KR20050088182A - Addresses generation for interleavers in turbo encoders and decoders - Google Patents

Abstract

An arrangement for generating addresses for interleaving/de-interleaving sequences (X1, X2, X3,..., XK) including a given number (K) of items, wherein each value for said given number (K) identifies a corresponding set of interleaving parameters (R, C, p, v). The arrangement comprises at least one memory unit (22) wherein a plurality of records are stored, each record being indicative of a respective set of parameters (R, C, p, v) corresponding to at least one value for said given number (K). Sets of interleaving parameters (R, C, p, v) are thus available in the memory unit (22) to be promptly and directly retrieved for all possible values of said given number of items (K). A preferred use is in turbo encoders /decoders for advanced telecommunications applications such as UMTS.

Description

Address generator for interleaver in turbo encoder and decoder {Addresses Generation For Interleavers In Turbo Encoders And Decoders}

TECHNICAL FIELD The present invention relates to address generation techniques and has been developed with particular attention to possible applications in interleaving / deinterleaving modules.

A preferred application of the present invention is in a third generation mobile communication system, more specifically in a turbo decoder device as described in the 1999 3GPP TS 25.212 V.3.5.0. (2000-12). It is a mother interleaver.

Turbo encoders grow the transmitted information in addition to possible control codes in such a way that the result of an impairment source that adversely affects transmission can be compensated by complementary decoding at the decoder. A component used in a communication system having a purpose of heating.

An example of a communication system suitable for using the turbo encoding / decoding scheme is a system adopting a so-called Universal Mobile Telecommunications System (UMTS) standard.

The key element in a turbo encoder / decoder (eg, as defined in the 3GPP TS 25.212 protocol) is the so-called interleaver / deinterleaver block. The operation of such a block is provided in order of the elementary items of information (bits) contained in a given length K block that are essentially variable in accordance with certain laws.

In EP-A-1 195 910, an interleaver for use in a turbo encoder for a UMTS device is disclosed. The interleaver may include: a register for updating and registering a plurality of parameters to set operating conditions of the interleaver; A controller for generating a control signal for controlling the operation of the system by accepting an operating condition from the register; A final interleaver address using an inter-row permutation pattern, an intra-row permutation pattern increasing device value, and an intra-row substitution base sequence provided by the register in accordance with a control signal generated by the controller; And a data storage device for storing data sequentially input to the turbo encoder and outputting data according to the address generated by the address calculator.

Thus, such a device updates and registers a plurality of interleaving parameters (i.e., the number of input data items K, and parameters designated as R, C, p, μ and type D) for setting the operating conditions of the interleaver. Is based on an interleaver containing registers for Thus, such conventional devices are based on the assumption that these parameters are generated and obtained from the outside of the device.

Therefore, this prior reference describes a method for implementing an interleaver in which the interleaving parameters can be dynamically and rapidly applied, taking into account possible changes in the number of input data items, e.g., which may actually change with each packet transmitted. Not teaching.

September 2001 paper by Masera, Mazza, Piccinini, Viglione, Zamboni Proceedings of ESSCIRC 2001, Villach, "Low-Cost IP-blocks for UMTS turbo decoders In the present invention, an apparatus for executing some of the elements having an interleaver in hardware form is disclosed. This prior art also implements an interleaver in which the parameters can be applied dynamically and quickly, and more generally, an apparatus in which the interleaving parameters can be extracted directly and dynamically from the information transmitted (original K value). It does not provide useful guidance.

The invention is illustrated by way of example only with reference to the accompanying drawings, in which:

1 is a general block diagram illustrating an interleaver device;

2 is a block diagram showing parameter generation in a first step in an embodiment of the present invention;

3 illustrates the block diagram of FIG. 2 in more detail; And

4 is an overall block diagram of an address generator according to the present invention.

It is therefore an object of the present invention that the interleaving parameters can be extracted directly and dynamically, whereby these parameters are possible, for example, in the number K of interleaved / deinterleaved input data items. It is to provide an address generator for use in an interleaver / deinterleaver that can be applied dynamically and quickly to change.

According to the invention, this object is achieved in particular by an address generation system having the features claimed in the claims. The present invention also relates to a turbo encoder / decoder comprising the address generator of the present invention as part of an interleaver / deterrier block as well as a corresponding operating method. Finally, the present invention can be loaded directly into the internal memory of a digital computer (such as a processor, for example) and includes a software code portion for executing the method of the invention when executed on the digital computer. It is about a product.

A preferred embodiment of the present invention is an address generator for an interleaving / deinterleaving sequence comprising a predetermined number K of items (e.g., in a turbo encoder / decoder), the apparatus comprising a plurality of records stored therein. At least one memory device having a. Each record includes an information signal indicative of a set of respective interleaving parameters corresponding to at least one value for a predetermined number K of said items.

Preferably, each value for the predetermined number of data items (K) is a matrix (R × C) that arranges the input sequence to produce a substituted matrix for output address generation and performs in-row and inter-row substitution of the matrix. Identify the relevant parameter sets (R, C, p, V) that make up Thus, the set of parameters R, C, p, V can be used immediately for all possible values of the number K of a given data item.

The value of the number K of said predetermined data items being interleaver / deinterleaved is usually contained within a given value range, and said memory device preferably comprises a record number substantially smaller than said given value range. Each record in the memory device is identified by a respective pointer. In the context of a memory device, there is a pointer retrieval circuit configured to generate a corresponding pointer for each record in the memory device for each value of the given number of data items within the given range.

Embodiments of the present invention allow code for the generation of interleaving addresses, both as a software solution with reduced computation (using a commercial DSP or Micro Controller) and preferably as an ASIC solution. This important feature leads to a standalone approach in the design of turbo decoder processors used in UMTS receivers.

As an introduction to the description of an exemplary embodiment of the invention, the basic operating principle of the turbo code internal interleaver is briefly described as provided in the 1999 publication 3GPP TS 25.212 V.3.5.0. (2002-12) specification.想起)

Essentially, the turbo code internal interleaver is a bit-input and purging of the rectangular matrix with padding, intra-raw, inter-raw, and in-row substitution of the rectangular matrix. It consists of bits-output from the rectangular matrix using pruning.

The bits input to the turbo code internal interleaver are x ₁ , x ₂ , x ₃ ,. It is denoted by x _K , where K has a value of one of the integer number of data being interleaved (data item) K≤5114.

The relationship between bits input to the turbo code internal interleaver and bits input to channel coding is defined as x _k = o _irk and K = K _i . In particular, the following indications are used:

Number of bits input to the K turbo code internal interleaver;

The number of rows in the R rectangular matrix;

Number of columns in the C rectangular matrix;

p prime number;

V primitive root;

<s (j)> _{j ∈ {0, 1,... , p-2}} base sequence for in _-line substitution;

q _i minimum prime;

r _i substituted decimal;

<T (i)> _{i ∈ {0, 1,... , R-1} in-line} substitution pattern;

<U _i (j)> _{j ∈ {0, 1,... , C-1} in-} row substitution pattern of the i-th column;

i row number index of the rectangular matrix;

j column number index of the rectangular matrix; And

Index of k-bit sequence.

As a first step, the bit sequence x ₁ , x ₂ , x ₃ ,... x _K is written to the rectangular matrix by assigning values 5, 10 and 20 to R according to the K value. The rows of the rectangular matrix are from top to bottom 0, 1,... R-1 is numbered.

The second step includes determining the number of columns of the prime p and rectangular matrix C used for intra-permutation. This is done as a function of K and R. The columns of the rectangular matrix are from left to right 0, 1,... , C-1 is numbered.

The third step is to input the sequence of bit sequences x ₁ , x ₂ , x ₃ ,... Into the row by row of the R × C rectangular matrix starting with bit y ₁ in column 0 of row 0 . creating x _K , where k = 1, 2,... , K _k = x _k with respect to K and R x C> K, then k = K + 1, K + 2,. , Dummy bits are filled such that y _k = 0 or 1 for R × C.

These dummy bits are truncated from the output of the rectangular matrix after in-row and inter-line substitution.

After bit input to the RxC rectangular matrix, in-row and interline substitution for the RxC rectangular matrix is performed step by step using the six-step algorithm detailed in the protocol.

Haengnae and after leading substituted, the bit of the substituted rectangular matrix is represented by y _'k. The output of the turbo code interleaver is columnar from the in-row and interlined R × C rectangular matrix, beginning with bit y ' ₁ in column 0 of row 0 and ending with bit y' _CR in column C-1 of row R-1. by cloumn).

The output is truncated by deleting the dummy bits that filled the input of the rectangular matrix before in-row and inter-line substitution; That is, the bit y ' _k corresponding to the bit y _k where k> K is removed from the output.

The output bits from the turbo code internal interleaver are x ' ₁ , x' ₂ ,... x ' _K , where x' ₁ corresponds to bit y ' _k with minimum index k after pruning, x' ₂ corresponds to bit y ' _k with second minimum index k after pruning _{, and} so on. Corresponding.

The number of bit outputs from the turbo code internal interleaver is K and the total number of bluish bits is R × C−K.

All of the foregoing are in accordance with treatment criteria well known to those skilled in the art, and thus do not need to be discussed in more detail herein.

Essentially, the interleaving scheme described in Technical Protocol 3GPP TS 25.212 V.3.5.0. (2000-12) comprises constructing a matrix and inputting data entering the matrix; Intra and interline substitution; And reading the substituted data from the matrix obtained through in-row and inter-line substitution.

The interleaver can be executed using data memory in which incoming data is written sequentially and read in a specific order.

In fact, any interleaving / deinterleaving process may simply be regarded as a result of any of the following operations:

a) taking an interleaver / deinterleaver item (e.g., a bit) from the input file in the order in which it appears in the file at this input and rearranging the item in an output file at each location each identified by a given address; ;

b) taking an item from an input file in the sequence identified by each address and rearranging the item in turn into an output file when taken from the input file.

Further, it will be appreciated that the step of executing all of the previously indicated operations in turn (in any order) corresponds to a complete interleaving / deinterleaving process.

In the following, the criterion for the interleaving process is made for the sake of simplicity, and in other cases it will be appreciated that what is stated in connection with interleaving also applies directly to the deinterleaving.

The deinterleaver module as well as the interleaver module can be more easily executed by the processing structure corresponding to either method.

This is shown schematically in FIG. 1. Here, reference numeral 10 denotes an address generator having an associated data memory 12 intended to cooperate with a turbo encoder / decoder (codec) 14.

The codec 14 is adapted to process the data arranged in the blocks (in a known manner which does not need to be described in detail herein), and each block is composed of K terms which are interleaver / deinterleaved.

Information relating to the K value (which may be changed to each other data packet transmitted and received, as indicated) is sent from the codec 14 to the address generator 10.

The data memory 12 is a basic element for performing interleaving / deinterleaving operations.

For example, this can be caused by having the codec 14 write the data to the memory 12 in order and then read the data from the memory 12 corresponding to the address sequence defined by the address generator 10. Can be. Alternatively, the codec 14 may cause the data to be written to the memory 12 corresponding to the address sequence defined by the address generator 10 and then to read the data from the memory 12 in order.

The address generator 10 described herein consists essentially of three devices: a parameter generator 10a, a substitution pattern generator 10b, and an address generator 10c.

A possible overall arrangement of the parameter generator 10a for use in the apparatus disclosed herein is shown in FIG.

The arrangement shown shows that when a look-up table (LUT) scheme is used, the computational effort required to generate the parameters R, C, p, V required for the construction of the interleaving matrix and the substitution pattern is reduced. It is based on perception.

In other words, instead of actively generating (eg, calculating) these parameters each time a new value of K is encountered (because the length given by K of the interleaved data block changes), for all possible values of K, Once-consider the possibility of calculating these parameters. This parameter can then be explicitly stored in a memory that is selectively retrieved whenever a new value for K is handled.

This approach entails the significant advantage of allowing the address generator to be implemented in essentially hardware form without the need to provide any particular computational software to operate on each new value for K.

This reduces processing time and thus represents a significant advantage since the interleaving process is already relatively time consuming in itself: the entire block of interleaved data actually does not have the proper interleaving process since the last input data for processing may be the first output data. It can be used before it starts.

In the look-up table device under consideration, the input to the look-up table may be the size of the data block entering the turbo encoder K. This can vary within the range [40, 5114] assuming 5075 possible values. The mapping concatenation of the other parameters C, R, p, and v may include the output of the lookup table. The execution of this lookup table may be the memory of 5075 recorders, each recorder containing a respective set of parameters for each possible K value.

However, this approach, which is advantageous, is not very efficient in terms of the required memory size, and may inevitably disadvantage important chip area occupancy parameters for mobile applications such as, for example, UMTS mobile terminals.

The implementation of the reference algorithm described in Technical Protocol 3GPP TS 25.212 V3.5.0. Shows that even though the possible K value is 5075, the number of other records in memory is much smaller than 5075, and in fact is about 200.

As a result, a presently preferred embodiment of the device shown here is provided for a memory having the above considered look-up table which is previously divided into two different memories or memory regions:

A first memory area 20 essentially containing a storage location for a pointer addressed by a K value; And

Second memory area (22) comprising (second) records for each different set of interleaving parameters, each addressed by a pointer.

The number of such (second) records is less than 5075, and in practice is limited to about 200 (ie, the number of other data sets possible due to the execution of the algorithm defined in the protocol).

Thus, the apparatus allows for each given K value (5075 possible values) to identify a corresponding pointer for each record (ie, each set of interleaving parameters) in the second memory area.

Another step to reduce the memory requirements can be taken by noting that the pointers are stored in increasing order in memory, and thus can be arranged in order in groups of adjacent pointers having the same most significant bits (MSB).

This feature can be used to further divide the memory area 20 into two subunits consisting of MSB (most significant bit) calculating circuit 20a and LSB (least significant bit) pointer memory 20b.

By this arrangement, a given number of MSBs for a pointer (e.g., the first two bits in each pointer) compares a number of thresholds and K values in a group of comparators 201, 202, and 203, and in logical unit 204 Can be determined by combining the results of the comparator. Therefore, the remaining, least significant bits (LSBs) can be retrieved from the LSB memory 20b, and the size of the LSB memory is also incidentally reduced. For example, considering 5075 different values for K, the size of memory 20b can be reduced from 5075x8 bits to 5075x6 bits since the two MSBs are already determined by circuit 20a.

By way of example, the thresholds of comparators 201, 202 and 203 can be set to values T1, T2 and T3, respectively, where T1 is lower than T2 and in turn T2 is lower than T3.

Thus, the range of possible K values is divided into four parts or bands. T1 becomes an upper range for the first band, while a lower range is also provided for the second band between T1 and T2. The third band is between T2 and T3, while T3 also represents a subrange for the fourth part or band.

Thus, comparing the instantaneous values of K against the thresholds T1, T2, and T3 in the comparators 201-203 and processing the respective results of the comparison in the circuit 204 are two MSBs of corresponding pointers. Place K in one of these four bands for and generate the values 00, 01, 10 and 11, respectively.

Accordingly, the disclosed pointer search apparatus

Circuitry (20a) for sensing a predetermined number (K) of items to be interleaver / deinterleaved to be derived from (two) most significant bits of the pointer; And

It consists of another memory device 20b having (six) remaining, least significant bits (LSB) of the pointer stored therein.

Of course, a lower or higher number of MSBs can be determined by the circuit 20a, the main purpose of which is to reach the best trade off between memory occupancy and combinatorial logic complexity.

To reduce the computational effort required by subsequent modules included in the interleaver, other information may be stored in a memory location of memory 20.

For example, a binary flag set to "1" when C is p + 1 and K is R * C can be used to generate the output interleaving address without any extra complexity. For reference, this binary flag originally corresponds to the bit designated as Type D in EP-A-1 195 910.

By adding such a flag to a record in memory 22, the number of such records is generally increased, in that the same record may have a different value for the flag in other cases.

However, experiments conducted by the Applicant show that, in this case too, the number of records in the memory 22 slightly exceeds 200, and generally 219 (relative to 5075 other values for K).

Starting from the set of parameters generated by module 10a (these are actually stored as individual records in memory 22), two kinds of substitution patterns must be calculated.

The first pattern, called vector s, is an in-row substitution pattern. According to the standard procedure defined in the 3GPP TS 25.212 V3.5.0. Protocol, this vector initially seeded the value V provided by the earlier processing step (ie, the first module 10a in the embodiment shown herein). Generated as a pseudo random sequence used as an initial seed.

In the apparatus shown here, pseudorandom sequence a * b mod c is generated in module 10b by using a linear algorithm that does not include any multiplication or division. These algorithms are in the field of cryptography as evidenced by the Adnan Gutub's CE 575 project report, Oregon State University's Department of Electrical and Computer Engineering, Winter 2000, article "A Modulo Multiplication Hardware Design." Known in

Moreover, 3GPP TS 25.212 V3.5.0 to have different substitutions for each row. In the standard procedure defined in the protocol, the s vector is addressed by a pseudorandom sequence and the seed of the sequence is different for each row.

In particular, the algorithm described in the protocol involves the generation of a sequence of seeds (r sequence) starting from a few other sequences (q sequences).

In contrast, in the apparatus shown here, the q sequence is used directly without calculation of the r sequence. You can do this by simply swapping the order of intra-line and interline substitution.

This optimization is done in 3GPP TS 25.212 V3.5.0. The same result is achieved without the reverse substitution described in Section 4.2.3.2.3.2- (4) of the Code.

For direct reference, this

r _{T (i)} = q _i , i = 0, 1,... , R-1,

So that the sequence <r _i > _{i∈ {0, 1,... , R-1}} to produce the sequence <q _i > _{i∈ {0, 1,... , R-1}} is required to be substituted, where <T (i)> _{i∈ {0, 1,... , R-1}} is a line substitution pattern defined as one of four types of patterns shown in Table 3 of 3GPP TS 25.212 V3.5.0. As a function of the number of input bits K.

This approach avoids unnecessary reading procedures from memory (not shown) where the T values are stored.

The sequence q is 3GPP TS 25.212 V3.5.0. Can be generated by reading a small lookup table for the management of the decimal table and the greatest common divisor operation described in section 4.2.3.2.3.2- (3) of the protocol.

Also, for direct reference, the operation is performed by the sequence <q _i > _{i∈ {0, 1,... , R-1}} assigns q ₀ = 1 to be the first prime number, and each i = 1,2,... , The sequence <q _i > _{i∈ {0, 1,... So} that gcd (q _i , p-1) = 1, q _i > 6 and q _i > q _{(i-1) for R-1 . , R-1}} determines that prime q _i is at least prime. Where gcd is the greatest common divisor.

In particular, in the block diagram of FIG. 4, reference numeral 30 denotes a block that starts with the parameters derived from block 10b, creating vectors s and q. These vectors are stored in respective memory blocks 31 and 32.

Finally, reference numeral 10c denotes a block that executes the following relationship:

X _r = T (i)

U _c = S (j * q (i) mod (p-1))

X _c = f (U _c )

Address_out = C * X _r + X _c

Where s, q and p are already identified entities, X _r and X _c are the row and column indices of the matrix before the substitution operation, i and j are the substituted matrix indices, and U _c is <U _i (j)> _{j ∈ {0, 1,... , C-1}} , the in-row substitution parameter, and f is a function of U _c that executes the algorithm steps as described in section 4.2.3.2.3.2-5 of the 3GPP TS 25.212 Technical Specification.

The output of block 10c is represented by the interleaving / deinterleaving address used by turbo codec 14 (see FIG. 1).

The techniques described herein to obtain interleaving parameters such as R, C, p, v, flags (type D) directly from K, in conjunction with other arrangements for achieving interline and intraline substitutions already known in the art. It is recognized that it is suitable for use.

Of course, without departing from the basic principles of the invention, the details and embodiments may be changed to the embodiments described by way of example only, without departing from the scope of the invention as defined in the claims below. .

Included in the Detailed Description of the Invention.

Claims (24)

An address generator for an interleaving / deinterleaving sequence (x ₁ , x ₂ , x ₃ , ..., x _K ) comprising a predetermined number K of items, At least one memory device (22) for storing a plurality of records, each record corresponding to at least one of a predetermined number (K) of said items, generated by said at least one value An address generator for an interleaving / deinterleaving sequence, characterized by each set of interleaving / deinterleaving parameters (R, C, p, v). The method of claim 1, Each value for the predetermined number of items (K) arranges the sequence to produce a substituted output matrix for the address generation and intra-row and inter-row permutation of the matrix. Identify corresponding parameter sets (R, C, p, v) for constructing a matrix (R × C) affecting Address generator for an interleaving / deinterleaving sequence, characterized in that it can be used as a record in the at least one memory device (22) for all possible values of a predetermined number (K). The method according to claim 1 or 2, The value of the predetermined number K of the items is within a predetermined value range, and the at least one memory device 22 has the number of records stored substantially smaller than the predetermined value range. An address generator for an interleaving sequence. The method of claim 3, wherein Each said record in said at least one memory device 22 is identified by a respective pointer, and said at least one memory device 22 is assigned to each value of a predetermined number K of said items in said predetermined range. And a pointer retrieval circuit (20) configured to generate a corresponding pointer to each record in said at least one memory device (22) for the interleaving / deinterleaving sequence. The method of claim 4, wherein The pointer search circuit 20 Circuitry (20a) for sensing a predetermined number (K) of items for deriving from a set of most significant bits (MSB) of the pointer; And And a respective memory device (20b) having residual significant bits (LSB) of said pointer stored therein. The method of claim 5, The circuit 20a A plurality of comparators (201, 202, 203) for comparing a predetermined number of said items with a given plurality of thresholds; And And a logic operation unit (204) for combining the results of the comparisons carried out in said comparators (201, 202, 203) and for deriving the most significant bit set of said pointer therefrom. The method of claim 2, The at least one memory device 22 includes, for each said record, at least one flag signal which takes one of a first (0) and a second (1) logical value, said flag being The predetermined number (K) of the items for the corresponding record is equal to the product of the number of rows (R) and the number of columns (C) in the matrix, and the number of columns (C) in the matrix is equal to the in-row substitution plus one. And the second logic value (1) when equal to the value of the used parameter (P). The method according to any one of claims 1 to 7, Generating an arithmetic circuit with a multiplier and a divider for generating a pseudorandom sequence of type a * b mode c to form a substitution pattern for use in said in-row substitution. Device. The method according to any one of claims 1 to 8, An interleaving / d, comprising first and second substitution modules for performing said in-line and said interline substitution, wherein said module performing inter-line substitution is arranged upstream of said module performing in-line substitution. An address generator for an interleaving sequence. The method of claim 9, An in-row module that creates a sequence q that performs the in-row substitution, wherein the in-row module assigns q ₀ = 1 to be a first prime number in the sequence q, each i = 1, 2,… The first prime number q _i in the sequence such that the greatest common divisor (q _i , p-1) = 1, q _i > 6 and q _i > q _{(i-1) for R-1} And an address generator for an interleaving / deinterleaving sequence. The method of claim 10, The in-row module A decimal table for reading at least one prime number; And An address generator for an interleaving / deinterleaving sequence, comprising a look up table that handles the greatest common divisor operation. An address generation method for an interleaving / deinterleaving sequence (x ₁ , x ₂ , x ₃ , ..., x _K ) including a predetermined number K of items, Generating a record representing each set of interleaving / deinterleaving parameters (R, C, p, v) based on a value of at least one of the predetermined number of items (K); And Storing (22) the set of interleaving / deinterleaving parameters (R, C, p, v) in at least one memory device. The method of claim 12, Each value for the predetermined number of items (K) constitutes a matrix (R × C) that arranges the sequence to produce a substituted output matrix for the address generation and affects in-row and inter-row substitution of the matrix. Identify corresponding parameter sets (R, C, p, v) for each of said parameter sets (R, C, p, v) for said at least for all possible values of said predetermined number (K) of said item; A method of generating an address for an interleaving / deinterleaving sequence, comprising the step of enabling it to be used as a record in one memory device (22). The method according to claim 12 or 13, A value of the predetermined number K of the items is included in a given value range, and storing the number of records in the at least one memory device 22 substantially smaller than the given value range. An address generation method for an interleaving / deinterleaving sequence. The method of claim 14, Identifying each said record in said at least one memory device 22 by a respective pointer, and corresponding pointers pointing to each record in said at least one memory device 22 in said predetermined range. Generating for each value of a predetermined number of items (K). The method of claim 15, Derive the most significant bit set of the pointer from a predetermined number K of the items, Retrieving the pointer by storing the remaining, least significant bits of the pointer in each memory device (21b). The method of claim 16, Comparing (201), (202), a predetermined number (K) of said items with a predetermined number of thresholds; And Combining (204) the results of the comparators (201, 202, 203) and deriving therefrom from the most significant bit set of the pointers (204). The method of claim 13, For each record, storing at least one flag signal taking one of a first (0) and a second (1) logical value, wherein the flag is a predetermined number of the items for the corresponding record. (K) is equal to the product of the number of rows (R) and the number of columns (C) in the matrix, and the number of columns (C) in the matrix is equal to the value of the parameter (P) used for the in-row substitution plus one. When the second logical value (1) is set to an interleaving / deinterleaving sequence. The method according to any one of claims 12 to 18, Interleaving / during a pseudorandom sequence of type a * b modec to produce a substitution pattern for use in the in-row substitution, wherein the step is performed with a linear algorithm without multipliers and dividers. Method of generating address for interleaving sequence. The method according to any one of claims 12 to 19, And said interline substitution is performed before said inline substitution. The method of claim 20, Creating a sequence q for performing the in-row substitution, wherein the in-row module assigns q ₀ = 1 to be the first prime number in the sequence q, each i = 1, 2,... The first prime number q _i in the sequence such that the greatest common divisor (q _i , p-1) = 1, q _i > 6 and q _i > q _{(i-1) for R-1} And an address generation method for an interleaving / deinterleaving sequence. 12. A turbo encoder comprising at least one of an interleaver module and a deinterleaver module, said at least one module comprising an arrangement for generating an address according to any one of claims 1-11. 12. A turbo decoder comprising at least one of an interleaver module and a deinterleaver module, said at least one module comprising an arrangement for generating an address according to any one of claims 1-11. 22. A computer program product comprising a portion of software code that can be loaded directly into a memory of a digital computer and when executed on a computer performs a method according to any one of claims 1-21.