KR100797456B1 - Data Rate Matching Method - Google Patents

Abstract

The present invention relates to next generation mobile communication, and more particularly, to a data rate matching method of next generation dual communication for supporting a variable or variable data rate for an information bit string having an arbitrary data rate. In the process of interleaving such that the information bit strings having no normal data rate are mapped to the physical layer according to the present invention, information bit strings having different bit rates are converted into different codes having "1 / inverse of coding rate". Performing channel coding, symbol repetition of the channel coded information bit string when the channel coded information bit string is smaller than a desired length of the interleaver, and puncturing the channel coded information bit string when the channel coded information bit string is large according to the symbol repetition factor. And matching the length of the channel coded information bit stream to the interleaving length. Therefore, in addition to the time diversity gain due to symbol repetition, a coding gain may be obtained due to the actual reduction of the code rate, thereby lowering the required transmit power.

Variable data rates, variable data rates, systematic bits, parity bits,

Description

Data Rate Matching Method

1 illustrates a radio architecture (RC) in the forward link to support variable and multivariate data rates according to the prior art;

2 illustrates a transmission chain of variable data rate and multivariate data rate for a turbo code according to the present invention.

3 illustrates a transmission chain of variable data rate and multivariate data rate for a convolutional code in accordance with the present invention.

The present invention relates to next generation mobile communication, and more particularly, to a data rate matching method of next generation dual communication for supporting a variable or variable data rate for an information bit string having an arbitrary data rate.

Generally, the next generation mobile communication; data transmission mode (The 3 ^rd Generation Partnership Part 2 hereinafter abbreviated as 3GPP2) has in addition to the regular data rate mode, there are two transfer modes called the variable data rate mode and the multivariate data rate mode.

The normal data rate mode refers to a transmission mode that operates on a fixed chain called a radio configuration (hereinafter referred to as RC).

RC can be thought of as a type of transmission chain in which the length of the information data, the length of the channel interleaver, and the length of the output string of the channel encoder according to the code rate of the channel code are matched. At this time, there is a formal rule between the length of the channel interleaver, the coding rate of the channel encoder, and the length of the Walsh code of the channel.

In other words, when the chip rate to be used is determined, the number of chips into one modulation symbol is determined according to the length of the channel interleaver, which can be defined as a spreading factor. The length of the Walsh code that can code multiplex the channels is determined.

At this time, the number of available Walsh codes is in proportion to the length of the Walsh code. Accordingly, the number of channels that can be accommodated in the multiplexing chain varies according to the Walsh code.

Consider the length after channel coding for input information bits having the same length. In this case, the ability of an error correction code to correct an error that may occur in a channel is stronger as the coding rate of the channel encoder is lower.

In other words, the lower the coding rate of the channel encoder is, the better the error correction capability is, and thus the lower the transmission power can be used. However, when a channel coder having a low code rate is used, the length of the output string of the channel encoder is increased, thereby increasing the length of the channel interleaver.

This results in an increase in the modulation symbol rate, thereby reducing the number of chips entering one modulation symbol at a fixed chip rate.

This results in a reduction in the number of useful Walsh codes.

On the contrary, if the channel coding technique of high coding rate is used for the input string of the same length channel encoder, the error correction capability is reduced but the length of the output string of the channel encoder is shortened. Length of channel interleaver may be used, resulting in an increase in the number of Walsh codes that may be useful.

As described herein, it can be seen that there is a trade-off relationship between the coding rate of the channel encoder and the Walsh code space. Considering this trade-off relationship, RC can be thought of as a formalization of a transmission chain that can be used to secure Walsh code space, or a transmission chain that can be used when lower transmission power is required. .

Currently, 3GPP2 defines several RCs for 1X systems using a chip rate of 1.2888Mcps and some formal RCs for 3X systems using a chip rate of 3.6864Mcps.                         

One thing to note here is that the spreading factor has an exponential value of 2, so that the input data rate and the length of the interleaver defined in each RC are also increased by twice.

Before the traffic channel is established between the terminal and the base station, the terminal and the base station determine the RC to be used through the negotiation process and the length of the spreading factor on each RC, that is, the channel interleaver, and the communication process proceeds according to the chain.

The modes of using a transmission chain other than the transmission chain defined on the RC are the variable data rate mode and the multivariate data rate mode.

Variable data rate mode is a transmission method that can support any data rate in addition to the standard data rate supported by each RC. This variable data rate was introduced to support Adaptive Multi-Rate (hereinafter abbreviated as AMR) codec, which is one of the 3GPP speech codecs on the physical layer of 3GPP2.

That is, in the case of AMR, data bits that do not match the standard data rate supported by each RC of 3GPP2 can be dropped during the frame period for 20ms.

Next, there is what is called a multivariate data rate mode. The purpose of this mode is to:

In the 3GPP2 system, the base station schedules transmission on the forward auxiliary channel. At this time, the base station allocates a fixed data rate to the terminal through a message.                         

However, the channel situation between the base station and the specific terminal may change during that particular time, and the system load of the base station may also change.

For example, as the terminal moves away from the base station, the situation of the channel may change, and the channel environment may deteriorate, and thus, the base station may not have sufficient transmit power for transmitting at a current data rate to a specific terminal.

To solve this problem, the base station may stop transmitting to the auxiliary channel during this time. This solution, however, causes delays in data transmission and can also cause unnecessary waste of useful transmission power and Walsh codes.

Another solution is to consider how the base station reschedules the transmission data rate after a certain amount of time. However, this also causes problems such as time delay and waste of Walsh code.

This situation is not unique to the forward link. Likewise, in the reverse link, the channel state between the terminal and the base station may change according to the movement of the terminal, and thus, there may be a lack of transmission power to maintain proper quality.

Thus, a mode called multivariate data rate has been used to solve this situation. In this mode, the transmission rate is changed in units of frames depending on the situation. That is, if it is determined that the channel environment is worsened, the base station lowers the transmission rate of the auxiliary channel. If it is determined that the channel environment is recovered again, it can be considered that the transmission mode is performed again at the previous transmission rate. Using this multivariate data rate mode, the base station can use the power available at the base station without frequent rescheduling.

In order to support the variable data rate mode and the variable data rate mode, each RC of 3GPP2 currently configures a transmission chain using the following method.

As described above, the length of the channel interleaver used in each RC is determined according to the spreading factor. At this time, since the spreading factor has a value increasing in the form of an exponential power of 2, the length of the interleaver determined for a certain spreading factor and the length of the interleaver determined for the spreading factor one step lower than that is exactly 1: 2. Will have a relationship with

In this case, let the large spreading factor A be the small spreading factor B. Then, in each RC, a 1: 1 mapping relationship is established between the spreading factor and the information bit string input to the channel encoder. If the length of the information bit string input to the channel encoder for the spreading factor A is I _A , and the length of the information bit string input to the channel encoder for the spreading factor B is I _B , then I _B = 2 * I _A Will have a relationship with Also, if the length of each channel interleaver to be used is N _A and N _B , there will be a relationship of N _B = 2 * N _A.

In this case, consider a variable data rate mode in which I satisfying a relationship of "I _A <I <I _B " rather than the length of data to be normalized becomes the length of an input string of a channel encoder.

Assuming that the coding rate of the channel incubator used in the current RC is 1 / n, the output of "n * I" will be sent for the input of I.

At this time, the relationship of "N _A <n * I <N _B " is satisfied. Therefore, some work is required to match the length "n * I" of the output string of the channel encoder to the interleaver length.

The current approach taken by 3GPP2 is to match the length L (= n * I) of the output string of the channel encoder to the interleaver of N = N _B. Accordingly, bit repetition of "N _B -n * I" is performed. The method of doing this performs a uniform iteration process of the following form.

번째 입력 비트열의 코드 심볼로부터 추정 가능하다.That is, the kth output symbol of the iteration block is for index k that increases from 0 to N-1.

It can be estimated from the code symbols of the first input bit string.

Next, a method for supporting the multivariate data rate mode will be described.

In multivariate data rate mode, the maximum data rate, one level lower data rate, and two levels lower data rate that can be supported during the initial negotiation process are defined as a set of transmission data rates.

Therefore, in the multivariate data rate mode for the current auxiliary channel, the data rate of the auxiliary channel is adjusted between the support rate up to two levels below. At this time, if the forward channel is considered, the terminal must determine the rate variation through the blind rate detection. Accordingly, if the data transmission rate is too large, a problem arises that increases the complexity of the terminal.

The length of the channel interleaver and the Walsh code used at the maximum data rate should not be changed. In other words, the interleaver and Walsh codes specified for the maximum transmission rate currently used are used as they are.

Therefore, if the data rate is lowered to 1/2 of the maximum rate, the symbol repetition is doubled to match the length of the interleaver to be used in the channel and the length of the output string of the channel encoder.

Similarly, if the data rate is lowered to 1/4 of the maximum data rate, four times symbol repetition is performed to match the length of the interleaver to be used in the channel and the length of the output string of the channel encoder.

The previous example is an example of a variable data rate in the secondary channel of the forward channel.

Multivariate data rate modes for variable data rates may be supported in the secondary channel.

In this case, however, the definitions of the variable data rate and the multivariate data rate mode are blurred. That is, even if the maximum data rate in the multivariate data rate mode is the normal rate set on the current RC, and the data rate one step lower is the data rate set on the current RC, it is actually one step lower. The data rate can also be seen as a variable data rate that does not use a chain defined on the current RC. This is because the length of the interleaver, that is, the spreading factor, in the multivariate data rate mode is fixed at that of the maximum transmission rate.

As an example, consider the RC4 of the current forward channel. At this time, a half rate turbo code or a convolution code is used. Assume that the maximum data rate that can be used in the multivariate data rate mode is 76.8 kbps. At this time, the length of the interleaver used in the forward RC4 is fixed to 3072. Consider a multivariate data rate method in this mode. Assume that the available data rate is used at an appropriate value among the set of {19.2 kbps, 38.4 kbps, 76.8 kbps}. The data rates of 38.4 kbps and 19.2 kbps are clearly defined at RC4, but the problem is that there is no chain connecting these rates at 3072, the length of the current interleaver on this RC. As a result, this rate can also be viewed as a variable data rate.

As a result, if the length of the interleaver is fixed to N even in the multivariate data rate mode, and the data rate being transmitted is not defined on the current RC, the output of the channel encoder through the uniform iteration algorithm described above You can match the length of a row to the length of a given interleaver.                         

That is, the variable data rate and the variable data rate should be considered in the same context, and a method for supporting this on each RC can be described through the chain as shown in FIG. 1.

1 is a diagram illustrating a radio architecture (RC) in a forward link for supporting a variable data rate and a multivariate data rate according to the prior art.

First, in the CRC bit block S10, a CRC bit having an arbitrary length is added to the channel input bit for error detection. In this case, the method of determining the length of the CRC in the new RC may be signaled in the initial negotiation step between the transmitter and the receiver.

After the CRC bit is added to the channel input bit, tail bits or reserved bits are added to this bit string. In this case, 8 tail bits are added in the case of using a convolutional code, and 6 tail bits and 2 reserved bits are added in the case of the turbo code (S11).

Turbo codes will always have a code rate of 1/5 rate. In order to use this RC, the following things must be promised during the initial negotiation phase.

First, the information data rate and spreading factor to be used, that is, the length N of the channel interleaver must be promised. The length of the CRC bit to be used must be promised. It also promises to use turbo codes or convolutional codes.

In this case, when using a turbo code, only a 1/5 rate channel encoder is used. On the other hand, in case of using a convolutional code, the inverse n of the coding rate of the convolutional code may be determined as a ratio between the length I of the input string to the channel encoder and the length N of the channel interleaver as described above. May be signaled in an initial negotiation phase (S12)

However, the conventional method for supporting the variable data rate mode and the multivariate data rate mode is that the concept of the RC is ambiguous in the concept of the variable data rate mode and the multivariate data rate mode. As mentioned above, the concept of RC can be thought of as a formal rule that defines the relationship between the coding rate and the spreading factor of a channel encoder. However, in the variable data rate mode or the multivariate data rate mode, this relationship does not hold as a formal rule in the RC. That is, when the desired spreading factor is determined and the length of the channel interleaver to be used is determined, and the length of the input string to the channel encoder is determined, the channel encoding is performed according to the coding rate currently used on the RC. In addition, a symbol repetition process is performed for code symbols to match the length of a predetermined interleaver.

In other words, in such a variable data rate mode and a multivariate data rate mode, the formal relationship on RC is lost between the coding rate and the spreading factor of the channel encoder.

In the current method, a matching process with a specific spreading factor is performed by repeating the code symbol as described above.

)로 정의하고, 각각의 RC상에서 정의하고 있는 코드 레이트를 1/n이라고 가정하는 경우에 가변 데이터 레이트 모드나 다변 데이터 레이트 모드에 있어 RC상에서 정규화되어 있는 체인을 제외한 다른 모든 임의의 데이터 레이트 상에서는 다음의 관계를 가지게 된다는 것이다.One thing to note here is that the effective code rate is the ratio of the length N of the interleaver to the length I of the information bit string input to the channel encoder.

Assuming that the code rate defined on each RC is 1 / n, then on any other data rate except the chain normalized on the RC in either the variable data rate mode or the multivariate data rate mode: Will have a relationship.

In other words, in the variable data rate mode or the multivariate data rate mode, it means that the effective code rate is lowered, and that it is used repeatedly for the matching process with the interleaver. This means that the effective code rate is reduced, but the actual code rate is still 1 / n specified on RC.

Therefore, it can be considered that both the current variable data rate mode and the multivariate data rate mode have a problem of losing a gain that can be obtained on transmission power through an increase in coding gain due to a decrease in coding rate.

That is, if the length of the interleaver is set to a certain value, and the transmission rate of the data to be transmitted currently is determined, the coding rate for obtaining the maximum coding gain is selected according to the relationship between the two. There is a need to construct a new transmission chain that performs rate matching puncturing or data matching repetition to match the length of the channel interleaver.

 Accordingly, an object of the present invention has been devised in view of the above-mentioned problems of the prior art, and the length of the spreading factor or channel interleaver specified during negotiation not only in the normal data rate mode but also in the variable data rate mode and the multivariate data rate mode is determined. It is an object of the present invention to provide a data rate matching method for next-generation mobile communication that can make an effective coding rate as low as possible under a prescribed channel environment.

delete

delete

delete

delete

delete

delete

delete

Hereinafter, a configuration and an operation according to an exemplary embodiment of the present invention will be described with reference to the accompanying drawings.
According to an aspect of the present invention, there is provided a rate matching method for a bit string having a data rate different from a normal radio configuration (RC), the method comprising: performing turbo coding on an input bit string and performing the turbo coded bit string When performing bit repetition to match the interleaver size, the priority of the bit repetition is changed according to each bit component, and the bit repetition is performed first with respect to the systematic bits.

First, the first method for supporting variable data rate and multivariate data rate in each RC is as follows.

If the length of the information bit string input to the channel encoder defined in each RC is I, and the length of the channel interleaver defined for this is N, in the normal data rate mode, the ratio of 1: 1 between I and N is normal. The chain is formed.

However, it can be considered that the above 1: 1 chain condition is not satisfied in the variable data rate or the multivariate data rate mode. Therefore, a new chain must be made to connect the length N of the channel interleaver in the current RC with the length I of the information bit string input to the channel encoder. The method used for this purpose is proposed as a different embodiment of the turbo code and the convolution code.                     

First embodiment

2 is a diagram illustrating a transmission chain of a variable data rate and a multivariate data rate for a turbo code according to the present invention.

Referring to FIG. 2, when an information bit string (having length I) consisting of channel bits, CRC error bits, and tail bits is input to the channel encoder, a chain other than a chain of normal data rates determined on each RC must be formed. In the case of using a 1/5 rate turbo encoder as a channel encoder (S20). The output string length L of the channel encoder uses rate matching puncturing (5 * I> N) or rate matching repetition (5 * I <N) for matching between "5 * I" and the length N of the channel interleaver. (S21).

The rate matching puncturing or repetition method uses a method of uniformly performing an information bit string input to a channel encoder.

The above chain is possible because the 1/2, 1/3, and 1/4 rate turbo codes used in each RC are punctured from the 1/5 rate turbo code.

For example, consider the variable data rate mode of RC4, and consider the case where the length I of the information bit string input to the channel encoder is 769 bits.

In this case, the length N of the channel interleaver will be 3072 bits. Therefore, when the chain is constructed by the conventional method, the length L of the output string of the channel encoder is 1538 (= 769 * 2), and the bits of (3072-1538) are added to fit the channel interleaver length N (3072). Will have to be repeated.                     

This results in an effective coding rate of 3.994 (3072/769) but still an actual coding rate of 1/2.

However, when the transmission chain proposed in the present invention is used, L becomes 3845 (5 * 769), and puncturing as much as (3845-3072) will be performed in order to fit the channel interleaver length (N = 3072). will be.

As a result, the actual coding rate itself is 3.994 (3072/769), so that the actual coding gain can be obtained.

As another example, when the length I of the information bit string input to the channel encoder using the turbo code is 1535, the length N of the channel interleaver will be 3072 bits.

When the chain is constructed by the conventional method, the length L of the output string of the channel encoder is 3070 (1535 * 2), and bits (3072-3070) need to be repeated to fit the channel interleaver length (N = 3072). will be.

Accordingly, the effective coding rate is 2.001 (3072/1535), so that the actual coding rate is almost the coding rate used on RC using the proposed transmission chain.

Similarly, when using the transmission chain proposed in the present invention, L becomes 7675 (5 * 1535), and puncturing as much as (7675-3072) will be performed to fit the channel interleaver length (N = 3072). will be.

Accordingly, the actual coding rate itself is 2.001 (3072/1535), which is expected to give the same performance as the effective coding rate that can be obtained in a conventional transmission chain.

This is because the 1/2 rate turbo code itself is a code obtained by applying puncturing to the 1/5 rate turbo code. Therefore, using the transmission chain proposed in the present invention, it is possible to give good performance in all variable data rate domains.

What is needed to configure the transmission chain of FIG. 2 is a definition of turbo code at 1/5 rate. That is, the puncturing pattern of the data portion and the tail portion for making a turbo code of 1/5 rate. This can be solved through the following Table 1.

Data part Tail part x 111 111 (2 repetitions) y 111 111 z 111 111 (2 repetitions) x ' 000 111 (2 repetitions) y ' 111 111 z ' 111 111 (2 repetitions)

The next thing we need to do is decide how to perform symbol repetition and puncturing for the 1/5 rate turbo code. The basic premise used in the present invention is as follows.

(1) The puncturing is excluded for the systematic bits, and priority is given to symbol repetition.

(2) An equal amount of puncturing or symbol repetition is performed for the parity bits of the two constituent encoders forming the turbo encoder.

(3) In (2), puncturing or symbol repetition of uniform intervals for parity bits of each component encoder is performed.

(4) In (2), priorities are given to polynomials constituting parity of each component coder according to their importance.

The polynomials of the constituent encoders that make up the current 1/5 rate turbo code are '13' in octal for feedback, '15' in octal for (y, y ') parity bits, and (z, z ') For parity bits, use' 17 'in octal. In this case, the (y, y ') parity bits and the (z, z') parity bits of each component encoder are known to have different importance. Thus, in the case of puncturing or symbol repetition, it is desirable to give priority to (y, y ') parity bits or (z, z') parity bits under certain conditions.

Also, when constructing an algorithm, the size of the interleaver and the length of the output channel of the channel encoder are compared to determine whether to use a puncturing loop or a symbol repetition loop. It is desirable to be able to.

The algorithm proposed in the present invention to satisfy all of these prerequisites is as follows.

The rate matching method for the variable data rate and the multivariate data rate for the turbo code is performed by the following procedure.

)로 정의하고, 채널 부호화기의 출력열 중 하나의 심볼을 구성하는 시스테메틱 비트(x)와, 제1 패리티 비트들(y, y')과, 제2 패리티 비트들(z, z')에 대하여 우선 순위를 달리하여 서로 다른 R값을 할당한다. The length of the output string of the channel coder of the turbo code is L, the length of the information bit string including tail bits and spare bits is called I (L / 5), the desired interleaver size is N, and the other bits of the entire bit string. Assuming that the number of bits to be output once for R is R, the symbol repetition factor M is a maximum integer not exceeding N divided by L.

), The systematic bits (x), the first parity bits (y, y '), and the second parity bits (z, z') constituting one symbol of an output string of the channel encoder. Different R values are assigned with different priorities.

That is, R ₀ , R ₁ , R ₂ , for the cystematic bit (x), the first parity bits (y, y '), and the second parity bits (z, z'), respectively, A symbol repetition is performed in the repetition factor of M + 1 for R ₃ and R ₄ bits (having a value greater than or equal to 0), and M symbol repetition (or puncturing) is performed for the remaining bits.

The symbol factor M is calculated as 0 when N is smaller than L when puncturing, and as a value larger than 1 when symbol repetition is to be performed.

Accordingly, in the present invention, in order to make the length of the output string of the rate matching block N, in the case of puncturing, M becomes 0. Therefore, the corresponding bit is exported once according to the conditions of Table 2 below. Output 0 times. At this time, output 0 means puncturing.

In the case of symbol repetition, since M has a value of 1 or more, the corresponding bit is output M + 1 times according to the conditions of Table 2 below. Otherwise, M output is performed. That is, in the case of puncturing, it can be interpreted as an iterative algorithm in which M is zero.                     

R <I  I ≤R <3I   R ≥3I R ₀ (x) R I I R ₁ (y) 0

I R ₂ (z) 0 0

R ₃ (y ') 0

I R ₄ (z ') 0 0

That is, in order to give priority to the output of the systematic bit (x) in the case of symbol repetition in Table 2, if R is smaller than I, the system bit (x) is _{equal to} R (= R ₀ ) bits. M + 1 times and the remaining parity bits are output M times. In the case of puncturing, since M is 0, the output is performed once for the systematic bits x as many as the R bits, and the remaining parity bits perform puncturing.

Next, if the R value satisfies the condition of "I <R <3I" for symbol repetition to give priority to the output of the (y, y ') parity bits, , And output all the systematic bits of length M + 1 times, and in the case of (y, y ') parity bits, the amount of I allocated to the systematic bit string (x) is subtracted from the total R value. After that, M + 1 output is performed for half of this value (R ₁ or R ₃ ) (when one half of the value is odd, one bit of the first parity bits has the superior value). The R (R ₂ , R ₄ ) value is output M times for the (z, z ') parity bits of which 0 is zero. Similarly, in the case of puncturing, all the systematic bits x having the length of I are outputted once, and in the case of the (y, y ') parity bits, I is assigned to the systematic bit sequence x. After subtracting the amount from the total R value, output one half of this value (R ₁ or R ₃ ) once (if half of the value is odd, the superior value of one of the first parity bits is returned. Puncture is performed on the (z, z ') parity bits in which the remaining R (R ₂ , R ₄ ) becomes zero. The preceding conditions (1), (2) and (4) are satisfied by prioritizing these first parity bits (y, y ') over other second parity bits.

Finally, the part in which (z, z ') parity bits are output M + 1 times in the case of symbol repetition is limited only to the case of "R≥3I". However, R (R ₀ , R ₁ , R ₃ ) is output by repeating the length I of the information string by M + 1 with respect to the systematic bits (x) and the first parity bits (y, y '). For the second parity bits (z, z '), the amount of 3I allocated to the systematic bit string (x) and the first parity bits (y, y') is subtracted from the total R value. M + 1 is output for the bits corresponding to half (R ₂ , R ₄ ) of the value (when the half value is odd, one bit of the second parity bits has the superior value). In the case of puncturing, R (R ₀ , R ₁ , R ₃ ) is equal to the length of the information string once for the system bit (x) and the first parity bits (y, y ') once. Output and subtract the amount of 3I allocated to the systematic bit string (x) and the first parity bits (y, y ') from the total R value with respect to the second parity bits (z, z'), Puncture is performed on bits corresponding to half of this value (R ₂ , R ₄ ). (If the half value is odd, it has the superior value of any one of the second parity bits.)

The R _j value is assigned to each parity sequence sequence and systematic bit sequence, and then the algorithm is completed by applying the following uniform algorithm to satisfy condition (3).

Uniform Algorithm for Turbo Code

) output (5i+j)-th bit M+1 timesIf (

) output (5i + j) -th bit M + 1 times

Else, output (5i + j) -th bit M times, where i = 0 to I-1 and j = 0 to 4

" 관계를 만족하는 (5i+j) 번째 채널 부호화된 정보 비트는(터보 코드로 채널 부호화됨) M+1번 반복되어 출력되고, "

" 관계를 만족하지 못하는 (5i+j) 번째 채널 부호화된 정보 비트는(터보 코드로 부호화됨) M번 반복되어 출력된다. That is, for the i index from 0 to I-1 and the j index from 0 to 4, "

"The (5i + j) -th channel coded information bit that satisfies the relationship (channel coded with a turbo code) is repeatedly output M + 1 times, and"

"(5i + j) &lt; th &gt; channel-coded information bits that do not satisfy the relationship (encoded by turbo code) are repeatedly output M times.

However, the situation may be different when the convolution code is used. For example, consider forward RC4.

Consider the case of using a convolutional code of 1/4 code rate on all chains not defined on RC. In this case, as in the previous example, consider the case where the length I of the information bit string input to the channel encoder is 769 bits. In this case, the length N of the channel interleaver will be 3072.

When the transmission chain is constructed by the conventional method, the length L of the output string of the channel encoder is 1538 (769 * 2), and this bit is equal to (3072-1538) to fit the length of the channel interleaver (N = 3072). Will have to be repeated.

The effective coding rate will thus be 3.994 (3072/769) but the actual coding rate will still be half. However, if the transmission chain proposed in the present invention is used, L is 3076 (4 * 769), and puncturing as much as (3072-769) will be performed to fit the channel interleaver length (N = 3072). will be.

As a result, the actual coding rate itself is 3.994 (3072/769), so that the actual coding gain can be obtained.

As another example, however, if the length I of the information bit string input to the channel encoder is 1535, the length N of the channel interleaver will be 3072.

Therefore, if the chain is constructed by the conventional method, the length L of the output string of the channel encoder is 3070 (1535 * 2), and bits (3072-3070) will have to be repeated to fit N.

As a result, the effective coding rate is 2.001 (3072/1535), so that the actual coding rate is almost the coding rate currently used by RC.

However, if the transmission chain proposed in the present invention is used, L becomes 6140 (4 * 1535), and puncturing as much as (6140-3072) is performed to fit the channel interleaver length N (= 3072). Will be done.

As a result, the actual coding rate itself is 2.001 (3072/1535).

However, the coding rate here is the coding rate obtained through puncturing of about 50% from 1/4 rate. In the case of convolutional codes, unlike in the case of turbo codes, the optimal code polynomials for 1/2 rate, 1/3 rate, and 1/4 rate are defined respectively. The chain's method can have parts that can degrade performance. In order to solve this problem, the present invention proposes two solutions.

Second embodiment

Here, a method of configuring a transmission chain of RC for supporting a variable or multivariable data rate mode of another method when using a convolution code instead of a turbo code in the first embodiment is proposed.

3 is a diagram illustrating a transmission chain of a variable data rate and a multivariate data rate for a convolutional code according to the present invention.

Referring to FIG. 3, in the convolutional code, when a chain other than the normal data rate chain is to be formed on each RC, a convolutional coder of 1 / n rate is used as the channel encoder (S30). Rate matching puncturing (n * I> N) or rate matching repetition (n * I <N) is used for matching between the length L (= n * I) of the output string of the channel encoder and the length N of the channel interleaver. (S31). In this case, the inverse n of the coding rate is a value signaled at the upper level and is a value determined in the negotiation process.

That is, by adding the inverse of the coding rate to the parameter signaled in the negotiation process for the conventional variable data rate or the multivariate data rate mode, it is possible to configure a variable chain for the convolutional code and a transmission chain for the multivariate data rate.

Here, as in the first embodiment, the rate matching puncturing or rate matching repetition method is performed uniformly with respect to the information bit string input to the channel encoder.

For example, assume that a negotiation process is performed to use the variable data rate mode on the current forward RC4. If the interleaver size N is 3072 and I is 1535, n may be signaled as two.

This signaling enables the construction of a transport chain that always provides the same or better performance than the current variable data rate mode or multivariate data rate mode.

To complete the transmission chain of FIG. 2, a rate matching algorithm for the convolutional code must be configured. This may be composed of the following uniform rate matching algorithm.

번째 비트 인덱스를 배치시킨다. 이때, 상기 k는 0부터 N-1까지 증가하는 값이 된다.The kth output symbol of the symbol repetition block is one of the I information bit strings.

Place the first bit index. In this case, k is a value increasing from 0 to N-1.

As described above, in the current 3GPP2 system, when the variable data rate, the variable data rate mode for the auxiliary channel, or both are supported, the coding gain due to the actual code rate reduction in addition to the time diversity gain due to symbol repetition It can be obtained, thereby lowering the required transmit power.

In addition, it is possible to give different priority to each parity bit and systematic bits to give better performance in the construction of a rate matching algorithm for the turbo code. In addition, the increase in the coding gain results in a decrease in power, thereby making it possible to more efficiently manage the loading of the system.

Those skilled in the art will appreciate that various changes and modifications can be made without departing from the spirit of the present invention.

Therefore, the technical scope of the present invention should not be limited to the contents described in the examples, but should be defined by the claims.

Claims (9)

A method of rate matching in a mobile communication system supporting a regular data rate mode through radio configuration (RC), Turbo coding at a minimum coding rate among coding rates supported by the mobile communication system for the input bit string without considering the radio configuration to support a data rate different from the normal data rate. Performing); And Performing rate matching by performing one or more of bit puncturing and bit repetition such that the turbo coded bit string matches the interleaver size defined in the current wireless configuration. Rate matching method comprising a. The method of claim 1, When the minimum coding rate is 1/5, the turbo coded bit string includes a systematic bit (x), a first parity bit (y, y '), and a second parity bit (z, z'). A rate matching method (x is a systematic bit output from a first constituent encoder, y and z are parity bits output from a first constituent encoder, and y 'and z' are a second constituent encoder) Parity bits output from). The method of claim 2, Turbo coding having the 1/5 coding rate is a puncturing pattern Data part Tail part x 111 111 (2 reps) y 111 111 z 111 111 (2 reps) x ' 000 111 (2 reps) y ' 111 111 z ' 111 111 (2 reps) The rate matching method (x 'is a systematic bit output from the second constituent encoder), characterized in that is performed using. The method of claim 2, The priority of the bit repetition has priority in the order of the systematic bits (x), the first parity bits (y, y '), and the second parity bits (z, z'). Way. The method of claim 4, wherein According to the priority, for each component bit of the turbo coded bit stream, the number of bits to perform the iteration by applying the first iteration number, and the number of bits to perform the iteration by applying the second iteration number step; And Performing bit repetition by applying the determined number of bits to each component bit Rate matching method comprising a. The method of claim 5, The first repetition number of times

+ 1, and the second repetition number is

Rate matching method (N is the desired interleaver size, L is the output string length of the turbo encoder). The method of claim 6, The repetition of the component bits may include performing a bit repetition on a specific component bit such that the bits to which the first repetition number is applied are uniformly distributed throughout the turbo coded bit string. Matching method. The method of claim 6, Performing iteration for each component bit,

Bit repetition is performed by applying a first repetition factor to a bit satisfying a value, and bit repetition is performed by applying a second repetition factor to a bit that is not satisfied. Length of the information bit stream including the spare bits, R is the number of bits to be output once more, Rj is the jth channel coded information bits, j is an index according to each bit component, i is a bit index according to the input bit string) The method of claim 5, The number of bits to perform the iteration by applying the first iteration number is R <I  I ≤R <3I   R ≥3I R ₀ (x) R I I R ₁ (y) 0

I R ₂ (z) 0 0

R ₃ (y ') 0

I R ₄ (z ') 0 0

The rate matching method of claim 1, wherein I is an information bit string length including tail bits and spare bits, R is the number of bits to be output once more, and Rj is the jth channel coded information bits.

Images