KR20010009145A - Method for rate matching in mobile communication system - Google Patents

Abstract

PURPOSE: A method for performing a rate matching in a mobile communication system is provided to supply an integrated rate matching algorithm capable of being commonly used in convolution codes and turbo codes, to supply an optimized puncturing/repetition pattern in accordance with a code rate so as to prevent code capacity deteriorations, and to supply the optimized puncturing/repetition pattern according to importance of code symbols when the convolution codes and the turbo codes are rate-matched. CONSTITUTION: A mobile communication system identifies channel codes of bit strings according to given identification information, and differently decides a puncturing or a repetition location within code symbols according to the identified channel codes. If forms of the identified codes are convolution codes, a puncturing is sequentially performed from a code symbol of a bit string having a lowest importance, among bit strings of the convolution codes. A puncturing algorithm performs the puncturing by making an initialization value as 1.

Description

Method for rate matching in mobile communication system

The present invention relates to a mobile communication system, and more particularly, to a rate matching method in a mobile communication system which can be commonly used for a convolutional code and a turbo code in a rate matching method of a W-CDMA mobile communication system.

First, the rate matching algorithm is briefly mentioned.

Rate matching algorithm is an algorithm that punctures or repetitions an output bit string of a channel encoder in order to match a code rate of a bit string output from the channel encoder to an arbitrary value. It is being studied in the Third Generation Partnership Project (3GPP), which defines technical standards for next-generation mobile communication systems.

Currently, the third generation common project defines a puncturing / repetition algorithm for convolution code or turbo code as a rate matching method for transport multiplexing of a W-CDMA mobile communication system. It is proposed in the CDMA standard. In the rate matching algorithm proposed in the W-CDMA standard, puncturing / repetition is performed on a bit basis, and the same algorithm is applied to a convolution code or a turbo code.

The rate matching method of the convolutional code or turbo code proposed in the W-CDMA specification will now be described in more detail.

First, puncturing for the convolutional code is performed uniformly at equal intervals as possible considering the positions of the punctured bits. Therefore, in order to satisfy the uniform puncturing condition for the convolutional code, a puncturing distance n is calculated in bits for the entire sequence, and puncturing is performed for each of the obtained nth bits.

For example, if the puncturing distance is 5, puncturing is performed once every 5 distances for the entire sequence. However, if the puncturing distance is 5.5, puncturing with a distance of 5 is performed by half and puncturing with a distance of 6 is performed by half.

At this time, according to the puncturing algorithm currently proposed for the turbo code, the puncturing is performed to maintain uniformity among the punctured code symbols on the basis of the code symbol unit rather than the bit unit. do. Therefore, the puncturing algorithm for the turbo code operates to first determine the location of the code symbol to be punctured and then puncture the first parity component and the second parity component alternately.

In contrast, the puncturing can occur in code symbol units at a distance of 2 times from the previous position, and then the first parity component and the second parity component are simultaneously punctured for each code symbol where each puncturing occurs. A method of making is also proposed.

However, such a rate matching method for a convolutional code or a turbo code does not consider optimization for puncturing / repetition patterns in code symbols that vary according to the parameters of an algorithm, resulting in the following problems.

First, the turbo code is punctured / repeated without considering that each code bit component has a different importance. That is, considering that the systematic bit components play an important role as compared to the parity bit components, excluding puncturing for the systematic bit components may provide better performance, but the conventional rate matching method Does not provide a protection mechanism for puncturing the systematic bit components, resulting in degradation of code performance.

In addition, the turbo code may configure a 1/2 code rate by puncturing, but the conventional method does not provide an optimized puncturing / repetition pattern for the 1/2 code rate.

Second, in the case of convolutional codes, convolutional codes, like turbo codes, have bits that play relatively important roles in code symbols, whereas conventional rate matching methods do not provide puncturing / repetition patterns for them. For example, considering a convolutional code with 256 state 1/3 code rates, which are widely used in mobile communication systems, the third bitstream of the convolutional code plays the most important role relative to the first and second bitstreams. In the bits or the conventional rate matching method, only a third bit string is punctured according to a configuration parameter, thereby causing a decrease in code performance.

Accordingly, an object of the present invention has been made in view of the above-mentioned problems of the prior art, it is configured in an integrated form to be commonly used in convolutional code and turbo code, and also optimized according to the code rate of the code It is to provide a rate matching method in a mobile communication system that can generate a processing / repeating pattern.

In addition, according to another aspect of the present invention for achieving the other object as described above, the code rate and the position of the code symbol that puncturing / repetition occurs when puncturing / repetition as a symbol unit with respect to the convolution code The purpose of the present invention is to provide a rate matching method in a mobile communication system that can improve code performance by varying the importance of each code symbol.

According to an aspect of the present invention for achieving the above object, a rate matching method in a mobile communication system is characterized in that a puncturing or iterative algorithm is performed on the basis of code symbols for a predetermined bit string having an arbitrary code rate. In the mobile communication system, a channel code of the bit string is identified according to given identification information, and a puncturing or repeating position in the code symbol is determined differently according to the identified channel code.

Preferably, when the type of the identified channel code is a convolutional code, the puncturing is performed sequentially from the code symbol of the bit string having the lowest importance among the bit sequences of the convolutional code.

1 is a conceptual diagram illustrating a rate matching method in a mobile communication system according to the present invention.

Hereinafter, a configuration and an operation according to an exemplary embodiment of the present invention will be described with reference to the accompanying drawings.

The present invention proposes an integrated rate matching method that can be commonly used for convolutional codes and turbo codes. In this case, the rate matching method according to the present invention is a code symbol unit, and the puncturing / repetition position for each code is configured differently according to the type of channel code to which the algorithm is applied.

1 is a conceptual diagram illustrating a rate matching method in a mobile communication system according to the present invention.

Referring to FIG. 1, in the rate matching method according to the present invention, the puncturing / repetition is performed based on a code symbol unit, and at this time, the puncturing / repetition is performed according to the identification information parameter 100 of the channel code applied to the algorithm. Change the position in the symbol.

Accordingly, in block 102 to be rate matched, the channel code is identified according to the identification information parameter 100 of the channel code. Accordingly, the rate matching algorithm 101 based on the code symbol performs puncturing / repetition. Perform 102.

The rate matching method for the convolutional code and the turbo code in units of such code symbols is implemented according to the following flow.

Step 1, determine the initial value.

Step 2, calculate the decrement value.

Step 3 proceeds to the next loop in units of one or more symbols.

Step 4, Subtract the decrement value continuously from the initial value, and if the result value is smaller than '0', execute the puncturing / repeat algorithm. In this case, the algorithm may be configured to have a range other than '0'. In the algorithm to be presented in the following examples, the initial value is represented as 'e', and the second step is represented by 'e = e-decrement value'.

The rate matching method according to the present invention will now be described by dividing into a case of a convolutional code and a turbo code.

First embodiment

The first embodiment describes a rate matching method for convolutional codes having 256 state numbers and having 1/2 or 1/3 code rates. (The convolutional polynomials of the convolution code are (561, 753), (557, 663, 711) and are assumed to be terminated.)

First, in the case of a convolutional code having a 1/2 code rate, it has priority in the order of the first and second in the bit position in the code symbol. At this time, the importance of the convolution code is determined as the Hamming distance. That is, the Hamming distance of the first bit string and the Hamming distance of the second bit string are calculated to determine that the bit string having the high hamming distance is of high importance.

Therefore, in a convolutional code with a 1/2 code rate, the first bit components are positioned second because the hamming distance of the first bit components in the code symbol is higher than that of the second bit components. It can be judged that it plays an important role than the bit components.

Therefore, the first and second bit components are punctured first, followed by the first and second bit components. It can be seen that the position of puncturing in the code symbol is changed differently according to the importance of each component bit.

On the other hand, the repetition is performed first with respect to the first bit component, as opposed to the case of puncturing, and then repeats the second bit components. This is to first iterate over important bit components.

On the other hand, in the case of a convolutional code having a 1/3 code rate, the order of importance is the third, second and first order of the bit position in the code symbol. This may be determined by the fact that the first bit string has a Hamming distance of 5, the second bit string has a Hamming distance of 6, and the third bit string has a Hamming distance of 7.

Therefore, since the third bit components in the code symbol play a more important role than the second and first, the first and second most significant bit components are punctured first, followed by the second and third Sequentially puncture the bits located at.

On the other hand, the repetition is performed first with respect to the important code symbol as opposed to the case of puncturing, so that the third bit component is repeated first, followed by the second and first bit components.

The above case corresponds to a case where the Viterbi decoding process is configured by using the termination information, and the puncturing pattern in the symbol in the case of not using the termination information shows another aspect and the pattern is obtained through simulation. If the situation is different from the above case, it is related to the 1/3 convolution code. If the termination information is not used, puncturing the second bit in the symbol shows the best performance.

The rate matching algorithm for the convolution code configured to have such an order is configured as follows.

First, for convenience of explanation of the rate matching algorithm for the convolution code, the following definition is made. (Code rate = 1 / n)

1.S _N = {N ₁ , N ₂ , ....., N _L } = an ordered set of possible numbers per block (in ascending order)

2. N _C = block size for rate matching = nN

3.S ₀ = {d ₁ , d ₂ , ....., d _Nc } = N _C data bit set

Then, the rate matching algorithm for the convolution code is as follows.

-For puncturing

y = N _c -N _i

e = N

m = l

count = 0

if (n == 2)

X _n = 0

endif

if (n == 3)

X _n = 2

endif

do while m ≤ N

e = e-2 * y

if e ≤ 0 then

puncture bit n * m-X _n from set S ₀

count = count + 1

e = e + 2 * N

endif

m = m + 1

end do

-For repetition

y = N _i -N _c

e = N

m = 1

if (n == 2)

X _n1 = 1, X _n2 = 0

endif

if (n == 3)

X _n1 = 0, X _n2 = 1, X _n3 = 2

endif

do while m≤ N

e = e-2 * y

if (e ≤-4 * N) & (n == 3) then

repeat bit n * m-X _n3 from set S ₀

e = e + 2 * N

end if

if e ≤ -2 * N then

repeat bit n * m-X _n2 from set S ₀

e = e + 2 * N

endif

if e ≤ 0 then

repeat bit n * m-X _n1 from set S ₀

e = e + 2 * N

end if

m = m + 1

end do

end if

On the other hand, when the termination information is not used, the algorithm is appropriately transformed when puncturing the first bit position in the symbol is changed to the second bit position in the above algorithm.

Considering the conventional algorithm, the third bit is always punctured when puncturing of A / (6 * n) (a natural number greater than n = 1) occurs when the initial value is A. The puncturing of the third bit causes performance degradation, so the initial value is set to 1 to completely prevent this problem.

Second embodiment

The second embodiment describes a rate matching method for turbo codes having arbitrary code rates.

The turbo code has the following rules in consideration of the fact that the uncoded systemic component bits among the bit components output from the convolutional encoder have a more important property in terms of information transmission than the parity component bits. do.

First, the cystic component bits do not puncture.

Second, balanced puncturing between the parity component bits is performed. That is, the first parity component bit and the second parity component bit have the same number of punctured bits.

Third, uniform puncturing is performed between each parity component bit. This means that the first parity component bit and the second parity component bit are punctured in a pattern having a constant distance.

Fourth, in the case of repetition, only the cymatic component bits are repeated.

In this case, when puncturing / repetition more than the number of code symbols is performed, in order to balance the first parity component bit and the second parity component bit, the decrement value is continuously subtracted from the initial value and various ranges other than '0' are added. Set additionally to perform the puncturing / repeat optimized according to the applicable range.

Therefore, when puncturing / repetition of more than this number of code symbols is performed, puncturing alternates between the first parity component bit and the second parity component bit. The first parity component bit and the second parity component bit are punctured simultaneously. The repetition repeats the cymatic component bits first, and repeats the second parity component bit and the first parity component bit after the repetition of the number of code symbols or more.

In addition, since termination bits play a relatively important role in the turbo code, the rate matching method according to the present invention adjusts the initial value of the puncturing algorithm so that puncturing does not occur in the termination bit portion. Preferably, the initial value is set to '1' to eliminate puncturing in the termination bit portion as much as possible.

The rate matching method for a turbo code that provides a puncturing / repetition pattern in units of code symbols can be configured by convolutional code and is as follows.

First, an integrated rate matching algorithm of a convolution code and a turbo code is defined as follows for convenience of description. (Code rate = 1 / n)

1.S _N = {N ₁ , N ₂ , ....., N _L } = an ordered set of possible numbers per block (in ascending order)

2. N _C = block size for rate matching = nN

3.S ₀ = {d ₁ , d ₂ , ....., d _Nc } = N _C data bit set

Then, the integrated rate matching algorithm for the convolutional code and the turbo code is as follows.

-For puncturing

y = N _c -N _i

e = N

m = l

flag = 0

if (n == 2)

X _n = 0

endif

if (n == 3)

X _n = 2

do while m ≤ N

e = e-2 * y

if e ≤ -2 * N then

puncture bit n * m, n * m-1 from set S ₀

e = e + 4 * N

endif

if e ≤ 0 then

if (convolutional code)

puncture bit n * m-Xn from set S ₀

endif

if (turbo code)

if flag == 0

puncture bit n * m from set S ₀

flag = 1

endif

if flag == 1

puncture bit n * m-1 from set S ₀

flag = 0

endif

endif

e = e + 2 * N

endif

m = m + 1

end do

-For repetition

y = N _i -N _c

e = N

m = l

if (n == 2)

X _n1 = 1, X _n2 = 0

endif

if (n == 3)

X _n1 = 0, X _n2 = 1, X _n3 = 2

endif

if (turbo code)

X _n1 = 2, X _n2 = 0, X _n3 = 1 or X _n1 = 2, X _n2 = 1, X _n3 = 0

endif

do while m ≤ N

e = e-2 * y

if (e ≤ -4N) && (n == 3) then

repeat bit n * m-X _n3 from set S ₀

e = e + 2 * N

end if

if e ≤ -2 * N then

repeat bit n * m-X _n2 from set S ₀

e = e + 2 * N

end if

if e ≤ 0 then

repeat bit n * m-X _n1 from set S ₀

e = e + 2 * N

end if

m = m + 1

end do

end if

Third embodiment

In the third embodiment, a case in which the convolutional code is performed by different positions within the symbol based on a plurality of code symbols during puncturing / repetition will be described.

That is, the conventional puncturing of the convolutional code is uniformly performed only at one position within the code symbol of the bit string, but in the present invention, the puncturing may be alternately performed based on the code symbol of the other bit string. This means that the rate matching algorithm for the turbo code can be applied to the convolutional code.

The rate matching algorithm for this convolutional code is shown below. The algorithm shown below is based on a convolutional code with a half code rate.

-For puncturing

y = N _c -N _i

e = N

m = l

flag = 0

do while m ≤ N

e = e-2 * y

if e ≤ 0 then

if flag == 0

puncture bit 2 * m from set S ₀

flag = 1

end if

if flag == 1

puncture bit 2 * m-1 from set S ₀

flag = 0

end if

e = e + 2 * N

endif

m = m + 1

end do

In this case, the integrated rate matching algorithm of the convolutional code and the turbo code is as follows.

-For puncturing

y = N _c -N _i

e = N

m = l

flag = 0

if (n == 3)

X _n = 2

do while m ≤ N

e = e-2 * y

if e ≤ -2 * N then

puncture bit n * m, n * m-1 from set S ₀

e = e + 4 * N

endif

if e ≤ 0 then

if ((convolutional code) && (n = 3))

puncture bit n * m-Xn from set S ₀

endif

if ((turbo code)) ∥ ((convolutional code) && (n = 2)))

if flag == 0

puncture bit n * m from set S ₀

flag = 1

endif

if flag == 1

puncture bit n * m-1 from set S ₀

flag = 0

endif

endif

e = e + 2 * N

endif

m = m + 1

end do

As described above, the present invention performs the puncturing / repetition having an optimized pattern according to the channel code in units of code symbols, and thus has the following effects.

First, by providing an integrated rate matching algorithm that can be commonly used in convolutional code and turbo code, an optimized puncturing / repetition pattern according to the code rate can be provided to prevent performance degradation of the code.

Second, it is possible to provide an optimized puncturing / repetition pattern according to the importance of code symbols in rate matching for convolutional codes and turbo codes.

Thirdly, the required amount is reduced in the computational complexity and complexity of the rate matching algorithm.

Claims (3)

In a mobile communication system, a puncturing or iterative algorithm is performed based on a code symbol for a predetermined bit string having an arbitrary code rate. Identifying a channel code of the bit string according to a given identification information, and determining a puncturing or repeating position in the code symbol according to the identified channel code. The mobile communication method of claim 1, wherein when the identified channel code is a convolutional code, the puncturing is performed sequentially from code symbols of a bit string having the lowest significance among the bit strings of the convolutional code. Rate matching method in the system. The method of claim 1, wherein the puncturing algorithm performs puncturing by setting an initial value of 1.