EA033569B1 - Method of digital data transfer - Google Patents

Abstract

The invention relates to the field of digital communication and control devices, and it may find use in constructing highly-effective intellectual systems with remote transfer of data streams. Utilization of the device allows to create a high reliability system for multi-channel transfer of digital data, avoid laying multiple highly expensive cable lines or allocating sufficient number of valuable radio frequencies carved out of the extremely deficit radio frequency spectrum. The method for data transfer is based on forming a Hadamard matrix at the transmitting side of the line, its dimension corresponding to the number of data sources, and the elements of its strings representing the values of orthogonal Walsh functions, attributing each such function to specific information source. Then, next bits with the same numbers are picked from each initial data package, converted into time-coded representations using the abovementioned orthogonal Wash functions and the data part of the bit frame is filled with such coded representations. After that an error-tolerant block code is superimposed by forming the control part of the bit frame, both parts of the bit frame flagged and/or attributed with "start-stop" signals and then the data are sent down the communication line. Once the adornments are removed at the receiving end of the line, the data and control parts of the bit frame are checked for correctness, adjusted, if necessary, and the bit values of individual data packages are restored using the abovementioned Wash functions and the newly formed data packages are transferred to the data user.

Description

The invention relates to the field of creating digital communication and control devices and can be widely used in the construction of highly efficient intelligent systems with remote transmission of information flows, as well as in the development of multi-functional mechatronic systems for various industries, in particular for mechanical engineering. Using the present invention allows to create a multi-channel digital information transmission system with high reliability, while avoiding the laying of a large number of expensive cable communication lines or allocating a sufficient number of frequencies from the extremely scarce radio frequency spectrum.

Currently, in the practical implementation of projects to create digital information transmission systems and control systems, frequency, time or code division of channel resources between subscribers - sources of information flows is used. The problem of collective use of information transmission channels by groups of subscribers is most effectively solved by introducing the principles of code division multiple access (CDMA).

According to their purpose, well-known design decisions in the field of code division of channels, as a rule, belong to the field of creating cellular communications and, to one degree or another, repeat the technical methods of the widely used recommendation IS-95 [1]. The main feature of the recommended technical solution is the application of the principle of correlation signal processing on the receiving side of the communication line in order to isolate individual messages from different data sources sent by different information sources.

A significant disadvantage of the known method of code division multiplexing is the relatively low level of reliability of identification of information flows from a large number of simultaneously functioning sources on the transmitting side of the communication line.

In accordance with this, in order to increase the noise immunity of known communication systems, the process of correlation processing of the received signal is improved. In particular, the input mixture containing the signals of all channels is multiplied in each channel with the corresponding synchronous reference signal, the result of the multiplication is filtered, its detection and simultaneous comparison of the selected voltage envelope with two (low and high) threshold values [2]. The filtered signal is multiplied with the same synchronous reference signal, whereby the signal received by this channel is restored, which is subtracted from the mixture at the input of the channels in which a temporary excess of two thresholds is not detected.

The essence of the known method of transmitting information with code division multiplexing is that when it is implemented, the detection of powerful signals of nearby subscribers and the formation of estimates of structural interference from them are ensured. At the same time, the structural interference estimates generated in each channel are used to compensate for the corresponding structural noise in the input mixture. In accordance with this method, the signals of remote subscribers are processed in the information transfer channels after the powerful signals of the near subscribers (structural interference) are compensated in the input mixture.

A disadvantage of the known method of transmitting information is the relatively weak suppression of spatially concentrated and / or narrowband interference. The presence of narrow-band interference leads to a significant reduction in the capacity of a communication system with code division multiplexing. This fact is explained by the fact that all the code channels for transmitting information share a common frequency range, upon defeat of which all code channels are blocked immediately. A decrease in the capacity of a communication system is especially unacceptable when constructing mechatronic systems used to control potentially dangerous objects, for example, energy and defense ones.

In addition, the manifestation of this drawback is unacceptable in dual-use communication systems, that is, in systems that simultaneously provide commercial services and are used by government agencies for official communication. Especially significant may be the loss of capacity of the communication system during its operation during natural disasters, major industrial accidents or other emergency situations. In such a situation, a large number of different means of communication work at the same time, primarily narrowband, which cannot be coordinated and their frequencies can be planned due to an emergency.

Known methods for ensuring noise immunity of information transmission systems with code division multiplexing from narrowband spatially concentrated interference. In practice, the following methods are used to combat the indicated interference: signal spectrum rejection (frequency selection) and spatial selection using multi-element antenna arrays. With this in mind, the closest technical option to the claimed method closest in technical essence to the claimed method was chosen to protect communication systems from narrow-band spatially concentrated interference by means of joint spatial-frequency selection [3].

The prototype method of transmitting digital information with code division multiplexing is implemented according to the well-known general algorithm. On the transmitting side of the communication line, the original data packets are received from digital information sources, an error-correcting code is applied to the indicated data packets, they are framed with flags and / or start-stop signals, and the data are sent to the communication line. At the receiving side of the communication line, said frame is removed, data decoding is performed, and

- 1,033569 data packets to the consumer of information.

In the known method, increasing the capacity and noise immunity of communication systems with code division multiplexing under the action of spatially concentrated or narrow-band interference is achieved through the use of joint spatial-frequency selection. Joint spatial-frequency selection is carried out using an adaptive antenna array with an arbitrary number of elements on the receiving side of the communication line, which is necessary for spatial selection in each of the frequency channels. In addition, the phase-frequency structure of the useful signal and the optimal frequency weighting are also corrected. The solution to these problems is achieved by introducing algorithms for calculating weight coefficients and algorithms for reconstructing the phase structure of a useful signal and optimal frequency weighting when combining the signals of frequency channels.

The disadvantage of this method is the relatively high complexity of the algorithms used for signal processing. Frequency selection is inevitably associated with the loss of part of the energy of the useful signal and distortions of its autocorrelation function, which leads to an additional decrease in its quality. The use of spatial selection methods leads to the need to use complex antenna arrays (for example, adaptive ones), to increase the complexity of the receiving paths of communication lines, and, accordingly, requires complex adaptation algorithms. In addition, the algorithms used in this case do not have sufficient speed for working in real information transfer channels, for example, at high speeds of the movement of the information source or change in the state of the control object.

Thus, correlation signal processing at the receiving side of the communication line is the dominant way of organizing multi-channel communication in code division multiplexing. Known methods for improving such information transfer systems do not affect this basic principle of their creation. Most of the options for improving the known method of transmitting digital information relate to increasing the reliability of decoding data on the receiving side of the communication line. As a rule, such a rationalization of the communication system is carried out due to a significant complication of equipment and algorithms for processing received signals.

The developed method for transmitting digital information leads to the rejection of the use of correlation processing of received signals and is implemented exclusively by software. The proposed method can significantly increase the reliability of data transfer even when using channels with a relatively low quality and high values of the probability of occurrence of bit errors (for example, wired communication channels and radio channels).

The method of transmitting digital information is that a Hadamard matrix is formed on the transmitting side of the communication line, the order of which corresponds to the number of information sources, and the line elements display the values of the Walsh orthogonal functions and assign each function to a specific information source. Then from each source data packet the next bits with the same numbers are selected, converted into clock code representations using the above Walsh functions, and the information part of the bit frame is filled with the received values of the clock code representations. After that, a block error-correcting code is imposed on it by forming the control part of the bit-frame, frame both parts of the bit-frame with flags and / or start-stop signals and send data to the communication line. After removing the specified frame, the correctness of the information and control part of the bit frame is checked on the receiving side of the communication line, if necessary, they are corrected, bit values of individual data packets are restored based on the use of the above Walsh functions, and the generated data packets are transmitted to the information consumer.

The bits are converted into clock code representations on the transmitting side of the communication line according to clock cycles, the number of which is equal to the number of information sources, and in each clock cycle, the value of the code representation is determined by the formula {p /, v = В _а Λι _v + Bnhiv + · + B _iN h _Nv }. (1)

Here p _iv are the values of the clock code representations;

i is the number of the bit frame;

v = 1, 2, ..., N - numbers of consecutive clock cycles of transmission of individual bits from the composition of the bit frame;

N is the number of information sources;

B _iv - bit of the source of digital information;

h _kv = ± 1 are the values of each of the kth order Walsh functions in transmission clocks.

If there is no next bit in the source data packet, the place of the specified bit in the clock code representations is replaced with zero.

On the receiving side of the communication line, using the obtained set of clock code representations, bit values are restored by recursively solving a system of linear algebraic equations

- 2 033569 {- ^ zl ^ lv - ^ i'2 ^ 2v 4 ... + Bijqhyj _v p, · _v }. (2)

The specified solution is found due to pairwise algebraic addition of each of the above equations with number i and equations with number i + N / 2, where i = 1, 2, ..., N / 2, and also due to operations of pairwise algebraic subtraction of the same equations.

Implementation of the proposed method for transmitting digital information is as follows.

Information technology code division of the channel resource between different subscribers - data sources suggests that the number of subscribers is limited to a certain value of n. The transmission of bits over the physical channel is carried out using successively constructed information blocks in the form of bit frames. Bit frames provide the formation in a single physical channel of a number of N logical channels through which the source subscribers transmit their information packets, with N> n. Each of the logical channels simultaneously receives one bit from the packet of the corresponding subscriber - the source with the original speed

V = Q ^ / kt, where Q _6-k is the number of bits in a bit frame;

At - the duration of the formation of the bit frame.

The flow of information in the form of bit streams from its sources is as follows: source bit stream 1 bi [t + / AH], bi [t + (/ - 1) AH], b \ [t + (/ - 2) Ar] , ..., bi [t + Ar], t> i [/], b \ [t - Ar];

source bit stream 2 b ₂ [t + / Ar], b ₂ [t + (/ - 1) Ar], b ₂ [t + (/ - 2) Ar], ..., b ₂ [t + Ar] , Z> ₂ [r], b ₂ [t - Ar];

source bit stream nb _n [t + / AG], b _n [t + (i - l) Ar], b _n [t + (i - 2) Ar], ..., b _n [t + Ar], b _n \ t \, b ₂ \ t - Ar]; source bit stream n + 1

0 _and + 1 [r + / Ar], 0 + 1 [T + (/ - 1) Ar], 0 _and + 1 [r + (/ - 2) Ar], ...

• · · 0 „ ₊ ! [Г + АГ], 0 _{st +} i [r], 0 _{and +} 1 [Г - АГ];

channel bit stream N

0 # [Г + / АГ], 0 # [Г + (/ - 1) АГ], ОДГ + (/ - 2) АГ], ..., ОД / + АГ], ОДГ], 0 # [Г - AH].

Here bj [t + iAt] is the value of the source bit j at time t + iAt;

i = 1, 2, ..., N - current time reference number corresponding to the number of the bit frame;

j = 1, 2, ..., N is the number of the source.

The value of At is determined depending on the required bit rate V in the information flows of the sources. The source of the information stream with a specific number j can be assigned to a certain logical channel both statically and dynamically. At the same time, the logical channel number on the receiving side should uniquely identify the source number of the information flow.

The value N defines the potential number of logical channels, that is, the number of sources that can simultaneously use the resource of one physical channel. In the unused logical channels of the system, the zero codes are permanently transmitted until the code combination of the start packet bits appears. For a given value n of the number of real sources, the number of logical channels of the system N is selected based on the following condition:

# = 2 “; and> Ant [log ₂ “], (3) where Ant [log ₂ n] is the Antje function of the upper integer boundary of the value of log ₂ n.

In the case of N> n, redundant channels are considered as a channel reserve, which can also be specially introduced in order to provide prospects for the development of the system. Each bit-frame consists of an information and control part.

The initial group of operations to implement the developed method for transmitting digital information is that on the transmitting side of the communication line, a Hadamard matrix is formed, the order of which corresponds to the number of information sources, and each row of the matrix is assigned to a specific information source.

The elements of the corresponding rows of the Hadamard matrix display the values of the Walsh functions. The converse is also true: the values of the elements of the rows of the Hadamard matrix correspond to sequences of values of certain Walsh functions.

Walsh functions are discrete functions having the following meanings:

- 3 033569 (4) ^ _Y = VD = ± 1 for k = 1, 2, ..., N at the points v = 1, 2, ..., N.

For each Walsh function, these values are determined by the elements of the corresponding kth rows of the Hadamard matrix:

A feature of the Hadamard matrix is that N in this case is a power of 2 (2, 4,

8, ...), and the elements of the rows are formed according to the recurrence rule:

The next operation of constructing a bit frame on the transmitting side of the communication line is the formation of bits of the information part of the bit frame, which is performed as follows.

The total number of physical bits in the information part of each bit frame is constant and equal. Its value is determined by the relation ^ = N (1 + log ₂ N), which takes into account the number of bits necessary to obtain the resulting parameters of the clock code representations in the information part of the bit frame.

For any time t + iAt, the first significant bits of the streams of all n functioning sources and the zero bits of N-n backup logical channels are grouped by a multiplexer into a single bit-frame. From individual bits {bj [t + iAt]} of information flows, a bit-vector string is preliminarily formed:

D [ί + iAt] = {By [t + ζΔί]}. (8)

For technological purposes, line elements are represented by bipolar unit values:

By [t + ζΔ /] = + 1 if bj [t + ζΔ /] = 1; (9)

By [t + ζΔ /] = -1 if bj [t + ζ'Δ /] = 0. (10)

When generating information bits, a time-clocked set of kth order Walsh functions {W _kv } is used as the signature of the bit frame. In each step v, the transmission is not of the values of the elements of the bit vector string B _i [t + iAt] = {B _i j [t + iAt]}, but of the resulting parameters of the clock code representations {p _v } with which the information part i- go bit frame:

{ _Ρί ·, _ν } = Β, · [ί + ζΔί] χ # _No. (11) with p _{(> v} = Bi \ hx _v + B, 2 ^ 2v + .. · + Biuhffr. (12)

The range of values of the resulting parameters of the clock code representations {p _iv } is ± N. Therefore, when filling out the information part of the ith bit frame for placing each parameter, taking into account the sign, (1 + log ₂ N) bits should be physically used, which determines the total number of N bits in the information part of the packet.

For a more compact presentation, the elements of the bit vector string B ^ t + iAt] are further represented by the corresponding By values.

The next operation of constructing a bit frame on the transmitting side of the communication line is the formation of bits of its control part.

The control part of the bit frame contains R control bits. The value of R is determined by the multiplicity of errors q, which are supposed to be detected and corrected in each bit frame:

R> qM, r ¹ -! <^ <2, (13) where 2 ^m is the upper integer boundary of the length of the information part.

In this case, the probability of residual bit errors in the bit-frame Р _{Ош б-к} after adjustments will be:

- 4 033569

where R _osh to - the probability of a bit error when transmitting information on a symmetric channel;

c ^{q N} * ^{+ R} is the number of combinations of Xf + R bits per q bit.

An acceptable embodiment of this operation is, for example, the use of a corrective cyclic code based on a generating polynomial G _r (x) of degree g. If the polynomial P _X f-1 ( ^x ) is a bit polynomial representing the code combination of the information part of a bitframe of degree Xf-1, then the representing polynomial C _r-1 (x) bits of the code combination of the control part of the bit-frame of degree g-1 is determined by the remainder of the division of the representing polynomial of the information part of the bit-frame shifted by g positions to the left by the generating polynomial:

xT „ _t _, (x) / G, (x) = C,., (x). (fifteen)

The coefficients of the polynomial C _r-1 (x) correspond to the bits of the code combination of the control part of the bit-frame. The coefficients of the depicting polynomial I ^ _{+ r-1} (x) of the cyclic code corresponding to the bits of the information and control parts are the result of the following element-wise operations:

Probability R _{osh b} . _for packets associated with the volume Q _n transmission data subscribers - sources and a confidence level of P _Ven information contained therein in the following ratio:

(17)

The final operations of constructing a bit frame on the transmitting side of the communication line are the formation of code combinations of flags and / or signals for the start of a bit frame and a freeze frame, as well as the direct assembly of all components with their packaging in a single information unit. Code combinations are built in accordance with the requirements of the protocol of informational compatibility of the transmitting and receiving sides of the communication line and are tightly logically mounted in the block together with the information and control parts of the bit-frame using concatenation operations. Prepared by the proposed method, the information block is ready for transmission over the communication line.

The initial operation of unpacking a bit-frame on the receiving side of the communication line is the allocation in atomic form of its components. To this end, a codeword bit frame logically multiplied by the four constant-extractor their parts form 111 ... 11 with appropriate lengths Q ^ pr _b-a, Hf, R and Q ^ _{b. to} . Components using logical shift operations to the right are pressed to the right borders of the discharge grid of the equipment. The completeness and correctness of the receipt of the bit frame is formally checked.

The subsequent operation of unpacking the bit frame is to verify the mutual correspondence of the information and control parts of the bit frame and perform corrective actions if it is established that there are errors. For this purpose, the image polynomial J ^ + r-1 (x) of degree Хф + r-1 is synthesized based on the values of the received bits of the combination of information and control parts of the bit-frame. Find the syndrome S _r-1 (x) in the form of the remainder of the division of the synthesized depicting polynomial J ^ _{+ r-1} (x) by the generating polynomial G _r (x):

(18)

With identically zero syndrome S _r-1 (x) = 0, the hypothesis of the absence of errors and ^ + r-1 ^(x) = - ^ + r-1 ^{(x) is} accepted.

If S _r-1 (x) syndrome is non-zero, then the hypothesis of the need for adjustments is accepted. One of the known methods for processing cyclic codes at the reception point is the one depicting the polynomial E ^ _{+ r-1} (x) to eliminate errors, and the following corrective actions are carried out:

^ ₊ ri (*) = ^ ₊ r-iW + ^ ₊ , -iW} mod2. (19)

Then, on the receiving side of the communication line, bits of information flows are restored based on the information part of the bit frame.

Using the vector of parameters of the clock code transformations {ρ ^ _ν } obtained in the i-th bit frame and corrected by the cyclic code, which corresponds to the information part of the bit frame, and the well-known set of Walsh functions JW _kv j, the value of the elements of the bit vector row Bi is restored [t + iAt] = {Bj [t + iAt]}. To restore this value, it is proposed to use the separation method, which involves solving a system of linear algebraic equations with respect to By values.

- 5 033569 {B, \ h \ _v + Bnfax + ... + B ^ h} f _V = p ,; _v }. (20)

The direct solution of these equations when implementing the proposed method is not possible to use for the following reason. With an increase in the number of information sources from which data are transmitted to the communication line to N = 32, 64, 128 and higher, the time required to solve the equations turns out to be extremely large, while they should not exceed the boundary value At.

In this regard, the following acceleration scheme is proposed for a method for directly recovering the values of elements of a bit-vector string. For this, the formation features and properties of the above Hadamard matrices H _{N are used} .

One of the properties of Hadamard matrices is that when transposed, they do not change. In addition, the application under consideration takes into account the fact that the set of Walsh functions is defined by the products of the Rademacher functions:

r _k , _v = Sign {sin [H2 * ⁺ ']}, (21) ^ 0, v 1 »^ 1, v И), vj ^ 2, v Π, νί ^ 3, v 7 * i _{> v} ro _{) V} j ^ 4, vi ”2, vj ^ 5, v ^, γΤ'ο, νί ^ 6, v = r ₂ , vH, v; ^ 7, v = Ζο, ν. (22)

Moreover, the values of the elements of certain rows of Hadamard matrices in terms of the number of alternating signs correspond to the values of the Walsh functions.

To implement the proposed method for transmitting digital information, a set of multilevel downward equivalent transformations of a system of linear algebraic equations is performed. The calculation scheme can be represented by a dichotomous graph with u hierarchical levels of vertices. With a pre-fixed value of N = 2 ^g at each level g = 1,2, ..., u, it is necessary to ensure that 2 ^{g of the} same steps are performed with the computational operations of addition or subtraction. Wherein *

for level g and step s, the step parameters p ( _gs yij in the addition operations and the parameters ρ in the subtraction operations can be recursively determined for each bit frame upon its arrival on the receiving side of the communication line. This actually makes the implementation of the downward process of direct matrix transformation equations and drastically reduces the overall computational complexity of the problem.

The downward equivalent first-level transformation is realized due to the operations of the first step of algebraic addition of equations with numbers 1, 2, ..., N / 2 and equations with numbers N / 2 + 1, N / 2 + 2, ..., N. С taking into account the technology for constructing Hadamard matrices and Walsh functions, the equivalent system of equations has the following form:

2 [Д1Л1 (Л72) + Bi2 ^ 2 (N / 2) ⁺ · · + D (А / 2) ^ (М2) (^ 2)] = Р (1, 1), i, (N / 2) , (23) where for the first level of transformations (u = 1) and the first step (s = 1) of this level

P (1,1), i, 1 ~ Pi, 1 ⁺ Pi, (N / 2 + l) i P (l, 1), i, 2 ⁼ Pi, 2 ⁺ Pi, (N / 2 + 2), ···; Ρ (ϊ, 1), ί, (ΝΙ2) -Pi, (NI2) + Ρί, Ν- (24)

In step addition operations, the right half of the set of unknowns B _i ( _{N / 2 + 1} ), B _{i (N / 2 + 2} ), ... B ^ is suppressed and the unknown coefficients of the left half of the set Bi1, Bi2, .. are doubled. ., Bi (N / 2).

At the second step, subtraction operations are performed from equations with numbers 1, 2, ..., N / 2 of equations with numbers N / 2 + 1, N / 2 + 2, N. The left half of the set of unknown unknowns is suppressed B _ib B _i2 , ..., Bi ( _{N / 2} ) with doubling the unknown coefficients of the right half of the set Bi (N / 2 + 1), Bi _(N / _{2 + 2} ), ..., Bflsf. Wherein

2 [Д (У / 2 + 1) ^ 1 (Я / 2 + 1) + Д (М2 + 1) ^ 2 (М2 + 2) ⁺ · · ⁺ B _iN h _NN ] ~ P (i, ₂ ) i, N 'where (25)

Ρ (ι.2> /. (ΛΤ / 2 + 1) ~ Pi, 1 ~ pi, (Y / 2 + 1); Ρ (1,2) |, (Y / 2 + 2) ~ Pi, 2 - pi, (N12 + 2) ', · · · (26) ··· »P (l, 2) i, (N = JV / 2 + JV / 2) Pi, (N / 2) pi, N'

One level of equivalent transformations leads to the division of each system of algebraic equations of this level into two systems and at the same time reduces the order of each newly obtained system by half, but the matrices of equivalent systems without multiplication by a factor of 2 remain Hadamard matrices and their elements do not require any recalculations . This allows using

- 6 033569 operations of addition and subtraction to perform the steps of the second and subsequent levels of downward equivalent transformations. For example, successive equivalent transformations of the first step of the second level using the addition operations form the following computational process:

4 [5 _ζ · ι / ίΐ (Λ74) + Bi2h ₂ (No. 4) + ... + B ^ / 4/1 ^ / 4 ^ / 4)] = p ( ₂ , 1), i, (Л74) , (27) where for the first step of the second level of transformations we get:

P (2,1), 1,1 ⁼ P (1,1), i, 1 + P (1,1), i, (N14 + 1) J P (2,1), i, 2 ⁼ P ( 1,1), i, 2 + P (1,1), i, (N / 4 + 2) ··· 5 '> P (2, 1), i, (L / 74) ^_ P (1, 1), /, (М4) + Р (1,1), i, (N12)

After the first steps of downward equivalent transformations at (u-1) levels, the resulting system takes the following form:

where p (u - 1,1), i, 1 = P ("-2,1), i, 1 + P (" - 2,1), i, b P ("- 1,1), i, 2 ~ P (and -2.1), i, 2 + P (and - 2.1), i, 4

Addition and subtraction of these equations leads to the following results:

[2 ⁽ >] (А _М + / ϊ 12) ^ / 1 ^- Р ("-1,1), i, 1 4 · р (и-1,1), 1,2 ^- Р (и, 1) , i, b (31) [2 ⁽ “ ¹⁾ ] (/ ϊ21 - h ₂ 2) Bt2 = P (u-1,1), i, 1 - P (u-1,1), i, 2 = P („, 1) ,, 2 · (32)

Thus, a direct calculation of the values of two unknowns is provided:

B, \ = [p ( _II- v, ΐ), i, 1 + p ( _m -1,1), i, 2] / [(2 ⁽ “> 1 + Lie)] ⁼ p (α, 1) , i, 1/2 ^th ; (33)

Dt ⁼ [p (α - 1,1), i, 1 - P (u-1,1), i, 2] / [(2 ⁽ “^ (/ ζςι - / 122)] - P ( _u , 1 ) 1,1 / 2 ^m ; (34)

Parallel equivalent transformations using subtraction operations form the process of calculating values for another pair of unknowns:

[2 ⁽ ““ ¹⁾ ] [D (A72 + 1) ^ 1 (Y / 2 + 1) + Bi (N / 2 + 2 ^ l (N / 2 + 2)] = Ρ ( _Β _ι, [„- ι] ψ, (Λ72 + 1) ^;

2 ^(U ~ ^l} ] [Bi ( _{N /} 2 + i) h2 (N / 2 + l) + VTsm2 + 2) h2 (NI2 + 2)] = P („_ [ _and _ ϊ] *} ^ ( „/ 2 ₊ 2) ^; where

P (-l, [al] ² ) i, (N / 2 + l) P (u-2, [u - 2] ² ) ί, (Ν / 2 +1) ^_ % - 2, [k - 2 ] ² ) i, (JV / 2 + 3) ' ^P (-1. ["- 1G) <, (N / 2 + 2) P ( _U -2, [u-2] ² ) i, (N / 2 + 2) ^P ("-2, [u-2] ² ) i, (N / 2 + 4) '

Subtraction and addition of these equations leads to the following results:

Bi (N / 2 + 2) = P ( _{a> and} φ, (* / 2 + 1) ¹ [( ^{2 (I_1)} ( ^A 1 (No. 2 + 2) - ^ 2 (Y / 2 + 2)] = Ρ („, _Β ψ, („ _{/ 2 + 2} ) ^{Ζ 2} “ί

Bi (N / 2 + l) = Ρ ( _Μ , „ψ, („ _{/ 2 + 1} ) / [( ^{2 <М} ~ ^1} (Л1 (М2 + 1) + h ₂ (N / 2 + 1)] = Ρ ^ ψ, ^ + ι) ^{Ζ 2} “>

Where

P (,, u ² ) l, (N / 2 + 2) “Ρ (- - 1, [- - 1] ² ) /, (Ν / 2 + 1) - P (- - 1, ^ - 1 ^ ), -, ^ / 2 +2) '• - * I ♦

Ρ (, “ ² ) ί, (Ν / 2 + 1) ~ P (ul, [ul] ² j, i, (N / 2 + l) Ρ (“ - 1, [u-1] ² ) ί, (Ν / 2 + 2) '

In a similar way, final expressions are formed for calculating the values of the remaining unknowns based on the predefined B _i1 , B _i2 , Bi ( _N / _{2 + 1} ), Bi _(N / _{2 + 2} ). The values of clock code representations restored as indicated on the receiving side of the communication line as information packets are sent to information consumers.

The advantages of the proposed method

The proposed method for reliable transmission and allocation of bits of information flows with code- (35) (36) (37) (38) (39) (40) (41) (42)

- 7 033569 in the division of logical channel resources is based on the implementation of CDMA principles exclusively by software. This greatly simplifies the use of the developed method for transmitting digital information, since it does not require the search and application of complex and expensive hardware solutions. The construction in accordance with the proposed method of software intelligent multiplexers on the transmitting side and separators on the receiving side of the communication line showed high efficiency and stability of the functioning processes of the digital information transmission system.

Using the developed method in a digital data transmission system provides a high degree of reliability of information delivery to consumers. This is achieved due to individual protection by a block error-correcting code of each bit frame. When interference and other negative factors are affected by the communication line, one or several bit frames are affected, errors in each of which are detected by the error-correcting code and restores the broken bit frame. Studies have shown that on channels of relatively low quality (primarily on radio channels) with error rates of 1 x 10 ^-4 - 2x10 ^-3 for N = 16, the probability of residual bit errors in each bit frame does not exceed R _{os bq} = 5 , 6x10 ^-5 .

Even with a relatively large number of logical channels in the system (for example, N = 256), reliable separation of information bits and their separation by sources is possible at acceptable levels of delays in performing operations. A significant additional time gain arises due to the fact that on the transmitting and receiving sides of the communication line, all operations in real bit-frame processing processors are carried out only with the help of short integer arithmetic instructions.

An example implementation of the proposed method

To implement the proposed method for transmitting digital information in conditions as close as possible to the real situation, we used the MATLAB 6.0 system of calculations and modeling with the extension Simulink 4.0. In the model of the transmitting side of the communication line, the software reproduced the operations of constructing bit frames for N = 8 and N = 16, as well as the operations of their noise-resistant coding using cyclic codes and corresponding generating polynomials

Cis = X ¹⁵ + X ^th + X ¹⁰ + X ⁹ + X ⁸ + Χ ^Ί + X ⁵ + X ³ + X ¹ + X + 1 (107657), C ₂₁ = Ύ ²¹ + Ύ ¹⁸ + JT ¹⁷ + JT ¹⁵ + JT ¹⁴ + JT ¹² + Ύ ¹¹ + X * + X ¹ + X ⁶ + X ⁵ + + Ύ + 1 (11554743), which provide detection and correction of three-time errors in each bit frame (q = 3). In the model of the communication line, distortions of individual bits of the transmitted frame were reproduced, while an assumption was made about the symmetric distortions of zero and single bits with a probability of _Rshk .

Table 1. Reliability of information transfer.

Q Μ R Rom to Ρ OSH 6-κ β; (bit) at information confidence level 0.98 1 2 3 4 5 6 7 β = 3 8.32 5 fifteen 5 · 10 ' ⁴ 1,110 ^- ' 1 836 600 1 · 10 ' ³ 1.7 · 10 ^-7 118,840 2 · 10 ³ 2.7 · 10 “ ⁶ 748 16.80 7 21 5 · 10 ' ⁴ 2.5 · 10 ~ ⁷ 80 811 1 · 10 ~ ³ 3.8 · 10 ^-6 5317 2 · 10 ' ³ 5,610 ~ ⁵ 360

The results of modeling and calculations, comparing the probabilities R _{w bk of} residual bit errors in a bit frame after adjustments by an error-correcting code and rational volumes of bi packets ”at which the reliability of their transmission is achieved not lower than 0.98, are given in Table. 1. From the analysis of the data table. 1 it follows that the proposed method for all logical channels provides a sufficiently high level of reliability of information in the transmitted packets. At the same time, packages can have significant volumes even when using channels with relatively low quality and

- 8 033569 high values of probabilities of occurrence of bit errors (for example, for wire communication channels and radio channels).

Analysis of the simulation results showed that the computational process when transmitting digital information according to the developed method is organized using sequential operations and is quite homogeneous. Below is a detailed dichotomous computational process for processing bit frames on the transmitting and receiving sides of the communication line for N = 8, u = 3. The resulting parameters of the clock code representations are determined in accordance with expressions (23) - (42).

1. Hadamard matrix on the transmitting side of the communication line:

5c-1 h \ 2 -1

J21 - 1 A22 ⁼ - 1

L31 ⁼ 1 L32 ⁼ 1

L41 = 1 L42 ^{= -} 1

5si ⁼ 1 ^ 52 ⁼ 1

5b1 = 1 ^ 62 ~ - 1 hq \ - 1 L72 ⁼ 1

5 ₈ 1 = 1 5 ₈ 2 ⁼ - 1

513 = 1/214 = 1

J23 == 1 / ^ 24 ⁼ - 1

A33 = - 1 L34 = - 1

L43 = - 1 L44 = 1

L53 = 1/254 = 1

Lbz = 1 Lb4 = - 1/273 = - 1/274 = - 1

5 ₈₃ ⁼ - 1/284 ⁼ 1/215 ⁼ 1/216 ⁼ 1/225 = 1/226 ⁼ “1/235 = 1/236 - 1/245 = 1/246 ^{= -} 1/255 = - 1 L ₅₆ = -1 / 265 ^{= -} 1/266 ⁼ 1/275 = - 1/176 = –1 / 285 - ~ 1 / ^ 86 ⁼ 1/217 ⁼ 1/218 = 1

L27 = 1/228 ⁼ -1

= L37 - L38 1 = -1 / 247 = - 1 / d ₄₈ = 1/257 ⁼ - 1/258 ⁼ - 1/267 ^{= - 1/268 = 1/277 =} 1 / * 78 = 1/287 = 1 / * 88 ⁼ - 1

2. The composition of the bit frame on the transmitting side of the communication line:

5.1, 5.2, 5, ₃ , 5.4, Bfi, B ft,

Vts, V%,

3. Equalities for calculating code representations on the transmitting side of a communication line:

pi, 1 = B _a h _n + B _i2 h _2X + 5, ₃ 5 ₃₁ + 5, ₄ 5 ₄ i + B _i5 h _5i + B _i6 h _6i + Bnh _7l + 5, ₈ 5 ₈ ι pi, 2 = Bnhn + Bahii + B _i3 h ₃₂ + B _iit h ^ ₂ + B _i5 h ₅₂ + B _i6 h ₆₂ + B _i7 h ₇₂ + 5, ₈ 5 ₈₂ pi, 3 = B _n h _l3 + Д ₂ / * 2з + 5, s5 ₃₃ + B uh + B ^ hy, + Bfth ^ + B _i7 h ₇₃ + 5, ₈ 5 ₈₃ p. 4 = Bnhu + B ^ h-iu + 5.3 / 234 + B ^ h ^ + Bi ₃ h ^ + B _i6 h _M + B _i7 h _7i , + B ^ h _M

Pi, 5 ~ Bi \ h \ s + 5, - ₂ 5 ₂₅ + B _i3 h ₃₃ + 5, -4 / 245 + 5, -5 / 255 + B _i6 h _b5 + 5.7 / 275 + 5, ₈ 5 ₈₅

Pi, 6 = 5.1516 + 5, - ₂ 5 ₂₆ + 5, ₃ 5 ₃₆ + 5, ₄ 5 ₄₆ + 5, - ₅ 5 ₅₆ + B _i6 h _6f) + B _i7 h ₇₆ + B ^ h ^

Pt, 7 - Bt \ h \ ₇ + B _t2 h ₂₇ + B _i3 h ₃₇ + Β ^ ₇ + B ^ h ^ + B ^ h ^ + B _i7 h ₇₇ + B ^ h ^

Pt, 5 = 8, iAi ₈ + B ₂₈ + 5 _minutes 5 z5z8 + 5 · _{48 May 4} + 5, ₅ + 5 ₅₈ 5 ₆₈ -b5 + B _n h _7% + B, ₈ 5 _88.

4. The composition of the information part of the bit frame on the transmitting side of the communication line:

Pi, 1 Pi, 2 r /, 3 Pi, 4 Pi, 5 Pi, 6 Pi, 7 Ρί, 8

5. Noise-tolerant coding of the bit frame by a cyclic code with a generating polynomial on the transmitting side of the communication line.

6. Noise-resistant decoding of a bit frame by a cyclic code with a generating polynomial

G15 =% ¹⁵ + X ¹¹ + X ¹⁰ + X ⁹ + X ^s + X ¹ + X ⁵ + X ³ + X ² + X + 1 (107657) on the receiving side of the communication line.

7. Recursive selection of bits of information sources on the receiving side of the communication line according to the following equations:

5.15c + Β, ₂ 5 ₂ ι + 5, ₃ 5 ₃ 1 + B, ₄ 5 ₄ i + B _i5 h _5l + 5, · ₆ 5 ₆ ι + B _n h _lx + Β, ₈ 5 ₈ ι = p _(j 1

5.1512 + B _i2 h ₂₂ + B, ₃ 5c2 + B _i4 h ₄₂ + B _i5 h ₅₂ + B _i6 h ₆₂ + 5, ₇ 5 ₇₂ + B, ₈ 5 ₈ 2 = Pi, 2

- 9 033569

D1 Dz + B _a h ₂₃ + B _i3 h ₃₃ + B _iA h ^ + B _i5 h ₅ 3 + B _i6 h ₆₃ + B _n h ₇₃ + B ^ h ^ = p ₍ , ₃

D1 # 14 + B _i2 h ₂ 4 + B _i3 h ₃ 4 + Βί4 ^ 44 + B _i5 h ₅₄ + B _i6 h _M + B _n h-) 4 + Β ₍₈ Λ ₈₄ = ρ _(> 4

D1D5 + Bnfas + D3D5 + D4A45 + B _i5 h ₅₅ + Biffas + D7Y75 + B _No. h _{& 5} = p _{it 5}

WCC \ b + B _i2 h ₂₆ + D3A36 + DchDb + WCC ₆ + B _lb h _{6 (i} + B _n h _{7 (} , + D ₈ L ₈₆ = p ,, ₆

D1D7 + B _t2 h ₂₇ + D3L37 + D4D7 + B _i5 h ₅₇ + B _i6 h ₆₇ + B _l7 h ₇₇ + B _iS h _S7 = p _z , ₇ D1D 8 + D1D ₈ ⁺ Dz # 38 + Y4L48 ⁺ D'5 ^ 58 + Bibh ^ + B _n h ₇ % + Β _ι8 Λ ₈₈ = p _z , ₈ .

The first level of recurrent transformations (g = 1)

The operations of adding the first step (s = 1) of the first level:

2 [Д1Лц + B _i2 h ₂ \ + Д3Д1 + Д4Д1] = pi, 1 + Pi, 5 = р (1, i), t, 1 2 [Д1Д ₂ + В _а 14 ₂₂ + Д3Л32 + Д4Д2] = р / , 2 + Pi, 6 = P (i, i), t, 2 2 [Д1Дз + B _i2 h ₂₃ + B _i3 h33 + Д ₄ Дз] = p _z , 3 + Ρ ”, 7 = P (i, i ), /, 3 2 [.В / 1 /? 14 + 5/2 ^ 24 В 13 ^ 34 + £ / 4 / ^ 44] p _Z , 4 + p _Z , ₈ P (l, η, _z , 4 .

Subtraction operations of the second step (s = 2) of the first level:

2 [2? _z5 / i ₅₁ + B _i6 hf, \ + B _i7 h ₇ \ + B _/ 8/2 ₈ i] = p _z , ι - p _z , 5 = pj, 2) /, 5 2 [B _i5 / i52 + In _i6 h ₆₂ + B _i7 h ₇₂ + B _z8 / z ₈₂ ] = pz, 2 ~ Pi, 6 = P (i, 2), 6 2 [B _l5 h _S3 + B _i6 h ₆ 3 + B _i7 h ₇₃ + D ₈ / g ₈ s] = p _z , 3 - p _z , ₇ = p ( _{t 2} }, ₇ 2 [5i5 ^ 54 ⁺ DB # 64 + Bi7h74 + D8 / g84] = Pi, 4 ~ Pi, 8 ⁼ P (l, 2) /, 8

The second level of recurrent transformations (g = 2)

The operations of adding the first step (s = 1) of the second level:

4 [D1D 1 + D? D1] = p (1, I), /, I + p (1,1), i, s ⁼ p (2,1), i, 1

4 [D1A12 + D2D2] - p (1, 1), i, 2 + p (1, 1), i, 4 = P (2,1), /, 2 Subtraction operations of the second step (s = 2) of the second level:

4 [B; ₃ 3131 + Β _ζ4 / 2 ₄ 1] = p (1, 1), i, 1 - p (1,1), i, s = P (2,2) /, 3

4 [B _i3 / l32 + 5/4 ^ 42] = p (1,1), i, 2 ~ P (1,1), i, 4 = P (2,2) /, 4 ·

The operations of adding the third step (s = 3) of the second level:

4 [D ₅ / g ₅ 1 + B / 6 ^ 61] = P (1,2) /, 5 + P (1,2) /, 7

- P (2,3), i, 5 ”P (2, 3), i, 6 ·

P (2,4) /, 84 [Β, ₅ λ ₅ 2 + Bi ^ lei] - P (l, 2) /, 6 ⁺ P (1,2) /, 8

Subtraction operations of the fourth step (s = 4) of the second level:

4 [B / 7 // 71 + B _; - ₈ L ₈ 1] = P (1, ₂ ) /, 5 ^- P (1,2) /, 7

4 [B _i7 h ₇₂ + B% h $ 2 \ = Ρ (v, 2) /, 6 “P (1,2) /, 8

The third level of recurrent transformations (g = u = 3)

The operation of adding the first step (s = 1) of the third level:

D1 ⁼ [p (2,1), i, 1 ⁺ p (2,1), v, r] / 8 ~ p (3,1), i, 1 Subtraction operation of the second step (s = 1) of the second level:

D-2 ⁼ [p (2,1), /, 1 “P (2,1), i, r] / 8 = p ( _{3 2} \ i, 2 The operation of adding the third step (s = 3) of the third level:

Dz ~ [P (2,2) i, 3 ⁺ P (2,2) /, 4] ^{/ 8 =} P (3, 3), 1.3 The operation of subtracting the fourth step (s = 4) of the third level:

- 10 033569 ^ / 4 [P (2,2) /, 3 “P (2,2) /, 4 TO p (3,4) _if 4 ·

The operation of adding the fifth step (s = 5) of the third level:

D-5 ⁼ [p (2, s) /, 5 ⁺ P (2,3) i, 6 ⁼ P (3, 5), /, 5 The operation of subtracting the sixth step (s = 6) of the second level:

Bi6 = [P (2,3) /, 5 “P (2,3) /, 6 Y8 = P (3,6) /, 5 ·

The operation of adding the seventh step (s = 7) of the third level:

Bn ⁼ [P (2,4ji, 7 ⁺ P (2. 8 VS = P (3, 7), (, 7) The operation of subtracting the eighth step (s = 8) of the third level:

De ⁼ [p ( ₂ , 4) /, 7 “P (2,4) /, 8 ^ ⁼ P (3,8) /, 7 ·

The values of bits B _ib B _i2 , B _i3 , B _i4 , B _i5 , B _i6 , B _i7 , B _i8 recovered by the indicated method on the receiving side of the communication line send information to consumers.

Similarly, these parameters are calculated for almost any number of information sources involved in the formation of data on the transmitting side of the communication line. Sources of information used in compiling the description

1. Feer, K. Wireless Digital Communications. Methods of modulation and spreading / Per. from English M .: Radio and communications, 2000 .-- 520 p.

2. Chugaev V.I. The method of correlation processing of broadband signals. RF patent 2185658, cl. G06G 7/19, priority 30.08.2000, publ. 07/20/2002.

3. Savinkov A. Yu. Algorithm for quasi-optimal two-antenna reception of a broadband signal in the conditions of spatially concentrated interference // Theory and technique of radio communication, 1998, issue 2 (prototype).

Claims (4)

CLAIM 1. The method of transmitting digital information, which consists in the fact that on the transmitting side of the communication line receive the original data packets from digital information sources, impose a noise-resistant code on these data packets, frame them with flags and / or start-stop signals and send data to the communication line , and on the receiving side of the communication line, the indicated frame is removed, data decoding is performed, and data packets are sent to the information consumer, characterized in that the Hadamard matrix is formed on the transmitting side, the order of which with corresponds to the number of information sources, and the row elements display the values of the orthogonal Walsh functions, and assign each row of the matrix to a specific information source, then the next bits with the same numbers are selected from each source data packet, they are converted into clock code representations using the above Walsh functions obtained by the values clock code representations fill the information part of the bit-frame, impose a block noise-resistant code on it by forming a control part bit frame, frame both parts of the bit frame with start-stop flags and / or signals and send data to the communication line, and on the receiving side after removing this frame check the correctness of the information and control part of the bit frame, if necessary, make corrections, restore the bit values of individual data packets based on the use of the above Walsh functions and transmit the generated data packets to the information consumer. 2. The method of transmitting digital information according to claim 1, characterized in that on the transmitting side of the communication line, the conversion of bits into clock code representations is performed by clock cycles, the number of which is equal to the number of information sources, and in each clock cycle, the value of the code representation is determined by the formula where p ^ _v are the values of the clock code representations; i is the number of the bit frame; v = 1, 2 ... N is the number of information sources; Biv - bit of the source of digital information; h _kv = ± 1 are the values of each of the kth order Walsh functions in transmission clocks. 3. The method of transmitting digital information according to claim 1, characterized in that on the transmitting side of the communication line in the absence of the next bit in the original data packet, the place of the specified bit in the clock code representations is replaced by zero. 4. The method of transmitting digital information according to claim 1, characterized in that on the receiving side of the communication line from the received set of clock code representations, bit values are restored by recursively solving a system of linear algebraic equations - 11 033569 {Β, ιΛίν + Β _Ω Λ ₂ ν + ... + B _iN h _Nv = pi, ν}, and their solution is found due to pairwise algebraic addition of each of the above equations with number i and equations with number i + N / 2, where i = 1, 2, ..., N / 2, and also due to the operations of pairwise algebraic subtraction of the same equations. Eurasian Patent Organization, EAPO Russia, 109012, Moscow, Maly Cherkassky per., 2