RU2254673C2 - Data transfer method - Google Patents

Abstract

FIELD: information science; digital data transfer systems and radio communications.

SUBSTANCE: proposed method includes generation of signal to be transmitted using message source that has two similar-length double-level characters characterized in similar probability of their choice, as well as two similar-length additional characters with added third electrical level; logical one bit is encoded by both positive and negative signal changes in center of bit interval and logical zero bit, by both positive and negative signal changes at first one third of bit interval with third level being added.

EFFECT: enhanced speed of code capable of partial solution of transmission line overload problem.

1 cl, 8 dwg

Description

The invention relates to computer science and can be used in digital information transmission systems and in radio communications.

The most important indicator of the development of information technology is the constant increase in the volume of transmitted and processed information, which is caused by the rapid development of computer science in the last decade and, as a result of this, an ever-increasing load on the transmission lines. An increase in the speed of transmitted information, as is known, directly depends on the channel capacity - an insufficient level of which, when it is less than the performance of the message source, leads to the loss of some information. An increase in channel capacity due to the expansion of bandwidth is not always technically possible, as is the case with the use of telephone lines on the Internet. Therefore, sometimes the most effective way to increase the speed of transmitted information is to maximize the approximation of the source performance to the existing channel bandwidth. Source performance, in turn, can increase either through the use of codes that reduce its redundancy to a minimum, or by the introduction of additional independent sources. An example of such a method of increasing the speed of transmitted information while maintaining the channel bandwidth is the proposed modified code.

The objective of the invention is to modify the code using one transmission channel, which is a phase-shifted code (Shevkoplyas BV Microprocessor structures. Engineering solutions. Radio and communication, 1990) by introducing two characters into the alphabet of the source forming this code (Fig. 4a) on the basis of which two independent sources are obtained using different encoding methods and having four characters in their alphabets with an equal probability of their choice, of which two form a logical unit, and two logical zero for transmitted binary messages (figb, c).

The technical result of the invention is the creation of a high-speed code that can partially solve the problem of congestion of transmission lines. If the data of the sent message comes from different sources, the number of bits arriving at the channel input per unit time increases on average 1.6 times in comparison with the phase-shifted code or AMI codes. Another area of application of the received code can be its use in radio communication as a code with high noise immunity, when a message sent via a radio link can be duplicated by transmission by a second source. In this case, the influence of interference, which can be represented as a separate source, is taken into account. The effect of interference leads to the fact that with known events (data) there is some uncertainty of their outcome at the output of the channel. The degree of uncertainty caused by the action of interference can vary significantly for each source individually, i.e. the entropy of noise at the output of the channel can be destructive for the message coming from one source and insignificant for another. In any case, at any level of interference, the reliability of the received message, and therefore the noise immunity of the signal transmitting information from two independent sources, doubles. The received code satisfies the basic requirements for transmitting codes, i.e. it does not contain a constant component in its spectrum, information is transmitted through only one channel and is completely restored by the decoder for any message duration.

The information capabilities of the resulting code increase significantly due to the combined effect of two independent sources that use different coding methods. In the first of them, the phase-manipulated method is used (figa, b). The encoding of the message in the second source is based on a method in which the bits are represented by a bipolar signal and a bipolar signal with the addition of a third "zero" level (application No. 2000104570/09 (004608)). Conventionally, such a code can be designated as a code TPZ (two poles - zero) (figa, b). Then H (fm) + H (tpz) = H (tm), where H (fm) is the source entropy using phase manipulation; H (tpz) - source entropy using the TPZ code; H (tm) - entropy of a system consisting of two sources (tm - abbr. From two messages). The phase-manipulated representation of the signal in the resulting code has a distinctive feature when comparing it with the Manchester-11 code - the bits of both logical values can have redundancy in the form of an additional electric level (third "zero" level), (figa, b). A three-level representation of the signal in the received code in comparison with the signal in the TPZ code gives redundancy, expressed in the representation of bits as positive and negative differences in the bit interval when switching from one polarity to another (figa, b). It follows from what has been said that if we consider a combined signal generated by two sources only as phase-shifted or only as represented by the TPZ code, in both cases it will lose to the TPZ and Manchester II codes in their “classic” version. In the first case, a second low-frequency component appears in the spectrum of the modified signal (Fig. 2 and Fig. 3), in the second case, the introduction of a third level eventually reduces the bit rate (Fig. 1 and Fig. 2).

The number of characters in the alphabets of both sources will be four with the same probability of their choice (figb, c). Of these, two characters transmit a logical unit, and two logical zero. In the source transmitting the phase-shifted signal (Fig. 4 c), the logical unit may be encoded with a negative signal drop in the middle of the bit interval or a negative signal drop in the first third of the bit interval, taking into account the added third level. The same applies to logical zero with the only difference being that in this case a positive difference is used. When transmitting from a second source of a bipolar three-level signal, a logical unit can be transmitted by two symbols of the source alphabet with different alternating polarities - both from minus to plus and from plus to minus (fig.4b). When transmitting logical zero, the same symbols are used, alternating in polarity, but with the addition of a third level. In other words, the bits of each source are transmitted by the same four characters, but in different combinations and encoding methods relative to the logical value of the bit. Since both logical values are equally probable, the entropies of both sources will be the same and equal to unity. It can be noted that the transmission of information from two sources by a combined signal is possible due to the use of two characters to form a bit, i.e. each bit is given a choice of one of two characters. It is theoretically possible to transmit an N-channel signal, provided that the sending of a bit of one logical value is carried out by N characters. At the same time, the presence of four symbols implies the assignment to each of them of an individual information value in the form of a separate event. In this case, the entropy of a system consisting of two sources becomes equal to the sum of the entropy of one source and the conditional entropy of another: H (tm) = H (fm) + H (tpz / fm) or H (tm) = H (tpz) + H (fm / tpz), i.e. will consist of two interdependent sources.

The durations of all three levels will be equal, while the bit durations of the same logical value may differ by one third, which implies that the bit rate, given that the initial phase for the signals of both channels is the same and equal to zero (Fig. 2a, b) will be the same for both the first and second sources. The phase-manipulated code alphabet consists of two characters, the duration of which can be designated: Ta - for the "single" bit and Tb - for the "zero" bit. If two characters are added here, including the third additional level, the bits of both logical values acquire a new encoding method, which will be equivalent to the appearance of a new source working in parallel with the existing one. In this case, the bit durations with equal probability can take two values - Ta and (3Ta) / 2; Tb and (3Tb) / 2. Then the averaged bit duration for both sources will be [(3Ta) / 2 + Ta] / 2 = Ta + Ta / 4 and [(3Tb) / 2 + Tb] / 2 = Tb + Tb / 4. If the message contains n bits, then the total time spent on its transmission will be equal to n (T + T / 4), where T = Ta = Tb is the bit duration. For the phase-shifted code, respectively, nТ. Then nТа / n (Т + Т / 4) = 4/5, i.e. the bit rate for each conditional channel is reduced by an average of 20% compared to the "Manchester II" code. In general, the speed of information transfer by the received code at a constant transmission line capacity increases on average 1.6 times.

According to Shannon’s theorem, if the transmission line capacity exceeds the amount of information coming from the source to its input per unit time, then by appropriate coding all incoming information can be transmitted over the line at a speed arbitrarily close to its capacity. In this case, the modified signal is an encoding example, the result of which will be an increase of 1.6 times in the information coming from the source compared to the phase-shifted signal along a line adapted to its characteristics, i.e. can be transmitted on the same channel as the FM code, since it contains the same transmit frequencies in its spectrum. The approximate spectrum width of the phase-shifted code is determined by the simplified file: P = 2.6Δf + 1.5B, where B is the bit rate, Δf is the difference between the upper and lower frequencies. For example, if the upper frequency is 10 MHz, then Δf = 5 MHz, and B = Fn = 5 MHz, then P = 2.6 × 5 + 1.5 × 5 = 20.5 MHz. The emission spectrum of the signal transmitted in the modified code in the case when the third level is absent becomes equal to the spectrum of the phase-shifted signal, i.e. this will be the maximum possible width of the spectrum for the resulting code. In the general case, if we calculate the frequency band from the shortest radio pulse of the modified signal, then at a frequency of 10 MHz we will have P = 1 / t = 10 MHz, where t is the duration of the radio pulse, equal to 0.1 μs.

One of the possible options for obtaining a modified code is shown in a simplified diagram of the encoder, Fig. 5 The encoder works as follows: the contents of the memory (RAM), on the basis of which the phase-shifted code of one channel is built, is bitwise read by the gates on a positive edge, previously modulated by the bits of the second channel, as shown in diagram "a)" of Fig.6. In order to eliminate the “sticking” of bits of one logical value, the bits read from the memory pass additional confirmation on elements 1 and 3 for units and zeros, respectively (in Fig. 6 of the diagram “c)” and “e)”). In this case, the "single" bits are supplied to the conjunctor 1, and the "zero" bits through the inverter 2 to the conjunctor 3 (diagrams "b)" and "e)" in Fig.6). Single-oscillators AG4 and AG5 using the elements “exclusive or” 6 and 7, respectively, form pulses that are phase shifted with respect to the pulses at the outputs of elements 1 and 3 by half the period of the “single” bit of the modulating signal (diagrams “g)” and “g ) "in Fig.6). The received signals are fed to opamps having supply voltages of different polarities depending on which bit is required to be received (U8, 9, 10, and 11) and then to a common output amplifier (U12), from the output of which the resulting signal is taken (diagram "h) "in Fig.6).

8. Signal decryption is somewhat more complicated. It consists of two stages: first, the modulating signal is reproduced - it is the signal of the first channel. The accuracy of its restoration is provided by the fronts of the phase-manipulated signal component related to the second channel, which is restored at the second stage of transformations by the recreated modulating signal of the first channel. A simplified diagram of the decoder is shown in Fig.7, and its operation is illustrated by the diagrams in Fig.8 in the time interval Tk-To. Initially, the input signal is converted from bipolar to unipolar (diagram "a)" of Fig. 8), and already in this form is fed to the op-amp (U1, 2, 3, and 4 in Fig. 8). U1, 2, 3, and 4 are analog comparators in which the reference voltages at U1 and U2 are close in magnitude to the amplitude. Diagram "b)" shows the double pulses at the outputs U1 and U2. Differences in their durations along the edges (diagram "c)") are highlighted at the output of element 5. The reference voltages at U3 and U4 are also close in value - one of them is slightly larger than the third intermediate voltage level, equal to half the amplitude level, the second is slightly less. The signals at their outputs are shown in diagrams d) and d), respectively. The result of the above transformations will be the appearance on the element 6 pulses synchronous with the appearance of the third intermediate levels (diagram "e)". With their help, filtering the "extra" pulses at the output of element 5 (diagram "c)"), for which they receive an increment in the form of shaded areas on the inverse output of the single-shot AG10 (this is not shown in Fig. 8). Diagram "g)" shows the filtered pulse sequence. In parallel with this, in the circuit "AG9 - element 8", negative-pulse pulses are removed. As a result of the transformations, the pulse sequence “c)” takes the final form, as shown in diagram “h),” and the reconstructed signal of the first channel is removed from the positive output AG9 (diagram “and)”). The last operation will be to decrypt the phase-manipulated code of the second channel and bring it to the form shown in diagram "k)". This can be done using the “exclusive or” element of the CMOS series, which will perceive the intermediate levels as logical zeros. Further conversion, not shown in the diagram, looks quite simple - the decoded signal in the diagram "k)" is entered into the memory or shift register with a modulated clock sequence (diagram "i)" or "h)"), which after a correctly selected time delay can be used already as reading from the same temporary storage device and thus again participating in the reproduction of the original two-channel code, i.e. serve as a link between the decoder and the encoder.

Claims (1)

A method of transmitting information using a single transmission channel, in which the transmitted signal is generated using a message source, the ensemble of which contains two equally long two-level symbols with the same probability of their choice, where the first symbol is a positive difference in the bit interval used to transmit logical zero, and the second is a negative difference, designed to transmit a logical unit, characterized in that two identical in duration up to additional characters with the added third electrical level equal to one third of the bit interval of the added characters or half of the bit interval of the available characters, in which the polarity is changed by a positive difference on the first one and the negative on the second, as a result of which the bits of the generated binary message can be encoded by two characters and in two ways, namely: a bit of a logical unit is encoded with a negative edge difference in the middle of the bit interval or a negative edge in the first third of the bits interval with the addition of a third level, and a logical zero bit - a positive difference in the middle of the bit interval or a positive difference in the first third of the bit interval with the addition of a third level, the bits of a logical unit are encoded with both positive and negative signal drops in the middle of the bit interval, and the bit logical zero - both positive and negative signal drops in the first third of the bit interval with the addition of a third level.

Images