DE20023807U1 - Rate matching device for data communication system has one of rate matching blocks that punctures received symbols based on predetermined puncturing pattern in order to rate match received symbols - Google Patents

Abstract

A channel encoder (200) processes the input information bits to obtain the encoded symbols. The rate matching blocks (231-239) whose number is equal to the denominator of the coding rate of the encoder, receive a number of encoded symbols separately. One of the rate matching blocks punctures the received symbols based on a predetermined puncturing pattern in order to rate match the received symbols. Independent claims are also included for the following: (a) a rate matching method; (b) and a symbol determining method.

Description

BACKGROUND OF THE INVENTION

1. Field of the invention

The The present invention relates generally to a channel coding apparatus for a Data transfer system and in particular a device for adapting the rates of channel-coded Symbols.

2. Description of the state of the technique

In digital transmission systems, e.g. Satellite Systems, ISDN (Integrated Services Digital Network) Systems, Digital Cell Systems, W-CDMA (Wideband Code Division Multiple Access) systems, UMTS (Universal Mobile Telecommunication Systems) and IMT-2000 (International Telecommunication-2000) systems, will generally be source user data before sending with one Error correction code encoded to increase the reliability of the system. to Channel encoding will typically be a convolutional code and a linear block code is used, and for the linear block code a single decoder used. In addition to such codes will last Time in big Scope also used a turbo code, which is to send and receive useful for data is.

In Multiple access transmission systems, the multiple users and multi-channel transmission systems with multiple channels support, Channel coded symbols are added to a given number of transmit channel symbols adapted to increase the efficiency of data transmission and to improve the system performance. Such a process is called "rate matching". A rate adjustment is also made to match the output symbol rate to the transmit symbol rate adapt. Typical rate matching techniques include puncturing or repeating parts of channel coded symbols.

A conventional rate matching device is disclosed in US Pat 1 shown. According to 1 encodes a channel coder 100 Input information bits (k) with a coding rate R = k / n and outputs coded symbols (n). A multiplexer (MUX) 110 multiplexes the coded symbols. A rate adjustment block 120 adjusts the rate of the multiplexed coded symbols by puncturing or repeating, and outputs the rate-adjusted symbols to a transmitter (not shown). The channel coder 100 at each cycle of a symbol clock operates at a rate CLOCK, and the multiplexer 110 and the rate matching block 120 operate at every predetermined period of a clock at a speed n × CLOCK.

It should be noted that they rate adjustment device of 1 for application to the case where a non-systematic code, eg, a convolutional code or a linear block code, is used for channel coding. For symbols encoded with a non-systematic code, eg, a convolutional code or a linear block code, because there is no weight between symbols, that is, because the error rate of the channel coder 100 output coded symbols for each symbol within a frame that is represented by the channel coder 100 encoded symbols the rate matching block 120 be fed without distinction and undergo a puncturing or repetition, as in 1 shown.

However, if systematic codes, such as a turbo code, are used, there is a weight between symbols, so it is important for the channel coded symbols that match the rate matching block 120 It is not good to experience dotting or repetition to the same degree. Because the weight between information symbols and parity symbols is not equal, it is recommended that the rate matching block 120 Parity symbols may puncture from the turbo-coded symbols, but should not puncture the information symbols. Alternatively, the rate matching block 120 Repeat the information symbols from the turbo-coded symbols to increase the energy of the symbols, but should not, if possible, repeat the parity symbols. That is, it is difficult to use the rate matching device of 1 to use when a turbo code is used. This is normal given the fact that the structure of 1 only for non-systematic codes, such as convolutional codes or linear block codes, and the turbo code has new features different from those of the convolutional codes and the linear block codes.

In order to solve such a problem, a method of rate matching of the turbo code channel coded symbols has recently been proposed. However, such a method can only be used when adjusting the rate of the turbo-coded symbols, and can not be used in adjusting the rate of the existing convolutional or linear block code channel coded symbols.

It There is therefore a need for a single device and a Method of rate matching of both with an existing not systematic code as well as channel-coded with a systematic code Symbols. For example needed a data transmission system, that is, both non-systematic code and systematic To support code two different structures to adjust both codes in the rate what an increase the complexity leads. If possible is the different codes by means of a single structure Adjusting the rate reduces the complexity of the implementation become.

It It is therefore an object of the present invention to provide a device for rate matching of both with a non-systematic code as well as channel-coded symbols using a systematic code a single structure in a data transmission system put.

It Another object of the present invention is an apparatus for the selective rate adaptation of non-systematic Code channel-coded symbols or with a systematic code to provide channel coded symbols in a data transmission system, this includes both a non-systematic code and a systematic one Code supported.

It Another object of the present invention is a device to provide rate matching of channel coded symbols, about the efficiency of data transmission increase and system performance in a data transmission system to improve.

Around To accomplish the above and other objects will become a device for adjusting a rate of channel coded symbols in a data transmission system proposed. The rate adaptation device may be based on a data transmission system applied either a non-systematic code (convolutional code or linear block code) or a systematic code (turbo code) or both codes are used. The rate adjustment device comprises a plurality of rate matching blocks, the number of which equals one Inverse of the coding rate of the channel coder. The rate adjustment device can be the symbols encoded with a non-systematic code or the symbols encoded with a systematic code in the rate adjust by output parameters that represent the number of input symbols, the number of output symbols and the punctuation / repetition scheme include determining parameters.

SUMMARY THE DRAWINGS

The Above and other objects, features and advantages of the present invention The invention will become more apparent from the following detailed description as shown in the accompanying drawings is brought.

1 Fig. 12 is a drawing illustrating a structure of a prior art rate matching device.

2 and 3 FIG. 13 is drawings illustrating structures of rate matching devices according to embodiments of the present invention. FIG.

4 Fig. 12 is a drawing illustrating a structure of a rate matching apparatus by puncturing according to an embodiment of the present invention.

5 Fig. 12 is a drawing illustrating a structure of a rate matching apparatus by puncturing according to another embodiment of the present invention.

6 is a detailed drawing showing a structure of the in 5 represents shown turbo encoder.

7 FIG. 10 is a flowchart illustrating a rate matching procedure by puncturing according to an embodiment of the present invention. FIG.

8th Fig. 12 is a drawing illustrating a structure of a rate matching device by puncturing according to another embodiment of the present invention.

9 Fig. 12 is a drawing illustrating a structure of a rate matching device by repeating according to an embodiment of the present invention.

10 Fig. 12 is a drawing illustrating a structure of a rate matching device by repeating according to another embodiment of the present invention.

11 FIG. 10 is a flowchart illustrating a rate matching procedure by repeating according to an embodiment of the present invention. FIG.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT

preferred versions The present invention will be described below with reference to the accompanying drawings. In the following description are not well-known functions or constructions in detail described because they would tarnish the invention in unnecessary detail.

When canceled a rate adjustment device required conditions

• Voraussetzung 1A: Eine Eingangssymbolfolge aus codierten Symbolen sollte mittels eines Punktierungsschemas mit einer spezifizierten Periode punktiert werden.
• Voraussetzung 2A: Die Anzahl punktierter Bits aus den Eingangssymbolen sollte, wenn möglich, minimiert werden.
• Voraussetzung 3A: Ein gleichmäßiges Punktierungsschema sollte benutzt werden, so dass die Eingangssymbolfolge, die aus von einem Codierer ausgegebenen codierten Symbolen besteht, gleichmäßig punktiert wird.
• Voraussetzung 1C: Eine Eingangssymbolfolge aus codierten Symbolen sollte mittels eines Wiederholungsschemas mit einer spezifizierten Periode wiederholt werden.
• Voraussetzung 2C: Ein gleichmäßiges Wiederholungsschema sollte benutzt werden, so dass die Eingangssymbolfolge, die aus von einem Codierer ausgegebenen Symbolen besteht, gleichmäßig wiederholt wird.

Before describing the invention, reference is first made to conditions which should be considered when adapting channel-coded symbols in rate with a non-systematic code, eg, a convolutional code or a linear block code (in the following description, the non-systematic code assumed as a convolutional code). The prerequisites 1A to 3A below are the prerequisites that should be considered when adjusting punctured coded symbols in the rate, and conditions 1C and 2C below are the prerequisites that should be considered when repeating coded symbols in the rate be adjusted.

• Condition 1A: An input symbol sequence of coded symbols should be punctured using a puncturing scheme with a specified period.
• Condition 2A: The number of punctured bits from the input symbols should be minimized, if possible.
• Prerequisite 3A: A uniform puncturing scheme should be used so that the input symbol sequence consisting of coded symbols output by an encoder is equally punctured.
• Condition 1C: An input symbol string of coded symbols should be repeated by means of a repetition scheme with a specified period.
• Condition 2C: A uniform repetition scheme should be used so that the input symbol sequence consisting of symbols output from an encoder is uniformly repeated.

These Prerequisites are based on the assumption that the error rate that output from the encoder using a convolutional code Symbols for every symbol within a frame (or codeword) is about the same is. Indeed It is known that if the above prerequisites are the main one Limiting factors when executing puncturing are used for rate matching, positive results as shown in the following references: [1] G. D. Forney, "Convolutional codes I: Algebraic structure ", IEEE Trans. Inform. Theory, Vol. IT-16, pp. 720-738, Nov. 1970, [2] J.B. Cain, G.C. Clark, and J.M. Geist, "Punctured convolutional codes of a rate (n-1) / n and simplified maximum likelihood decoding ", IEEE Trans. Inform. Theory, Vol. IT-25, Pages 97-100, Jan 1979.

• Voraussetzung 1B: Da ein Turbocode ein systematischer Code ist, sollte der Teil, der Informationssymbolen aus den durch den Codierer codierten Symbolen entspricht, nicht punktiert werden. Außerdem sollte aus dem weiteren Grund, dass ein iterativer Decoder als Decoder für den Turbocode verwendet wird, der den Informationssymbolen entsprechende Teil nicht punktiert werden.
• Voraussetzung 2B: Da ein Turbocodierer aus zwei parallel geschalteten Komponentencodierern besteht, ist vorzuziehen, den minimalen freien Abstand jedes der zwei Komponentencodierer für den minimalen Abstand des ganzen Codes zu maximieren. Um optimale Leistung zu erlangen, sollten deshalb die ausgegebenen Paritätssymbole der zwei Komponentencodierer gleichmäßig punktiert werden.
• Voraussetzung 3B: Da in den meisten iterativen Decodern die Decodierung von dem ersten internen Decoder durchgeführt wird, sollte das erste Ausgangssymbol des ersten Komponentencodierers nicht punktiert werden. Mit anderen Worten, das erste Symbol eines Codierers sollte nicht punktiert werden, ungeachtet ob es eine Systematik oder Paritätsbits sind, weil das erste Symbol den Anfangspunkt der Codierung angibt.
• Voraussetzung 4B: Die ausgegebenen Paritätssymbole jedes Komponentencodierers sollten mittels eines gleichmäßigen Punktierungsschemas punktiert werden, so dass die von dem Codierer ausgegebenen codierten Symbole, z.B. der bestehende Konvolutionalcode, gleichmäßig punktiert werden.
• Voraussetzung 5B: Die für den Turbocodierer benutzten Abschlussschwanzbits sollten wegen der nachteiligen Wirkung auf die Leistung des Decoders nicht punktiert werden. Zum Beispiel hat ein SOVA- (Soft Output Viterbi Algorithm) Decoder eine niedrige Leistung, wenn die Abschlussschwanzbits punktiert werden, verglichen mit dem Fall, wo die Abschlussschwanzbits nicht punktiert werden.
• Voraussetzung 1D: Da ein Turbocode ein systematischer Code ist, sollte ein Teil, der Informationssymbolen aus den durch den Codierer codierten Symbolen entspricht, wiederholt werden, um die Energie der Symbole zu steigern. Außerdem sollten, da ein iterativer Decoder als Decoder für den Turbocode benutzt wird, der den Informationssymbolen entsprechende Teil oft wiederholt werden.
• Voraussetzung 2D: Da ein Turbocodierer aus zwei parallel geschalteten Komponentencodierern besteht, ist vorzuziehen, den minimalen freien Abstand jedes der zwei Komponentencodierer für den minimalen freien Abstand des ganzen Codes zu maximieren. Deshalb sollten, wenn die Paritätssymbole wiederholt werden, die ausgegebenen Paritätssymbole der zwei Komponentencodierer gleichmäßig wiederholt werden, um optimale Leistung zu erzielen.
• Voraussetzung 3D: Da in den meisten iterativen Decodern die Decodierung von dem ersten internen Decoder durchgeführt wird, sollte das erste Ausgangssymbol des ersten Komponentendecoders vorzugsweise wiederholt werden, wenn die Paritätssymbole wiederholt werden.
• Voraussetzung 4D: Die ausgegebenen Paritätssymbole jedes Komponentencodierers sollten mittels eines gleichmäßigen Wiederholungsschemas wiederholt werden, so dass die von dem Codierer ausgegebenen codierten Symbole, z.B. der bestehende Konvolutionalcode, gleichmäßig wiederholt werden.
• Voraussetzung 5D: Die für den Turbocodierer benutzten Abschlussschwanzbits sollten wegen der Auswirkung auf die Leistung des Decoders wiederholt werden. Zum Beispiel hängt die Leistung eines SOVA- (Soft Output Viterbi Algoritm) Decoders davon ab, ob die Abschlussschwanzbits wiederholt werden oder nicht.

Next, reference is made to the requirements that should be considered when adapting channel-coded symbols in rate with a systematic code (in the following description, it is assumed that the systematic code is a turbo code). Conditions 1B through 5B below are the prerequisites that should be considered when rate matching the punctured coded symbols, and assumptions 1D through 5D are the prerequisites that should be considered when rate matching the symbols coded by repetition.

• Condition 1B: Since a turbo code is a systematic code, the portion corresponding to information symbols from the symbols encoded by the encoder should not be punctured. In addition, for the other reason that an iterative decoder is used as a decoder for the turbo code, the part corresponding to the information symbols should not be punctured.
• Condition 2B: Since a turbo encoder consists of two component encoders connected in parallel, it is preferable to maximize the minimum free space of each of the two component encoders for the minimum distance of the whole code. For optimal performance, therefore, the output parity symbols of the two component encoders should be dotted evenly.
• Prerequisite 3B: Since in most iterative decoders, the decoding of the first internal deco which is performed, the first output symbol of the first component coder should not be punctured. In other words, the first symbol of an encoder should not be punctured, regardless of whether it is a system or parity bits, because the first symbol indicates the starting point of the encoding.
• Condition 4B: The output parity symbols of each component coder should be punctured using a uniform puncturing scheme so that the coded symbols output by the coder, eg the existing convolutional code, are punctured evenly.
• Condition 5B: The tail tail bits used for the turbo encoder should not be punctured because of the adverse effect on the performance of the decoder. For example, a SOVA (Soft Output Viterbi Algorithm) decoder has low power when the tail tail bits are punctured compared to the case where the tail bits are not punctured.
• Condition 1D: Since a turbo code is a systematic code, a part corresponding to information symbols from the symbols encoded by the encoder should be repeated to increase the energy of the symbols. In addition, since an iterative decoder is used as a decoder for the turbo code, the part corresponding to the information symbols should be repeated often.
• Condition 2D: Since a turbo encoder consists of two component coders connected in parallel, it is preferable to maximize the minimum free space of each of the two component coders for the minimum free space of the whole code. Therefore, if the parity symbols are repeated, the output parity symbols of the two component encoders should be evenly repeated to obtain optimum performance.
• Requirement 3D: Since in most iterative decoders the decoding is performed by the first internal decoder, the first output symbol of the first component decoder should preferably be repeated when the parity symbols are repeated.
• Prerequisite 4D: The output parity symbols of each component coder should be repeated by means of a uniform repetition scheme so that the coded symbols output by the coder, eg the existing convolutional code, are evenly repeated.
• Requirement 5D: The tail tail bits used for the turbo encoder should be repeated because of the effect on the performance of the decoder. For example, the performance of a SOVA (Soft Output Viterbi Algoritm) decoder depends on whether the tail tail bits are repeated or not.

The The present invention is intended to provide a rate matching device to implement that not only meets 1A-3A and 1C-2C but also meets the requirements 1B-5B and 1D-5D. That is, one Rate adjustment device according to the invention by puncturing serves as a rate matching device which Requirements 1A to 3A for convolutionally encoded symbols, and also serves as a rate matching device satisfying the requirements 1B to 5B for met turbocoded symbols. The rate adjustment device according to the invention by repetition serves as a rate matching device which the requirements 1C to 2C for convolutionally encoded symbols, and also serves as a rate matching device satisfying the requirements 1D to 5D for met turbocoded symbols.

basic structure the rate adjustment device

Embodiments of the rate matching device structures according to the invention are disclosed in US Pat 2 and 3 shown. In particular shows 2 an example of a hardware implemented rate matching device according to an embodiment of the present invention, and 3 FIG. 12 shows an example of a software implemented rate adaptation device according to one embodiment of the present invention. FIG.

According to 2 Channel encodes a channel encoder 200 input information bits at a coding rate R = k / n and outputs coded symbols. Here, n is the number of coded symbols that make up a codeword, and k is the number of input information bits that make up an input information word. There are n rate adjustment blocks 231 - 239 , each receiving coded symbols separately from the channel coder 200 are output by a number of input symbols determined according to the coding rate, and the received symbols puncture or repeat. The n rate adjustment blocks 231 - 239 each separately receive the coded symbols received from the channel coder 200 by the number determined by multiplying the number of coded symbols in a frame by the coding rate. For example, if the number of coded symbols in a frame is 10 and the coding rate is R = 1/5, the 5 rate matching blocks each separately receive 2 symbols. Each of the rate adjustment blocks 231 - 239 punctures the received symbols according to a predetermined puncturing scheme or repeats the received symbols according to a predetermined repeating scheme. A multiplexer 240 multiplexes the rate-adjusted symbols from the rate matching blocks 231 - 239 and outputs the multiplexed symbols to a channel transmitter (not shown). Since the channel transmitter is beyond the scope of the present invention, a detailed description thereof will be obviated herein. The rate matching operation of the rate matching blocks 231 - 239 will become more apparent from the following detailed description of the embodiments of the present invention.

According to 3 Channel encodes a channel encoder 200 input information bits at a coding rate R = k / n and outputs the coded symbols. A digital signal processor (DSP) 250 with a rate matching module, rate matching (or puncturing / repeating) passes through the channel coder 200 channel-coded symbols by means of the rate matching module. The through the DSP 250 rate adjusted symbols are output to the channel transmitter. The Rate Matching DSP 250 separately receives the coded symbols of one frame of n separate data streams, the number of symbols received from each stream being equal to the number of input symbols determined according to the coding rate, and punctures / repeats the received symbols in the same manner as in FIG 2 shown. In other words, although the DSP 250 is a single element in hardware, it performs the same rate matching operation as the n rate matching blocks of 2 by. The DSP 250 can also be implemented with a CPU (Central Processing Unit), and the rate matching operation can be implemented by a subroutine. As used herein, the term "rate matching blocks" is also intended to be the rate matching modules in the DSP 250 affect.

As in 2 and 3 As shown, a rate matching apparatus of the present invention may have a structure including as many rate matching blocks as the number corresponding to the coding rate (ie, an inverse of the coding rate if k = 1, but if k <1, then the number of rate matching blocks may be equal to one reciprocal the coding rate multiplied by k), and each rate matching block receives as many symbols as the number determined by multiplying the number of coded symbols in a frame by the coding rate, and punctures the received symbols according to a predetermined puncturing scheme or repeats the received symbols according to a predetermined one rep scheme. This structure has the feature that the channel coded symbols are processed separately while the conventional rate matching apparatus of 1 processing the channel coded symbols in a frame unit. The rate matching device according to the invention can be used both for convolutional codes and for turbo codes. That is, the rate matching device of the present invention has a single structure that can be applied to both convolutional codes and turbo codes, although two different sets of prerequisites are needed.

A rate matching device according to the invention may also have a structure of 8th to have. This rate matching device has a combined structure of the conventional rate matching device of FIG 1 and the novel rate matching device of 2 and 3 , Including a single rate matching block, the rate matching device, though implemented by hardware, has a low complexity.

According to 8th Channel encodes a channel encoder 200 input information bits at a coding rate R = k / n and outputs coded symbols. The coded symbols are passed through a multiplexer 260 multiplexed, and the multiplexed symbols are sent to a rate matching block 230 output. The through the rate adjustment block 230 Symbols adjusted by puncturing / repeating are sent to a channel transmitter. A RAM (Random Access Memory) 270 stores an initial value that is entered by the rate matching block 230 rate adaptation received, and passes the output value to the rate matching block 230 , The channel coder 200 at each period of the symbol clock, it operates at a rate CLOCK, and the multiplexer 260 and the rate matching block 230 operate at a predetermined period of a clock at a speed of n × CLOCK. The one to the RAM 270 The output value passed includes the input symbol number Nc, the output symbol number Ni, the error value 'e', and the puncturing / repetition pattern determining parameter 'a' and 'b'. The number of symbols to be punctured for each frame of the coded symbols is determined by the input symbol number Nc and the output symbol number Ni. The RAM 270 stores the input symbol number Nc corresponding to each symbol clock in a predetermined period, the output symbol number Ni, the error value 'e', and the puncturing / repetition pattern determining parameters 'a' and 'b'. When the rate matching is done by puncturing, the rate matching block is received 230 the corresponding input symbol number Nc, the output symbol number Ni, the error value 'e' and the puncturing scheme determining parameters 'a' and 'b' stored in the RAM 270 are stored, at each symbol clock period, to determine whether the single symbol processed at each symbol clock period must be punctured, and puncturing according to the corresponding puncturing scheme. When the rate matching is done by repeating, the rate matching block is received 230 the corresponding input symbol number Nc, the output symbol number Ni, the Error value 'e' and the repetition scheme determining parameters 'a' and 'b', which are stored in RAM 270 at every symbol clock period, to determine whether the single symbol processed at each symbol clock period must be repeated, and performs the repetition according to the corresponding repetition scheme.

If a convolutional code or a linear block code in the channel coder 200 is used, the output value is in RAM 270 to a specific puncturing / repetition parameter (Nc, Ni, e, b, a). That is, the rate matching block (RMB) 230 works like in 1 shown without the ram 270 to update.

If in the channel coder 200 a turbo code is used should the rate matching block 230 sequentially from RMB1 to RMBn (each RMBx [x = 1 to n] connected to a set of values for Nc, Ni, e, b, and a) operate at each symbol clock period designated as period 'n' (ie, period n = the Period of a clock at a speed CLOCK). In other words, every period of a clock at the speed of n × CLOCK becomes the rate matching block 230 with the values for Nc, Ni, e, a and b of one of the RMBx [x = 1 to n] updated. For each period of n, the rate matching block thus becomes 230 updated with the values for Nc, Ni, e, b and a from each of the RMBx. For example, during a period of 1 / (n × CLOCK), the rate matching block 230 receive the values for Nc, Ni, e, a, and b from the RMB1, and then at the next period of 1 / (n × CLOCK) receive the values for Nc, Ni, e, a, and b from the RMB2, and so on until the values from the RMBn through the rate matching block 230 were received. The same cycle is then repeated again in the next period 'n'. The state values of RMBx processed at a particular time, that is, the parameter values (Nc, Ni, e, b, a) for determining the symbols and the puncturing / repeating schemes, therefore, become in RAM 270 saved for the process at the next time. Therefore, if this value is used when the RMBx is next processed again, it is possible to perform the operation of n RMBs (RMB1-RMBn) with a single RMB. As for a processing rate n × TAKT as in 1 and 2 shown, the complexity will not be increased.

Meanwhile, receive in 2 the rate adjustment blocks 231 - 239 each separated by the channel coder 200 coded symbols such as the number determined by multiplying the number of coded symbols in a frame by the coding rate. It should be noted, however, that each of the rate matching blocks 231 - 239 Also separated is a different number of channels through the channel 200 can receive encoded symbols. For example, one of the rate matching blocks 231 - 239 receive separately a number of coded symbols smaller than the number determined by multiplying the number of coded symbols in one frame by the coding rate, and another rate matching block could separately receive a number of coded symbols greater than that by multiplying the number of coded symbols coded symbols in a frame with the coding rate specific number. For the sake of simplicity, however, a case will be described where each of the rate matching blocks 231 - 239 separated by the channel coder 200 encoded symbols.

versions the rate adjustment device

It A description will now be given of the rate matching apparatus according to one execution of the present invention. For the sake of convenience, the Description assuming that the coding rate R = 1/3 is and 3 rates adjustment blocks to be provided. It should be noted, however, that the rate adjustment device according to the invention for each In the case where there are n rate matching blocks, that is, the coding rate R = k / n. Further referred to in the following description Ncs is the total number of coded symbols output by the channel encoder in a frame. Nc returns the number in each rate adjustment block entered symbols and the number of entered symbols is displayed as Nc = R × Ncs certainly. In the following description, R × Ncs = 1/3 × Ncs = Ncs / 3 system. Ni denotes the number of each rate matching block and the number of symbols output as Ni = RxNis determines what is in the description Nis / 3, where Nis is the total number is the symbol output after the rate matching process. The is called, Nis is the total number of the respective rate adjustment block output symbols. The number of times through each rate adjustment block to puncturing / repeating symbols (bits) is therefore called y = Nc - Ni certainly. The values Nc and Ni can vary.

Furthermore, the invention uses the parameters 'a' and 'b', which are integers determined according to a puncturing / repeating scheme in a frame, that is, integers for determining the puncturing / repeating scheme. The parameter 'a' is an offset value for determining the position of the first symbol in the punctuation / repetition scheme. That is, the parameter 'a' determines which of the coded symbols contained in a frame is to be taken as the first symbol of the puncturing / repeating scheme. When a value of the parameter 'a' increases, a symbol located at the front of the frame is punctured or repeated. The parameter 'b' is a value for controlling the puncturing or repetition period in the frame. By changing this parameter value, it is possible to puncture all the coded symbols contained in the frame.

As As described above, a rate matching device according to the invention the rate adjustment not only by puncturing, but also by Repeat. The description of a rate matching device according to the invention splits into a device for performing the rate adaptation Puncturing and a device for performing rate matching by To repeat.

A. Designs the rate adaptation device by puncturing

1. Execution of the Rate adaptation device by puncturing (for a convolutional code)

4 Fig. 10 shows the structure of a rate matching device by puncturing according to an embodiment of the present invention. This structure is used when the rate matching devices of 2 and 3 Adapt convolutionally encoded symbols by puncturing in the rate.

According to 4 encodes a convolutional encoder 210 input information bits Ik at a coding rate R = 1/3 and outputs encoded symbols C1k, C2k and C3k. The coded symbols C1k, C2k and C3k are separated into rate matching blocks 231 . 232 respectively. 233 to hand over. The first rate adjustment block 231 dotted the coded symbol C1k. Here, the puncturing process is performed on the basis of the dotted symbol number y = Nc-Ni determined by the input symbol number Nc, the output symbol number Ni, and the puncturing scheme determining parameters 'a' and 'b'. For example, the first rate matching block 231 output the symbols '... 11x10x01x ...' (where x indicates a dotted symbol). The second rate adjustment block 232 dotted the coded symbol C2k. Here, the puncturing process is performed on the basis of the dotted symbol number y = Nc-Ni determined by the input symbol number Nc, the output symbol number Ni, and the puncturing scheme determining parameters 'a' and 'b'. For example, the second rate matching block 232 output the symbols '... 11x11x10x ...' (where x indicates a dotted symbol). The third rate adjustment block 233 dotted the coded symbol C3k. Here, the puncturing process is performed on the basis of the dotted symbol number y = Nc-Ni determined by the input symbol number Nc, the output symbol number Ni, and the puncturing scheme determining parameters 'a' and 'b'. For example, the third rate matching block 233 output the symbols '... 01x11x11x ...' (where x indicates a dotted symbol). The through the rate adjustment blocks 231 . 232 and 233 rate-matched coded symbols are passed through a multiplexer 240 (in 4 not shown) and handed over to a channel transmitter.

In 4 For example, the input symbol number Nc and the output symbol number Ni are determined similarly as Nc = R × Ncs and Ni = R × Nis for each rate matching block. Each rate matching block separately punctures the same number of channel-coded symbols on the assumption that the error rate of coded symbols is nearly the same for each symbol in a frame. That is, within a frame, an almost uniform puncturing scheme is provided regardless of the various dotted numbers determined according to the service type. This is because of the ability to uniformly puncture all the symbols in a convolutional code frame.

According to an embodiment of the present invention, therefore, the convolutional coder 210 encoded symbols separated and in equal numbers to the rate matching blocks 231 . 232 and 233 to hand over. The rate adjustment blocks 231 . 232 and 233 each dot the same number of input symbols. At this point, the punctuation schema parameters can be determined to be the same or different. That is, the punctuation schemes may be for the rate matching blocks 231 . 232 and 233 be determined either the same or different.

2. Another version of the Rate adaptation device by puncturing for turbo code)

5 FIG. 10 shows a structure of a rate matching device according to another embodiment of the present invention. FIG. This structure is used when the rate matching devices of 2 and 3 adjust the turbocoded symbols by puncturing in the rate.

According to 5 encodes a turbo encoder 220 input information bits Ik at a coding rate R = 1/3 and outputs encoded symbols C1k, C2k and C3k. Among the coded symbols, the information symbol C1k is separated to the first rate matching block 231 and the parity symbols (or redundancy symbols) C2k and C3k are separated to the second and third rate matching block 232 respectively. 233 to hand over. The turbo encoder 220 consists of a first component encoder 222 a second component encoder 224 and a nestler 226 , as in 6 shown. The construction of the turbo coder 220 is well known to those skilled in the art. A detailed description is therefore omitted. The input X (t) into the turbo encoder 220 corresponds to the in 5 shown input information bits Ik. Outputs X (t), Y (t) and Y '(t) correspond to those in 5 shown encoded symbols C1k, C2k and C3k. For the first output of the turbo coder 220 the input information bits Ik = X (t) are output as they are, so that in 5 the input information bits Ik are output as C1k.

The first rate adjustment block 231 punctured coded symbols C1k based on the following criteria. Since the coding rate is R = 1/3, the input symbol number Nc is determined as Nc = R × Ncs = Ncs / 3, which is 1/3 of the total number of coded symbols. The output symbol number Ni is also determined as Ni = R × Ncs because, according to requirement 1B, the part corresponding to the information bits is not punctured. The puncturing scheme determining parameters 'a' and 'b' can be set to an integer, which is irrelevant because according to Condition 1B no puncturing is performed. For example, the first rate matching block 231 output the symbols '... 111101011 ...'.

The second rate adjustment block 232 punctured coded symbols C2k based on the following criteria. Since the coding rate is R = 1/3, the input symbol number Nc is determined as Nc = R × Ncs = Ncs / 3, which is 1/3 of the total number of coded symbols. According to conditions 2B and 4B, since the outputted parity symbols of the two component decoders are to be equally punctured and the total output symbol number after puncturing for the entire input symbols (Ncs) in one frame is Nis, the number Ni is that of the second rate matching block 232 Symbols output after puncturing Ni = [Nis - (R × Ncs)] / 2. When Ni = [Nis - (R × Ncs)] / 2 is an odd number, the output symbol number becomes Ni = [Nis - (R × Ncs) + 1] / 2 or Ni = [Nis-R × Ncs) -1] / 2. One of the two values becomes according to the relationship between the second rate matching block 232 and the third rate matching block 233 selected. That is, when the output symbol number of the second rate matching block 232 when [Nis - (R × Ncs] + 1] / 2) is determined, the output symbol number of the third rate matching block becomes 233 as [Nis - (R × Ncs) - 1] / 2 be true. In contrast, when the output symbol number of the second rate matching block becomes 232 is determined as [Nis - (R × Ncs) - 1] / 2, the output symbol number of the third rate matching block 233 determined as [Nis - (R × Ncs) + 1] / 2.

The puncturing scheme determining parameters 'a' and 'b' may be selected as integers according to a desired puncturing scheme. These integers are determined only according to the punctuation scheme, and the parameters can be set to b = 1 and a = 2. A detailed description of a method of determining integers for the punctuation scheme determining parameter is provided by reference to the tables below. The second rate adjustment block 232 can, for example, output the symbols '... 11x11x10x ...' (where x indicates a dotted symbol).

The third rate adjustment block 233 punctuates the coded symbols based on the following criteria. Since the coding rate is R = 1/3, the input symbol number is determined as Nc as Nc = R × Ncs = Ncs / 3, which is 1/3 of the total number of input symbols (coded symbols). Because, according to conditions 2B and 4B, the entire output parity symbols of the two component decoders are to be dotted evenly and the total output symbol number after puncturing is Nis for all the input symbols in a frame, the number is that of the third rate matching block 233 output symbols after puncturing Ni = [Nis - (R × Ncs)] / 2. If this expression gives an odd number, the output symbol number becomes Ni = [Nis - (R × Ncs) + 1] / 2 or [Nis - (R × Ncs) -1] / 2. One of the two values becomes according to the relationship between the second rate matching block 232 and the third rate matching block 233 selected. That is, when the output symbol number of the second rate matching block 232 is determined as [Nis - (R × Ncs) + 1] / 2, the number of output symbols of the third rate matching block becomes 233 as [Nis - (R × Ncs) - 1] / 2. On the other hand, if the output symbol number of the second rate matching block 232 is determined as [Nis - (R × Ncs) -1] / 2, the output symbol number of the third rate matching block becomes 233 determined as [Nis - (R × Ncs) + 1] / 2.

The puncturing scheme determining parameters 'a' and 'b' may be selected as integers according to a desired puncturing scheme. These integers are determined only according to the punctuation scheme, and the parameters can be set to b = 1 and a = 2. A detailed Be A method of determining the integers for the punctuation scheme determining parameter is described with reference to the tables below. The third rate adjustment block 233 can, for example, output the symbols '... 01x11x11x ...' (where x indicates a dotted symbol).

In 5 be through the turbo encoder 220 coded symbols separated and then equal numbers to the rate matching blocks 231 . 232 and 233 to hand over. The first rate adjustment block 231 returns the input symbols as they are. The second and third rate matching block 232 and 233 punctuate the same number of input symbols. At this point, the punctuation schemes can be determined to be the same or different. That is, the punctuation scheme may be for the rate matching blocks 232 and 233 be determined the same or different.

3. Determination parameters for puncturing

In the embodiments of the present invention discussed herein, the rate matching blocks puncture the same number of symbols (except the rate matching block 231 from 5 ). The rate matching blocks, however, can puncture different numbers of symbols. When the number Ni of the symbols output from the respective rate matching blocks is set differently, the number of symbols punctured by the respective rate matching blocks is determined differently. Further, by changing the puncturing scheme determining parameters 'a' and 'b', the scheme of the punctured by the respective rate matching blocks can be determined to be the same or different. That is, although it has a single structure, a rate matching apparatus of the present invention may set parameters, such as the input symbol number, the output symbol number, the number of symbols to be punctured, and the puncturing scheme determining parameters differently. Table 1 below shows as an example different cases of the parameters. The coding rate is assumed here as R = 1/3. Therefore, three rate matching blocks are provided, and the respective rate matching blocks separately receive the same number of symbols, ie, Nc = Ncs / 3 symbols. The rate matching blocks herein separately receive the same symbol number which is determined by multiplying the number of coded symbols by the coding rate. It should be noted, however, that the present invention can also be applied to a case where the rate matching blocks separately receive a different number of symbols, that is, a number of symbols smaller than the number obtained by multiplying the number of symbols coded symbols in a frame is determined at the coding rate, or a number of symbols greater than the number determined by multiplying the number of coded symbols in a frame by the coding rate. In the description below, RMB1, RMB2 and RMB3 respectively denote the first to third rate matching blocks.

[Table 1]

• Fall 1, Fall 2: In Fall 1 und Fall 2 werden die Symbole in einem Rahmen nach einem gleichmäßigen Schema punktiert. Insbesondere haben im Fall 1 die Ratenanpassungsblöcke dasselbe Punktierungsschema, weil die Parameter "a" und "b" gleich sind, und im Fall 2 haben die Ratenanpassungsblöcke verschiedene Punktierungsschemas, weil die Parameter "a" und "b" verschieden sind.
• Fall 3: Beim systematischen Punktieren werden Informationssymbole nicht punktiert, sondern die Paritätssymbole werden punktiert. Da hier die das Punktierungsschema bestimmenden Parameter 'a' und 'b' gleich sind, führen RMB2 und RMB3 eine gleichmäßige Punktierung halb und halb mittels desselben Punktierungsschemas durch.
• Fall 4: Beim systematischen Punktieren werden Informationssymbole nicht punktiert, und die Paritätssymbole werden punktiert. Da hier die das Punktierungsschema bestimmenden Parameter 'a' und 'b' verschieden sind, führen RMB2 und RMB3 eine gleichmäßige Punktierung halb und halb mittels verschiedener Punktierungsschemas durch.
• Fall 5: Dies ist ein allgemeiner Fall für Fall 3. In diesem Fall wird der das Punktierungsschema bestimmende Parameter 'a' auf eine Ganzzahl 'p' gesetzt, so dass es möglich sein kann, die verschiedenen Punktierungsschemas festzulegen. Der Parameter 'a' wird für RMB2 und RMB3 auf denselben Wert gesetzt.
• Fall 6: Dies ist ein allgemeiner Fall für Fall 4. In diesem Fall wird der das Punktierungsschema bestimmende Parameter 'a' auf Ganzzahlen 'p' und 'q' gesetzt, so dass es möglich sein kann, die verschiedenen Punktierungsschemas festzulegen. Der Parameter 'a' wird für RMB2 auf 'p' und für RMB3 auf 'q' gesetzt.
• Fall 7: Dies ist ein weiterer allgemeiner Fall für Fall 5. In diesem Fall wird der das Punktierungsschema bestimmende Parameter 'a' auf eine Ganzzahl 'p' gesetzt, und der das Punktierungsschema bestimmende Parameter 'b' wird auf eine Ganzzahl 'q' gesetzt, so dass es möglich sein kann, die verschiedenen Punktierungsschemas festzulegen. Die Parameter 'a' und 'b' werden für RMB2 und RMB3 auf denselben Wert gesetzt.
• Fall 8: Dies ist ein weiterer allgemeiner Fall für Fall 6. In diesem Fall wird der das Punktierungsschema bestimmende Parameter 'a' für RMB2 und RMB3 auf Ganzzahlen 'p' bzw. 'r' gesetzt, und der das Punktierungsschema bestimmende Parameter 'b' wird für RMB2 und RMB3 auf Ganzzahlen 'q' bzw. 's' gesetzt, so dass es möglich sein kann, die verschiedenen Punktierungsschemas festzulegen. Die Parameter 'a' und 'b' werden für RMB2 auf 'p' und 'q' und für RMB3 auf 'r' und 's' gesetzt.
• Fall 9, Fall 10: In diesen Fällen werden alle möglichen Parameter geändert. Das heißt, die Ausgangssymbolzahl kann auf jede Ganzzahl gesetzt werden, und die das Punktierungsschema bestimmenden Parameter 'a' und 'b' können auch auf jede gegebene Ganzzahl gesetzt werden.

In Table 1, RMB1, RMB2, RMB3 denote rate matching blocks, and p, q, r, s, t, w, x, y and z are integers. In cases 9 and 10, (1 / p + 1 / q + 1 / r) = 1.0. This is because Nis (1 / p + 11q + 1 / t) = Nis. NA (not present) indicates that the input symbols are output as they are without puncturing, whereby the parameters 'a' and 'b' can be set to any value. The parameters 'a' and 'b' are positive numbers here. Further, the case is shown where the input symbols are punctured to perform the rate matching so that the number of input symbols is larger than the number of the output symbols (ie, Ncs> Nis). It will be referenced in any case.

• Case 1, Case 2: In Case 1 and Case 2, the symbols in a frame are dotted according to a uniform scheme. Specifically, in case 1, the rate matching blocks have the same punctuation scheme because the parameters "a" and "b" are the same, and in case 2, the rate matching blocks have different puncturing schemes because the parameters "a" and "b" are different.
• Case 3: In systematic puncturing, information symbols are not punctured, but the parity symbols are punctured. Since here the punctuation scheme determining parameters 'a' and 'b' are the same, RMB2 and RMB3 perform equal puncturing half and half by the same puncturing scheme.
• Case 4: In systematic puncturing, information symbols are not punctured and the parity symbols are punctured. Since here the puncturing scheme determining parameters 'a' and 'b' are different, RMB2 and RMB3 perform a uniform puncturing half and half by means of different puncturing schemes.
• Case 5: This is a general case for Case 3. In this case, the punctuation scheme determining parameter 'a' is set to an integer 'p' so that it may be possible to set the various puncturing schemes. Parameter 'a' is set to the same value for RMB2 and RMB3.
• Case 6: This is a general case for case 4. In this case, the one that determines the puncturing scheme Parameter 'a' is set to integers 'p' and 'q', so it may be possible to set the different punctuation schemes. Parameter 'a' is set to 'p' for RMB2 and 'q' for RMB3.
• Case 7: This is another general case for case 5. In this case, the punctuation scheme determining parameter 'a' is set to an integer 'p' and the punctuation scheme determining parameter 'b' is set to an integer 'q' so that it may be possible to set the different punctuation schemes. The parameters 'a' and 'b' are set to the same value for RMB2 and RMB3.
• Case 8: This is another general case for case 6. In this case, the parameter 'a' determining the puncturing scheme for RMB2 and RMB3 is set to integers 'p' and 'r' respectively, and the parameter determining the puncturing scheme 'b' is set to integers 'q' and 's' respectively for RMB2 and RMB3, so it may be possible to set the different punctuation schemes. Parameters 'a' and 'b' are set to 'p' and 'q' for RMB2 and 'r' and 's' for RMB3.
• Case 9, Case 10: In these cases all possible parameters are changed. That is, the output symbol number can be set to any integer, and the punctuation scheme determining parameters 'a' and 'b' can also be set to any given integer.

In Table 1 can Case 1 and Case 2 are applied when rate matching on the convolutionally encoded symbols, and cases 3 to 8 can be applied when the rate adaptation to the turbo-coded Symbols takes place.

The Punctuation scheme can according to a change of the punctuation scheme determining parameter 'a' changed become. Table 2 below shows a change in the punctuation scheme according to a change the parameter 'a'. In Table 2 becomes Suppose that Nc = 10, Ni = 8, y = Nc - Ni = 10 - 8 = 2 and b = 1. The according to the puncturing scheme dotted symbols are represented by "x".

[Table 2]

From Table 2 it can be seen that it is possible to obtain different puncturing schemes by setting 'b' to '1' and setting 'a' to different values. It can be understood that the first symbol of the punctuation scheme is at the front when the value 'a' is increased. It is of course possible to obtain more different punctuation schemes by also changing the parameter 'b'. In addition, it is possible to prevent the puncturing of the first symbol by setting the parameter 'b' to '1' and using a value satisfying Equation 1 below for the parameter 'a'. To satisfy condition 3B, the parameter 'a' should be set to a value within a range of Equation 1. 1 <= a <| Nc / y | (1) where | Nc / y | the largest integer is less than or equal to Nc / y.

In Equation 1 is for Nc = 10 and y = 2 Nc / y = 10/2 = 5. Therefore, the first symbols become not punctured if 'a' is 1, 2, 3 and 4 has.

To satisfy requirement 5B, the tail bits should not be punctured. To do this, Nc should be set to a value that is determined by subtracting the number of tail bits from it. That is, when the input symbol number Nc is set to Nc-NT, where NT denotes the number of tail bits, the tail bits are not punctured, thus satisfying requirement 5B. In other words, the tail bits do not get into the rate matching block. The rate matching scheme thus takes into account only one frame size of Nc-NT. After puncturing or repeating by the rate matching block, the tail bits are sequentially concatenated with the output symbols of the rate matching block. The Tail bits are not processed and only appended to the end of the output symbols.

4. Rate matching algorithm by puncturing

7 shows a rate matching procedure by puncturing according to an embodiment of the present invention. This procedure is performed on the basis of a rate matching algorithm shown in Table 3 below. "So = {d1, d2, ..., dNc}" denotes the symbols input for a rate matching block, that is, the symbol inputted in a frame unit for a rate matching block, and consists of Nc symbols as a whole. A sliding parameter S (k) is an initial value used in the algorithm and is permanently set to '0' when a rate matching device of the invention is used in a downlink digital transmission system (ie, when rate matching is to be sent to the mobile station from the base station coded symbols). 'm' indicates the order of symbols entered for rate matching and has the order 1, 2, 3, ..., Nc. It can be seen from Table 3 that the parameters including the input symbol number, the output symbol number Ni, and the punctuation scheme determining parameters 'a' and 'b' can be changed. For example, the parameters may be changed as shown in Table 1. The input symbol number Nc may be determined according to the coding rate R as a value other than Ncs / 3. 7 corresponds to the case where the algorithm of Table 3 is applied to a downlink of a digital transmission system, ie, S (k) = 0.

[Table 3]

If the algorithm used in Table 3 will be the following Benefits provided.

First, it is possible, to variably puncture the coded symbols of a frame unit.

Secondly, it is possible, different puncturing schemes by adjusting the parameters Nc, To produce Ni, a and b.

Third, it is possible, the complexity and calculating time of each rate matching block by 1 / R. This is because when using a plurality of rate matching blocks is the number of points to be punctured by each rate matching block Symbols is reduced compared to the case where a rate matching block is used.

According to 7 be in step 701 initializes all kinds of parameters including the input symbol number Nc, the output symbol number Ni and the puncturing scheme determining parameters 'a' and 'b' for the rate matching process. When Nc and Ni are determined by parameter initialization, the number of symbols to be punctured becomes in step 702 determined by y = Nc - Ni. In step 703 an initial error 'e' between existing and desired puncturing ratios is calculated. The initial error is determined by e = b * Nc mod a * Nc.

Next will be in step 704 'm', which indicates the order of the input symbols, is set to '1'. After that, in the steps 705 to 709 the symbols from the beginning symbol to whether they should be punctured or not. When in step 707 it is determined that the calculated error value 'e' is less than or equal to '0', the corresponding symbol is punctured, and then in step 708 the error value is updated by e = e + a · Nc. Otherwise, if in step 707 it is determined that the calculated error value 'e' is greater than '0', no puncturing performed. The process of receiving the coded symbols in order, determining whether puncturing is performed on the received symbols, and performing the corresponding puncturing is performed repeatedly until in step 705 It is determined that all symbols in a frame have been completely received.

As shown by the above algorithm, the position of the first to puncturing or repeating symbols by the parameters (a, b) controlled (Initial_Offset_m = the position of the first to puncturing symbol). In the above algorithm, Initial_Offset_m is = 'm', if for the first Times 'e' <= 0. The table below shows an example for determining initial offset m. In the example below, it is assumed that bNc is less than aNc.

"Initial offset m = k = 4"

In the equations below, Ppnc denotes the period of puncturing or repeating in the above algorithm. Initial offset m = | bNc / ay | = | (b / a) · (Nc / y) | = | (b / a) · Ppnc Ppnc = | Nc / y | if Nc / y is an integer
= | Nc / y | +/- 1 if Nc / y is not an integer.

As Shown in the above equations, by controlling the parameters (a, b) the position of the first to be punctured or repeated Symbols are controlled.

To the For example, the value of Initial-Offset_m decreases while 'a' increases, while 'b' is constant remains. By elevating of 'a' is thus the position of the first symbol to be punctured / repeated closer to pushed first position. If 'a' is greater than selected by / Nc is, then initial offset m = 1, which means that the first Symbol will be punctured or repeated. Accordingly, the position of the first symbol to be punctured / repeated, by choosing a value for 'a' between 1 and Ppnc. If e.g. 'b' = 1 and 'a' = 2, the position of the first becomes puncturing / repeating symbol always be Ppnc / 2.

As for the parameter 'b', it controls the initial offset m together with 'a' and, as shown below, once the value for 'a' has been determined, the value of 'b' can be set to 1 <= ' b '<=' a 'are expressed. If 'a' remains constant, Initial_Offset_m will increase as 'b' increases, and will decrease as 'b' decreases. The puncturing / repeating positions can thus be controlled by manipulating the values of the parameters (a, b).

Although the value of 'b' may be arbitrary, it does not make sense to choose a value of 'b' greater than 'a', as shown below, because the initial value of 'e' will become cyclic as soon as the value of 'b 'becomes greater than' a '(ie, the value of' e 'repeats itself).
Let 'a' = 3;
the initial value of e = (2 * S (k) * y + Bnc) mod aNc;
e = bNc mod aNc, since S (k) = 0 in the downlink;
if b = 1 then e = Nc;
if b = 2 then e = 2Nc;
if b = 3 then e = 3Nc;
if b = 4 then e = Nc;
if b = 5 then e = 2Nc;
if b = 6 then e = 3Nc;

As shown in the above example changes the value of 'e', if the value of 'b' changes. But as soon as the value of 'b' gets bigger as 'a', repeats the Value of 'e' itself cyclically. It therefore does not make sense to assign 'b' a value greater than 'a'. In summary, the punctuation / repetition scheme can be controlled by manipulating the parameters (a, b).

B. designs the rate adjusting device by repeating

1. Execution of the Rate adaptation device by repeating (for a convolutional code)

9 shows the structure of a rate matching device of an inventive execution. This structure is used when the rate matching devices of 2 and 3 adjust convolutionally encoded symbols by repeating in the rate.

According to 9 encodes a convolutional encoder 210 input information bits Ik at a coding rate R = 1/3 and outputs encoded symbols C1k, C2k and C3k. The coded symbols C1k, C2k and C3k are separated into rate matching blocks 231 . 232 respectively. 233 to hand over. The first rate adjustment block 231 selectively repeats the encoded symbol C1k. At this point, the repetition process is performed on the basis of the symbol repetition number y = Ni-Nc which is determined by the input symbol number Nc, the output symbol number Ni, and the repetition pattern determining parameters 'a' and 'b'. The first rate adjustment block 231 may, for example, output the symbols '... 11 (11) 101 (00) 010 ...' (where (11) and (00) denote repeated symbols).

The second rate matching block selectively repeats the encoded symbol C2k. At this point, the repetition process is performed on the basis of the symbol repetition number y = Ni-Nc which is determined by the input symbol number Nc, the output symbol number Ni and the repetition pattern determining parameters 'a' and 'b'. The second rate adjustment block 232 can, for example, output the symbols '... (11) 01 (00) 1100 ...' (where (11) and (00) denote repeated symbols).

The third rate adjustment block 233 repeats the coded symbol C3k. At this point, the repetition process is based on the symbol repetition number y = Ni-Nc determined by the input symbol number Nc, the output symbol number Ni, and the repetition pattern determining parameters 'a' and 'b'. The third rate adjustment block 233 may, for example, output the symbols '... 0 (11) 1101 (11) ...' (where (11) indicates repeated symbols). The through the rate adjustment blocks 231 . 232 and 233 rate-matched coded symbols are passed through a multiplexer 240 multiplexed and handed over to a channel transmitter.

In 9 For example, the input symbol number Nc and the output symbol number Ni are determined for each rate matching block similarly as Nc = R × Ncs and Ni = R × Nis, respectively. It is determined that each rate matching block separately repeats the same number of the channel coded symbols on the assumption that the error rate of coded signals for each symbol in a frame is almost equal. That is, within one frame, an almost uniform repetition scheme is provided regardless of the various repetition bit numbers (y = Ni-Nc) determined according to the type of service. This is because it is possible to evenly repeat the entire symbols in a frame for the convolutional code.

According to the embodiment of the present invention, the symbols encoded by the convolutional encoder are separated in the same number and sent to the rate matching blocks 231 . 232 and 233 to hand over. The rate adjustment blocks 231 . 232 and 233 repeat each same number of input symbols. At this point, the repeat schema parameters can be determined to be the same or different. That is, for the rate matching blocks 231 . 232 and 233 the repetition patterns can be determined either equal or different.

2. Another embodiment of a Rate adaptation device by repetition (for a turbo code)

10 Fig. 10 shows the structure of a rate matching device by repeating according to another embodiment of the present invention. This structure is used when the rate matching devices of 2 and 3 to adjust the rates of turbo-coded symbols.

According to 10 encodes a turbo encoder 220 input information bits Ik at a coding rate of R = 1/3 and outputs encoded symbols C1k, C2k and C3k. Of the coded symbols, the information symbol C1k is separated to a first rate matching block 231 and the parity symbols (redundancy symbols) are separated to a second and third rate matching block 232 respectively. 233 to hand over. The turbo encoder 220 consists of a first component encoder 222 a second component encoder 224 and a nestler 226 , as in 6 shown. The component encoders 222 and 224 can use recursive systematic codes (RSC). The construction of the turbo coder 220 is well known to those skilled in the art. A detailed description is therefore omitted. The input X (t) into the turbo encoder 220 corresponds to the in 10 shown input information bits Ik. The outputs X (t), Y (t) and Y '(t) of the turbo coder 220 correspond to the in 10 shown encoded symbols C1 k, C2k and G3k. The first output of the turbo coder 220 returns the input information bits Ik as they are, so that in 10 the input information bits Ik are output as C1k.

The first rate adjustment block 231 repeats the coded symbols based on the following criteria. Since the coding rate is R = 1/3, the input symbol number Nc is determined as Nc = R × Ncs = Ncs / 3, which is 1/3 of the total number of input symbols (coded symbols). The output symbol number Ni is determined as Ni = Nis - (2R × Ncs), because the repetition is according to assumption 1D should be done. The repetition scheme determination parameters 'a' and 'b' may be set to given integers according to a desired repetition scheme. The integers can only be determined depending on the repetition scheme, and the parameters can typically be set to b = 1 and a = 2. A detailed description of a method for determining the integers for the repeat scheme determination parameters will be made with reference to the tables below. The first rate adjustment block 231 can, for example, output the symbols '... 1 (11) 101 (00) 11 ...' (where (11) and (00) indicate repeated symbols.

The second rate adjustment block 232 outputs the coded symbols Ck2 without repetition. The second rate adjustment block 232 However, under certain circumstances, such as severe repetition, it may repeat the coded symbols. Since the coding rate is R = 1/3, the input symbol number Nc is determined as Nc = R × Ncs = Ncs / 3, which is 1/3 of the total number of input symbols. The output symbol number Ni is determined to be Ni = R × Ncs, which is equal to the input symbol number, since according to the conditions 2D and 4D, the two types of parity symbols should not be repeated. The second rate adjustment block 232 can, for example, output the symbols '... 110111101 ...', in which there is no repetition.

The third rate adjustment block 233 outputs the coded symbols C3k without repetition. The third rate adjustment block 233 but under strict repetition can also repeat the coded symbols C3k. Since the coding rate is R = 1/3, the input symbol number Nc is determined as Nc = R × Ncs = Ncs / 3, which is 1/3 of the total number of input symbols. The output symbol number Ni is determined to be Ni = R × Ncs, which is equal to the input symbol number, since according to the conditions 2D and 4D, the two types of parity symbols should not be repeated. The repeat scheme determining parameters 'a' and 'b' may be set to given integers according to a desired repeat scheme. But if the blocks 232 and 233 do not use repetition, the parameters (a, b) are for the rate matching blocks 232 and 233 without meaning. The integers are only determined depending on the recurrence scheme, and the integers can typically be set to b = 1 and a = 2. A detailed description of a method for determining the integers for the recurrence scheme determination parameters will be made with reference to the tables below. The third rate adjustment block 233 can, for example, output the symbols '... 01011010 ...', which will not be repeated to have.

In 10 be through the turbo encoder 220 coded symbols separated in the same number and then to the rate matching blocks 231 . 232 and 233 to hand over. The first rate adjustment block 231 receives the information symbols from the coded symbols and repeats the received symbols according to a predetermined repetition scheme. The second and third rate matching block 232 and 233 receive the parity symbols from the encoded symbols and output the received symbols as they are without repetition.

3. Determination parameters to repeat

As described above the for the appropriate rate matching blocks used recurrence schemes be either the same or different. That is, that symbol repeating scheme used in the respective rate matching blocks and the number of repeated symbols can be variably determined. If the number Ni of the output from the respective rate matching blocks Icons set differently, the number of passes through the appropriate rate matching blocks repeated symbols differently be determined. Furthermore, the scheme of the concerned by the Rate matching blocks repeated symbols by changing the repeat scheme determination parameters 'a' and 'b' are either equal or different be determined. This means, Although it has a single structure, an inventive rate matching device Parameters, e.g. the input symbol number, the output symbol number, the number of symbols to be repeated and the repetition scheme determination parameters, determine differently.

table 4 below shows as an example different cases of parameters. The coding rate is assumed herein as R = 1/3. There are therefore provided three rate matching blocks, and the respective rate matching blocks receive the same separately Number of symbols, i.e., Nc = Ncs / 3 symbols. Here are the received Rate matching blocks the same number of symbols obtained by multiplying the number of coded symbols is determined with the coding rate. One should but note that the present invention also applies a case can be applied where the rate matching blocks are separated receive a different number of symbols, that is, a number of symbols smaller than the number by multiplying the number of coded symbols in a frame with the coding rate is determined, or a number of symbols that is larger as the number, by multiplying the number of coded symbols is determined in a frame with the coding rate. In the description At the bottom, RMB1, RMB2 and RMB3 denote the first to third, respectively Rate matching block.

[Table 4]

• Fall 1: Bei systematischer Wiederholung werden Informationssymbole wiederholt, aber die Paritätssymbole werden nicht wiederholt. Die Wiederholungsschema-Bestimmungsparameter werden auf a = 2 und b = 1 gesetzt.
• Fall 2: Bei systematischer Wiederholung werden Informationssymbole wiederholt. aber die Paritätssymbole werden nicht wiederholt. Die Wiederholungsschema-Bestimmungsparameter werden auf a = p und b = q gesetzt.
• Fall 1 und Fall 2 können nur angewandt werden, wenn die turbocodierten Informationssymbole wie in 10 gezeigt wiederholt werden.
• Fall 3: Sowohl die Informationssymbole als auch die Paritätssymbole werden wiederholt, und die Wiederholungsschemas werden für RMB1, RMB2 und RMB3 in gleicher Weise bestimmt. Die Zahl von wiederholten Symbolen ist für RMB1, RMB2 und RMB3 gleich.
• Fall 4: Sowohl die Informationssymbole als auch die Paritätssymbole werden wiederholt, und die Wiederholungsschemas werden für alle oder einige von RMB1, RMB2 und RMB3 unterschiedlich bestimmt. Die Zahl von wiederholten Symbolen ist für RMB2 und RMB3 gleich.

In Table 4, RMB1, RMB2 and RMB3 denote rate matching blocks, and p, q, r, s, t, w and x are given integers. NA (not present) indicates that the input symbols are output as they are without repetition, for which the parameters 'a' and 'b' can be set to any value. Here the parameters 'a' and 'b' are positive numbers. Further, the case is shown where the input symbols are repeated to perform the rate matching so that the number of input symbols is less than or equal to the number of output symbols (ie, Ncs <= Nis). It will definitely be referred to.

• Case 1: In systematic repetition, information symbols are repeated, but the parity symbols are repeated will not be repeated. The repetition scheme determination parameters are set to a = 2 and b = 1.
• Case 2: In systematic repetition, information symbols are repeated. but the parity symbols are not repeated. The repetition scheme determination parameters are set to a = p and b = q.
• Case 1 and Case 2 can only be applied if the turbo-coded information symbols as in 10 be repeated shown.
• Case 3: Both the information symbols and the parity symbols are repeated, and the repetition schemes are determined in the same way for RMB1, RMB2 and RMB3. The number of repeated symbols is the same for RMB1, RMB2 and RMB3.
• Case 4: Both the information symbols and the parity symbols are repeated, and the recurrence schemes are determined differently for all or some of RMB1, RMB2 and RMB3. The number of repeated symbols is the same for RMB2 and RMB3.

table 5 below shows the changes in recurrence schemes according to a change in parameter 'a'. In Table 5 becomes Suppose that Nc = 8, Ni = 10, y = Ni - Nc = 10 - 8 = 2 and b = 1. The according to the repeating scheme repeated symbols are represented by '()'.

[Table 5]

From Table 5 it can be seen that it is possible to obtain different repetition schemes by setting 'b' to '1' and setting 'a' to different values. Of course, it is possible to get more different recurrence schemes by also changing the parameter 'b'. In addition, it is possible to always repeat the first symbol by setting the parameter 'b' to 1 and using a value that satisfies Equation 2 below for the parameter 'a'. To satisfy the requirement of 3D, the parameter 'a' should be set to a value within a range of Equation 2. a> | Nc / y | (2) where | Nc / y | the largest integer is less than or equal to Nc / y.

In Equation 2 is for Nc = 8 and y = 2, Nc / y = 8/2 = 4. If 'a' is a Value greater than 4, the first symbols are repeated.

Around the requirement to fulfill 5D, the tail bits should be repeated. This should be Nc on one Value to be set by adding the number of tail bits is determined to Nc. That is, if the input symbol number Nc is set to Nc + Nt where NT is the number of tail bits become the tail bits for the information symbols always repeated, fulfilling condition 5D. In other words, the tail bits in the rate matching block will also repeat entered and considered for repetition.

4. Rate matching algorithm by repeating

11 FIG. 12 shows a rate matching procedure by repeating according to an embodiment of the present invention. FIG. This procedure is performed based on a rate matching algorithm shown in Table 6 below. In Table 6, "So = {d1, d2, ..., dNc}" denotes the symbols inputted for rate matching, that is, the symbols entered in a rate matching frame unit, and is composed of Nc symbols as a whole. A shift parameter S (k) is an initial value used in the algorithm and is permanently set to "0" when the rate matching device of the present invention is used in a downlink of a digital transmission system (ie, when the rate matching is performed on the Ba sisstation to the mobile station to be transmitted coded symbols). 'm' indicates the order of symbols entered for rate matching and has the order 1, 2, 3, ..., Nc. From Table 6, it can be seen that the parameters including the input symbol number Nc, the output symbol number Ni, and the repetition pattern determination parameters 'a' and 'b' can be changed. For example, the parameters may be changed as shown in Table 4. The input symbol number Nc may be determined according to the coding rate R as a value other than Ncs / 3. 11 corresponds to the case where the algorithm of Table 6 is applied to the downlink of a digital transmission system, ie, S (k) = 0.

[Table 6]

If The algorithm used by Table 6 will be the following Benefits provided.

First, it is possible, the coded symbols (or codeword symbols) of the frame unit are variable to repeat.

Secondly, it is possible, different repeat schemes by adjusting the parameters Nc, Ni, a and b to produce.

Third, it is possible, the complexity and calculating time of each rate matching block by 1 / R. This is because when using a plurality of rate matching blocks will be the number of repeats to be repeated by each rate matching block Symbols are reduced compared to the case where a rate matching block is used. For example, the number of symbols passing through each rate matching block can be repeated compared to In the case where a rate matching block is used, the coding rate R be reduced.

According to 11 be in step 1101 all kinds of parameters including the input symbol number Nc, the output symbol number Ni, and the repetition pattern determination parameters 'a' and 'b', for initializes the rate matching process. When Nc and Ni are determined by parameter initialization, in step 1102 the number of symbols to be repeated is determined by y = Ni - Nc. In step 1103 an initial error value 'e' is calculated between the current and desired repetition ratios. The initial error value is determined by e = b * Nc mod a * Nc.

Next will be in step 1104 'm', which indicates the order of the input symbols, is set to '1' (m = 1). After that, in the steps 1105 to 1109 examines the symbols from the beginning symbol to see whether they are to be repeated or not. When in step 1107 it is determined that the calculated error value 'e' is less than or equal to '0', the corresponding symbol is repeated, and in step 1108 then the error value is updated by e = e + a * Nc. Otherwise, if in step 1107 it is determined that the calculated error value 'e' is greater than '0', no repetition is performed. The process of receiving the coded symbols in order, determining whether to repeat the received symbols, and respectively performing the repetition is performed repeatedly until in step 1105 It is determined that all symbols in a frame have been completely received. During the repetition process will be in step 1106 the error value is updated by e = e - a · y.

As described above, the data transmission system according to the invention, the rate adaptation both on channel-coded symbols with a non-systematic code as well as channel-coded symbols with a systematic code by means of a single structure. The data transmission system, which supports both non-systematic and systematic codes Therefore, with a non-systematic code channel coded symbols or channel coded symbols selectively with a systematic code in the rate adjust, reducing the efficiency of data transmission increased and the system performance is improved.

The The present invention has the following advantages.

First, it is possible, by adjusting the parameters of the rate matching blocks Free punctuation / repetition schemes, and all prerequisites taken into account in the rate matching of the turbo-coded symbols should be able to can be met by simply setting the parameters.

Secondly, it is possible, all rate adjustment blocks according to the coding rate R using the same algorithm, and the rate adjustment blocks have a simple structure.

Third, a system that uses both convolutional codes and turbo codes, can convolutional codes as well as turbo codes by simple Support setting various initial parameters by a single rate matching device instead of different ones Rate matching devices is used.

Fourth, it is not necessary to change the rate matching blocks according to a convolutional code or a turbo code to implement differently.

Fifth, by setting the number of input symbols to a value that is determined by adding the number of its tail bits, so that the tail bits are repeated is the novel rate matching device useful, if a SOVA decoder is used or if the performance is due not repeating the tail bits would be degraded. By Set the number of input symbols to a value determined by Adding the number of tail bits to the number of non-tail bits is determined so that the tail bits are repeated is the novel rate adjustment device useful when using a SOVA decoder is used or if the performance as a result of not repeating the tail bits would be degraded. By setting the number of input symbols to a value by adding the number their tail bits is determined so that the tail bits are repeated the novel rate matching device is useful, if a SOVA decoder is used or if the performance is due to the Not repeating the tail bits would be worsened.

Sixth, by setting the punctuation pattern determining parameter 'b' to '1' and Set the parameter 'a' to a value within a specific area, can be prevented that the first Symbol is dotted in a frame. Furthermore, it is possible that repeat the first symbol in a frame by setting the repetition pattern determination parameter 'b' to '1' and the parameter 'a' to a value within a specific area is set.

Claims (57)

Rate adaptation device in a data transmission system, marked by: a channel coder, the coded bits generated; a control unit which determines whether a channel coding scheme, which is used to generate the coded bits, a systematic one Code or non-systematic code; a multiplexer, which multiplexes the coded bits, and if more systematic Code is used, the multiplexer encoded the bits into one divides systematic stream and at least one parity stream, and if one non-systematic power is used, the multiplexer one outputs coded current; and Rate matching functions that puncturing a portion of the coded bits, being, if more systematic Code is used, at least one rate matching function, the corresponds to a number of parity streams, each corresponding parity streams receives and dotted a part of corresponding parity streams, and the systematic stream is bypassed, if not more systematic Code is used, a rate matching function forms part of the dotted coded bits, and each of the rate matching functions the coded bits punctuate with rate matching parameters that according to everyone the rate matching functions are determined individually. The device of claim 1, wherein, if more systematic Code is used, every parity stream according to rate adjustment parameters dotted, which is every parity stream correspond. The device of claim 1, wherein the rate matching parameters for every Rate matching function according to a first parameter, which is a position of a first in a frame determined to puncturing bits, and a second parameter determined be a period of bits to be punctured in a frame certainly. The device of claim 1, wherein the numbers are dotted Bits in each of the rate matching functions are equal to each other. The device of claim 1, wherein the numbers are dotted Bits in each of the rate matching functions are different from each other are. The device of claim 1, wherein each rate matching function (A) a number "y" of the one to be punctured Symbols by receiving a number Nc of input symbols and a number Ni of output symbols is determined; (b) an initial one Error value "e" calculated, the one Difference value between a current puncturing ratio and a desired one Punktierungsverhältnis displays; (c) the error value for each of the input bits updated; (d) puncturing the corresponding input bit, if the error value is less than or equal to "0"; and (e) repeats the operations in (c) and (d), until the number of counted Bits is greater than "Nc". Apparatus according to claim 6, wherein the initial one Error value "e" by means of a formula [{(2 × S (k) × y) + (b × Nc)} mod {a × Nc}] is calculated and each of the rate matching functions one sentence of the coded bits. Apparatus according to claim 7, wherein S (k) is a displacement parameter indicates that in the downlink is set to "0", "a" indicates a parameter that contains a position of a bit to be punctured first in a frame, and "b" a parameter indicating a period of puncturing in a bitstream Bits determined. Rate matching apparatus in a data transmission system, characterized by: a channel encoder that generates coded bits; a controller that determines if a rate matching scheme is a repeat or punctuation scheme; a multiplexer that multiplexes the coded bits, wherein when puncturing is used, the multiplexer divides the coded bits into a systematic stream and at least one parity stream, and when repetition is used, the multiplexer outputs a coded stream; and rate matching functions that puncture a portion of the coded bits, wherein when puncturing is performed, at least one rate matching function corresponding to a number of parity streams each receives respective parity streams and a portion of corresponding Pa and the systematic stream is bypassed when repetition is performed, a rate matching function repeats a portion of the coded bits, and each of the rate matching functions punctures the coded bits with rate matching parameters individually determined according to each of the rate matching functions. The device of claim 9, wherein when puncturing carried out is, every parity stream according to rate adjustment parameters dotted, which is every parity stream correspond. Apparatus according to claim 9, wherein the rate matching parameters for every Rate matching function according to a first parameter, which is a position of a first in a frame determined to puncturing bits, and a second parameter determined be a period of bits to be punctured in a frame certainly. Apparatus according to claim 9, wherein the numbers punctured bits in each of the rate matching functions are the same. Apparatus according to claim 9, wherein the numbers dotted bits in each of the rate matching functions from each other. Apparatus according to claim 9, wherein each rate matching function (a) a number "y" of the points to be punctured Symbols by receiving a number Nc of input symbols and a number Ni of output symbols is determined; (b) an initial one Error value "e" calculated, the one Difference value between a current puncturing ratio and a desired one Punktierungsverhältnis displays; (c) the error value for each of the input bits is updated; (d) the corresponding input bit punctured if the error value is less than or equal to "0"; and (e) the operations in (c) and (d) repeatedly performs until the number of counted Bits is greater than "Nc". The device of claim 14, wherein the initial one Error value "e", which is a difference value between a current one Punktierungsverhältnis and a desired one Punktierungsverhältnis is calculated by a formula [{(2 × S (k) × y) + (b × Nc)} mod {a × Nc}] and each of the rate matching functions encodes a set of the coded ones Receives bits. Apparatus according to claim 15, wherein S (k) is a displacement parameter indicates that in the downlink is set to "0", "a" indicates a parameter that contains a position of a bit to be punctured first in a frame, and "b" a parameter indicating a period of the bitstreams to be punctured Bits determined. Data transfer system with a channel coder, the input information bits with predetermined Coding rate encoded and a coded bitstream for the entered Outputs information bits to coded bits of the input information bits and a rate matching function, rate matching a number of bits from the coded bits to a number of bits output from the rate matching function be adapted to a given bit number of the transmission channel: one Multiplexer that multiplexes coded bits and multiplexes bits outputs; a rate matching function, the rate adaptation of a predetermined number of the multiplexed bits according to rate matching parameters performs; and a memory that stores rate matching parameters, which according to one Type of bitstream to be determined where the rate matching function The Rate Matching parameter receives sequentially and rate adaptation multiplexed bits according to the received ones Rate matching parameters. The system of claim 17, wherein, when in the rate matching function the number of bits entered is greater as the number of bits output, the rate matching function the predetermined number of the multiplexed bits according to respective rate matching parameters punctured. The system of claim 17, wherein, when in the rate matching function the number of bits output is greater as the number of bits indicated, the rate matching function predetermined number of multiplexed bits according to corresponding rate matching parameters repeated. The system of claim 17, wherein, if the channel encoder is a turbo encoder, the rate adapt performs rate matching of the multiplexed bits with rate matching parameters corresponding to the type of bitstream, and the bitstream is a systematic bitstream or parity bitstreams. The system of claim 17, wherein when the channel coder a convolutional encoder is the rate matching function of the multiplexed rate adaptation Performs bits with the same rate matching parameters. The system of claim 17, wherein the rate matching parameters the number of input multiplexed bits and the number of outputs Include bits from the rate matching function. The system of claim 17, wherein the parameters of Another position parameter, determining a position of an input multiplexed bit, The first thing to do is rate matching of the multiplexed bits is, and a period parameter containing a period of entered by the multiplexed bits rate adaptation to be subjected determined by multiplexed bits. The system of claim 17, wherein the number of bits the rate adjustment are to be subjected to the number of entered multiplexed bits and the number of bits output. Data transmission system for rate adaptation of turbo codes, comprising: a turbo encoder ( 220 ) encoding input information bits at a predetermined coding rate and outputting a systematic bitstream, a first parity bitstream and a second parity bitstream for the input information bits to produce coded bits of the input bitstreams; and a rate matching device that punctures a number of bits from the coded bits, characterized in that the rate matching device comprises: a first rate matching device ( 231 ) receiving the first parity bitstream and puncturing a predetermined number of first parity bits in the first parity bitstream; and a second rate matching device ( 232 ) puncturing a predetermined number of second parity bits in the second parity bit stream; wherein each of the rate matching devices ( 231 . 232 ) punctures a predetermined number of parity bits in a corresponding parity bit stream with a corresponding puncturing scheme, and the puncturing scheme of each of the rate matching means can be controlled according to the puncturing scheme parameters, respectively. Datenüberragungssystem according to claim 25, wherein the system further comprises a multiplexer contains which collects the systematic bitstream and the rate-matched parity bitstreams. Data transfer system according to claim 25, wherein puncturing determining parameters comprise a first parameter (a), a position of one to be punctured first in a frame Bits, and a second parameter (b) is one period determines the bits to be punctured in a frame. A data transmission system according to claim 27, wherein each of the rate matching devices ( 231 . 232 ) punctures a predetermined number of parity bits in a respective parity bit stream in accordance with the corresponding puncturing scheme and the respective number of bits supplied by the rate matching means (12). 231 . 232 ). A data transmission system according to claim 27, wherein the respective numbers of bits supplied by said rate matching means ( 231 . 232 ) are equal to each other. A data transmission system according to claim 27, wherein the respective numbers of bits supplied by said rate matching means ( 231 . 232 ) are different from each other. Data transmission system according to claim 27, wherein the number of bits transmitted by at least one rate matching device ( 231 . 232 ) is different from the number of bits used by at least one other rate matching device ( 231 . 232 ). A data transmission system according to claim 27, wherein each of the rate matching devices ( 231 . 232 ) has the same corresponding punctuation scheme parameters. A data transmission system according to claim 27, wherein each of the rate matching devices ( 231 . 232 ) has corresponding punctuation scheme parameters and the first parameters of each of the corresponding punctuation scheme parameters are equal to each other. A data transmission system according to claim 27, wherein each of the rate matching devices ( 231 . 232 ) has corresponding punctuation scheme parameters and the first parameters of each of the corresponding punctuation scheme parameters are different from each other. A data transmission system according to claim 27, wherein each of the rate matching devices ( 231 . 232 ) has corresponding punctuation scheme parameters and the second parameters of each of the corresponding punctuation scheme parameters are equal to each other. A data transmission system according to claim 27, wherein each of the rate matching devices ( 231 . 232 ) has corresponding punctuation scheme parameters and the second parameters of each of the corresponding punctuation scheme parameters are different from each other. Data transmission system according to claim 25, wherein the punctuation schemes of the rate matching devices ( 231 . 232 ) are equal to each other. Data transmission system according to claim 25, wherein the punctuation schemes of the rate matching devices ( 231 . 232 ) are different from each other. Data transfer system according to claim 25, wherein each of said rate adjusting means (A) a number "y" of the one to be punctured Symbols by receiving a number Nc of input symbols and a number Ni of output symbols is determined; (b) an initial one Error value "e" calculated, the one Difference value between a current puncturing ratio and a desired one Puncturing ratio displays; (c) the error value for each of the input bits updated; (d) puncturing the corresponding input bit, if the error value is less than or equal to "0", and (e) the processes in (c) and (d) repeatedly performs until the number of counted Bits is greater than "Nc". Data transfer system according to claim 39, wherein the initial one Error value "e" by means of a formula [{(2 × S (k) × y) + (b × Nc)} mod {a × Nc}] is calculated and each of the rate matching means one sentence of the coded bits. Data transfer system according to claim 40, wherein S (k) indicates a shift parameter set to "0" in the downlink, "a" a parameter indicating a position of one to be punctured first in a frame Bits and "b" determines a parameter indicates that a period of bits to be punctured in a bitstream certainly. Data transmission system for rate adaptation of turbo codes, comprising: a turbo encoder ( 220 ) which encodes input information bits at a predetermined coding rate and outputs a systematic bit stream, a first parity bit stream and a second parity bit stream for the input information bits to produce coded bits of the input information bits; and rate matching means for rate matching a number of bits from the coded bits, characterized in that the rate matching means comprises: a first rate matching means ( 231 ), which receives the systematic bit stream and repeats a predetermined number of systematic bits in the systematic bit stream, a second rate adapter ( 232 ), which receives the first parity bit stream and repeats a predetermined number of first parity bits in the first parity bit stream, and a third rate match ( 233 ) which repeats a predetermined number of second parity bits in the second parity bit stream; wherein at least one of the rate matching devices ( 231 . 232 . 233 ) is arranged to repeat respective systematic or parity bits in accordance with a corresponding repetition scheme, and the repetition scheme of each of the rate matching means may be controlled according to the repetition scheme parameters, respectively. Data transmission system for rate adaptation of turbo codes, comprising: a turbo encoder ( 220 ) encoding the input information bits at a predetermined coding rate and output a systematic bit stream, a first parity bit stream and a second parity bit stream for the input information bits to produce coded bits of the input information bits; and rate matching means for rate matching the coded bits, characterized in that the rate matching means comprises: a first rate matching means ( 231 ) receiving the systematic bit stream and outputting the systematic bit stream, a second rate matching device ( 232 ) which receives the first parity bitstream and punctures a predetermined number of first parity bits in the first parity bitstream, and third rate matching means (16) 233 ) receiving the second parity bitstream and puncturing a predetermined number of second parity bits in the second parity bitstream; wherein each of the rate matching devices ( 232 . 233 ) puncturing a predetermined number of parity bits in a respective parity bit stream with a corresponding puncturing scheme, and the puncturing scheme of each of the rate matching means can be controlled according to the puncturing scheme parameters, respectively. Data transmission system according to claim 43, wherein the rate matching device ( 231 ) puncturing a predetermined number of systematic bits in the systematic bit stream with a corresponding puncturing scheme, and controlling the puncturing scheme of the rate matching means according to puncturing scheme parameters. Data transfer system according to claim 44, wherein the system further comprises a multiplexer contains the rate-matched bitstreams collects. Data transfer system according to claim 44, wherein the puncturing scheme parameters are a first parameter (a), a position of one to be punctured first in a frame Bits, and a second parameter (b) is one period determines the bits to be punctured in a frame. A data transmission system according to claim 46, wherein each of the rate matching devices ( 231 . 232 . 233 ) punctures a predetermined number of bits in a respective bit stream according to the corresponding puncturing scheme and the number of bits output from the rate matching means. A data transmission system according to claim 46, wherein the numbers of bits output by the rate matching means ( 231 . 232 . 233 ) are equal to each other. A data transmission system according to claim 46, wherein the numbers of bits output by the rate matching means ( 231 . 232 . 233 ) are different from each other. A data transmission system according to claim 46, wherein the number of bits output by at least one rate matching means is different from the number of bits used by at least one other rate matching means (16). 231 . 232 . 233 ). Data transfer system according to claim 47, wherein each of the rate matching means is the same corresponding puncturing scheme parameter has. A data transmission system according to claim 47, wherein each of the rate matching devices ( 231 . 232 . 233 ) has corresponding punctuation scheme parameters and the first parameters of each of the corresponding punctuation scheme parameters are equal to each other. A data transmission system according to claim 47, wherein each of said rate matching means ( 231 . 232 . 233 ) has corresponding punctuation scheme parameters and the first parameters of each of the corresponding punctuation scheme parameters are different from each other. Data transfer system according to claim 47, wherein each of the rate matching means corresponds Puncturing scheme parameter and the second parameters of each of the corresponding puncturing scheme parameters are equal to each other. A data transmission system according to claim 47, wherein each of the rate matching devices ( 231 . 232 . 233 ) has corresponding puncturing scheme parameters and the second parameters correspond to each of the the punctuation scheme parameters are different from each other. Data transfer system according to claim 47, wherein the puncturing schemes of the rate matching means are equal to each other. The data transmission system according to claim 47, wherein the punctuation schemes of the rate matching devices ( 231 . 232 . 233 ) are different from each other.