are practically observed, and at S/N = 13 - 14 dB the error free reception of data takes place, i.e. the PCM equipment perceives the rate of 64 kbit/s at such low signal/noise ratios. At this the error probability for the reception is the same as for the transmission. Thus the given method of the high data transmission and reception rate allows having practically always the data transmission rate of 64 kbit/s in the telephone line.