KR102205630B1 - Early termination apparatus for enhancing efficiency of code decoder and method thereof - Google Patents

Abstract

The present invention relates to an early termination device and a method thereof. According to an embodiment of the present invention, the early termination device includes: a codeword receiving unit for receiving a constituent code codeword input to a decoder (decoder); an expected reduction rate calculation unit detects a syndrome of a constituent protection codeword received in a half-iteration decoding process performed a preset number of times, counts the number of detected syndromes, counts the number of syndromes, and calculates the expected reduction rate of the syndrome based on the number of counted syndromes; and an early termination determination unit which calculates a difference in the detection rate of the syndrome in the hemispherical decoding process, and determines whether or not the hemispherical decoding process is early terminated based on the calculated difference in a detection rate and the calculated expected reduction rate.

Description

Early termination device and method for increasing the efficiency of a code decoder {EARLY TERMINATION APPARATUS FOR ENHANCING EFFICIENCY OF CODE DECODER AND METHOD THEREOF}

The present invention relates to an early termination apparatus and a method for increasing the efficiency of a code decoder, and more particularly, to a technical idea for determining whether or not a decoding process is early terminated using a code decoder.

The block turbo code (BTC) forms an error correction code word in the form of a two-dimensional or more matrix based on a short-length constituent code. In general, the code used in the code word composition is binary or non- As a binary block code, a hamming code, a bose-chaudhuri-hocquenghem (BCH) code, and a reed-solomon (RS) code correspond to this.

The constituent code is expressed using parameters of (n,k,d), where n is a codeword length, k is an information bit length, and d is a hamming distance.

Therefore, if a block turbo code in the form of a two-dimensional matrix is constructed using the same type of constituent code, the codeword is expressed as (n,k,d) ² , where the codeword length is n ² and the number of information bits is k. ² , the Hamming distance becomes d ² . It should be noted here that since the Hamming distance of the block turbo code is expressed as the product of the Hamming distance of each constituent code, the error correction performance is further increased.

The block turbo code generally performs error correction based on soft-input soft-output (SISO) information, and this algorithm is called Chase-Pyndiah algorithm. This decoding algorithm shows very good error correction capability and can achieve Shannon's channel limit capacity.

The Chase-Pyndiah decoding algorithm can be roughly divided into two stages, the first process is hard decision decoding (HDD) based on the Chase-2 algorithm, and the second process is iterative decoding. This is the process of calculating extrinsic information.

The first hard decision decoding process occupies most of the computational complexity in the decoding process of the block turbo code, and the computational complexity in this process is the location of the least reliable bits (LRB), the main parameters of the Chase-2 algorithm. It is determined by the number.

Here, the minimum reliability bit refers to a symbol value of a constituent code codeword input to the block turbo code decoder, that is, bits having a relatively small size among soft values. This is generally expressed as a variable called'p', and can be variously set to a size value of 3 to 5 depending on the communication environment. In the Chase-2 algorithm, 2 ^p test patterns are generated in the process of executing the algorithm based on the size of the p value, and the same number of hard decision decoding processes are performed.

On the other hand, the Chase-Pyndiah algorithm is evaluated to have a relatively high decoding complexity compared to other error correction codes, although it has excellent error correction capability, and various studies have been conducted to overcome this.

Specifically, methods of continuously determining whether error correction is successful during an iterative decoding process using a decoding algorithm, and ending error correction early if it is determined that all errors have been fixed are being studied.

However, the early termination techniques proposed so far can obtain a positive effect caused by the technique only when the channel environment is relatively good.

In other words, even if the initial error rate in the codeword is quite high due to the influence of the channel environment, the early termination technique should always perform error correction as many times as the maximum number of half-repeated decoding, but it performs error correction as many as the maximum number of half-repeated decoding. Even so, there is a problem that there is no guarantee that error correction will always be completed successfully through this.

Korean Patent Laid-Open Publication No. 2003-0047178, "Repetitive decoding method of block turbo codes and recording medium storing a program of repeating decoding of block turbo codes"

R, Pyndiah, "Near-optimum decoding of product codes: block turbo codes," IEEE Trans. Commun., vol 46, No. 8, pp. 1003-1010, Aug. 1998. Chen, Guo Tai, et al. "An efficient stopping criterion for turbo product codes." IEEE communications letters 11.6 (2007): 525-527. Tawfeek, Hazem, Ashraf Mahran, and Gamal M. Abdel-Hamid. "A Reliability-based Stopping Criterion for Turbo Product Codes." 2018 14th International Computer Engineering Conference (ICENCO). IEEE, 2018. AlMahamdy, Mohammed, and Jeffrey Dill. "Early Termination of Turbo Decoding by Identification of Undecodable Blocks." 2017 IEEE Wireless Communications and Networking Conference (WCNC). IEEE, 2017. Xia, Tian, Hsiao-Chun Wu, and Scott C-H. Huang. "A new stopping criterion for fast low-density parity-check decoders." 2013 IEEE Global Communications Conference (GLOBECOM). IEEE, 2013.

An object of the present invention is to provide an early termination apparatus and method for determining whether to terminate a decoding process early by using syndrome information of a constituent code codeword input to a block turbo code decoder.

In addition, the present invention is an early termination apparatus and method capable of effectively reducing the average number of half-repeated decoding without deteriorating the error correction ability by using the degree of reduction in the charge ratio of the double-error syndrome as the half-repeated decoding process proceeds. Want to provide.

An early termination device according to an embodiment includes a codeword receiver for receiving a constituent code codeword input to a decoder, and a constituent protection code received in a half-iteration decoding process performed a predetermined number of times. An expected reduction rate calculator that detects the syndrome of a word, counts the number of detected syndromes, calculates the expected reduction rate of the syndrome based on the counted number of syndromes, and the difference in the detection rate of the syndrome in the semi-repetitive decoding process And an early termination determination unit that determines whether or not the half-repeated decoding process is prematurely terminated based on the calculated detection rate difference and the calculated expected reduction rate.

According to one side, the component code code word may be an error correction code code word based on a block turbo code (BTC).

According to one side, the syndrome may be a double-error syndrome.

According to one side, the expected reduction rate calculation unit may initialize a tolerance counter value, a syndrome counter value, and a half-repeat counter variable for determining whether to terminate early or not before performing the half-iteration decoding process. have.

According to one side, the expected reduction rate calculator may set a determination threshold value of a size corresponding to at least one of a type, a code length, and a code rate of a component code of the received component code word.

According to one side, the expected reduction rate calculator may increase the syndrome counter value by '1' each time a syndrome is detected in a half-iteration decoding process.

According to one side, the expected reduction rate calculation unit may calculate an initial detection ratio by using the syndrome counter value increased through the first half-repeated decoding process, and calculate the expected reduction ratio by dividing the calculated initial detection ratio by the maximum number of half-repeated decoding. have.

According to one side, the early termination determination unit calculates the m-1th detection rate by using the syndrome counter value increased through the m-1th (where m is a positive integer greater than or equal to 2) half-repetitive decoding process, and The m-th detection rate may be calculated using the syndrome counter value increased through the semi-repetitive decoding process, and the detection rate difference may be calculated by subtracting the m-th detection rate from the m-1th detection rate.

According to one side, when the calculated detection rate difference is smaller than the calculated expected reduction rate, the early termination determination unit may increase the tolerance counter value.

According to one side, when the increased tolerance counter value coincides with a preset determination threshold value, the early termination determination unit may early terminate the half-repeated decoding process.

The early termination method according to an embodiment includes the steps of receiving a constituent code codeword input to a decoder in a codeword receiving unit, and a half-iteration performed by a predetermined number of times in an expected reduction rate calculation unit. ) In the step of detecting the syndrome of the configuration protection codeword received in the decoding process, counting the number of detected syndromes, calculating the expected reduction rate of the syndrome based on the counted number of syndromes, and in the early termination determination unit. , Calculating a difference in detection rate of the syndrome in the semi-repetitive decoding process, and determining whether or not the semi-repetitive decoding process is prematurely terminated based on the calculated detection rate difference and the calculated expected reduction rate.

According to an embodiment, it is possible to determine whether to terminate the decoding process early by using syndrome information of the constituent code codeword input to the block turbo code decoder.

According to an embodiment, as the half-repeated decoding process proceeds, the average number of half-repeated decoding can be effectively reduced without deteriorating the error correction capability by using a reduction in the charge ratio of the double-error syndrome.

1 is a view for explaining an early termination apparatus according to an embodiment.
FIG. 2 is a diagram illustrating a detection rate of each syndrome according to an increase in the number of half-repeated decoding in a process of performing error correction using the existing Chase-Pyndiah algorithm.
3 is a diagram for explaining error correction performance according to a determination threshold setting in an early termination apparatus according to an embodiment.
FIG. 4 is a diagram illustrating a measurement result of an average number of half-repeated decoding according to a determination threshold setting in an early termination apparatus according to an exemplary embodiment.
5 is a diagram for describing an early termination method according to an embodiment.

Hereinafter, various embodiments of the present document will be described with reference to the accompanying drawings.

The embodiments and terms used therein are not intended to limit the technology described in this document to a specific embodiment, and should be understood to include various changes, equivalents, and/or substitutes for the embodiment.

In the following description of various embodiments, when it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the subject matter of the invention, a detailed description thereof will be omitted.

In addition, terms to be described later are terms defined in consideration of functions in various embodiments and may vary according to the intention or custom of users or operators. Therefore, the definition should be made based on the contents throughout this specification.

In connection with the description of the drawings, similar reference numerals may be used for similar elements.

Singular expressions may include plural expressions unless the context clearly indicates otherwise.

In this document, expressions such as "A or B" or "at least one of A and/or B" may include all possible combinations of items listed together.

Expressions such as "first," "second," "first," or "second," can modify the corresponding elements regardless of their order or importance, and to distinguish one element from another It is used only and does not limit the components.

When any (eg, first) component is referred to as being “(functionally or communicatively) connected” or “connected” to another (eg, second) component, a component is It may be directly connected to the element, or may be connected through another element (eg, a third element).

In the present specification, "configured to (configured to)" is changed according to the situation, for example, hardware or software, "suitable for," "having the ability to," "... ," "made to," "can do," or "designed to" can be used interchangeably.

In some situations, the expression "a device configured to" may mean that the device "can" along with other devices or parts.

For example, the phrase “a processor configured (or configured) to perform A, B, and C” means a dedicated processor (eg, an embedded processor) for performing the operation, or by executing one or more software programs stored in a memory device. , May mean a general-purpose processor (eg, CPU or application processor) capable of performing corresponding operations.

In addition, the term'or' means an inclusive OR'inclusive or' rather than an exclusive OR'exclusive or'.

That is, unless otherwise stated or clear from the context, the expression'x uses a or b'means any one of natural inclusive permutations.

In the above-described specific embodiments, constituent elements included in the invention are expressed in the singular or plural according to the presented specific embodiments.

However, the singular or plural expression is selected appropriately for the situation presented for convenience of description, and the above-described embodiments are not limited to the singular or plural constituent elements, and even constituent elements expressed in plural are composed of the singular or However, even if it is a constituent element expressed in the singular, it can be composed of pluralities.

Meanwhile, although specific embodiments have been described in the description of the present invention, various modifications may be made without departing from the scope of the technical idea implied by various embodiments.

Therefore, the scope of the present invention is limited to the described embodiments and should not be defined, but should be defined by the claims and equivalents as well as the claims to be described later.

1 is a view for explaining an early termination apparatus according to an embodiment.

Referring to FIG. 1, the early termination apparatus 100 according to an embodiment may reduce the complexity of calculation compared to the conventional one by reducing the average number of decoding in the decoding process without deteriorating the error correction performance for the block turbo code. That is, the early termination apparatus 100 may provide an effective early termination technique in a relatively poor channel environment, that is, a low-intermediate signal to noise ratio (SNR) region.

The early termination device 100 may provide a technique for measuring an error syndrome reduction rate and predicting a possibility of error correction failure using syndrome information.

Specifically, an early termination technique in other error correction fields such as LDPC code or turbo code uses an average of prior probability values of each symbol. However, these techniques require excessive use of various arithmetic operators, such as multiplication and comparator operations, which may cause additional computational complexity.

However, the early termination device 100 uses only the syndrome information of each component code word.

This syndrome information calculation process is already included in all decoding algorithms of the block turbo code, and it can be seen that there is little increase in complexity added due to the application of the early termination technique.

Accordingly, the early termination device 100 can reduce the complexity of decoding calculation in a more efficient manner compared to the known early termination technology.

In addition, the early termination apparatus 100 may estimate an optimal threshold value for preventing a decrease in error correction capability.

More specifically, the early termination apparatus 100 may use an arbitrary judgment threshold in applying the early termination technique.

Here, the judgment threshold can be set to a different value depending on conditions such as the type of block turbo code and the code length, which can prevent the deterioration of error correction performance that may be caused by the application of the early termination technique and at the same time. Can be.

That is, the determination threshold value used by the early termination device 100 determines what size value should be used according to the type of code through the simulation process, and how the size of the determination threshold value varies according to the type of code, and the determination result is considered. Can be set.

To this end, the early termination apparatus 100 may include a codeword receiving unit 110, an expected reduction ratio calculating unit 120, and an early termination determining unit 130.

Meanwhile, the early termination apparatus 100 described below may be implemented in the form of an algorithm for determining whether to terminate early based on a semi-repetitive decoding process.

The codeword receiving unit 110 according to an embodiment may receive a constituent code codeword input to a decoder.

According to one side, the component code codeword may be an error correction code word based on a block turbo code (BTC), and the decoder may be a block turbo decoder.

The expected reduction rate calculation unit 120 according to an embodiment detects a syndrome of a configuration protection codeword received in a half-iteration decoding process performed a predetermined number of times, and the number of detected syndromes Is counted, and the expected reduction rate of the syndrome may be calculated based on the counted number of syndromes.

) 일 수 있다. 또한, 기대 감소율 산출부(120)를 통해 검출되는 신드롬은 노-에러(no-error) 신드롬, 싱글-에러(single-error) 신드롬, 더블-에러(double-error) 신드롬 중 적어도 하나 이상의 신드롬일 수 있다. For example, the preset number of half-repeated decoding processes is the maximum half-repeated decoding number (

) Can be. In addition, the syndrome detected through the expected reduction rate calculation unit 120 is at least one of a no-error syndrome, a single-error syndrome, and a double-error syndrome. I can.

Here, the no-error syndrome is a syndrome detected when no error exists in the component code codeword, and the single-error syndrome is a syndrome detected when an odd number of errors exists in the component code word, and double-error. The syndrome may be a syndrome that is detected when an even number of errors exists in the constituent code codeword.

Preferably, the expected reduction rate calculator 120 may detect a double-error syndrome to determine whether to terminate early.

) 및 반-반복 카운터 변수(

)를 초기화할 수 있으며, 여기서

는

의 조건을 만족하는 양의 정수일 수 있다.According to one side, the expectation reduction rate calculation unit 120 includes a tolerance counter value S set in advance for determining early termination, and a syndrome counter value before performing the half-iteration decoding process.

) And a half-repeated counter variable (

) Can be initialized, where

Is

It may be a positive integer that satisfies the condition of.

)을 각각 '0'으로, 반-반복 카운터 변수(

)를 '1'로 초기화할 수 있다. More specifically, the expected reduction rate calculation unit 120 includes a tolerance counter value S and a syndrome counter value corresponding to the double-error syndrome (

) As '0', and the half-repeated counter variable (

) Can be initialized to '1'.

)을 설정할 수 있다. According to one side, the expected reduction rate calculation unit 120 determines a threshold value of a size corresponding to the type of the component code of the received component code word (

) Can be set.

)은 구성 부호 종류에 따라 다른 크기를 가질 수 있으며, 이 값의 설정으로 인해 오류정정 성능과 복잡도 감소 효과가 달라질 수 있다.Specifically, the judgment threshold (

) May have different sizes depending on the type of constituent code, and the error correction performance and complexity reduction effect may vary due to the setting of this value.

)은 구성 부호의 길이 또는 부호율에 대응되는 크기로 설정될 수 있다. For example, the judgment threshold (

) May be set to a size corresponding to the length or code rate of the constituent code.

)을 설정할 수 있다. For a more specific example, the expected reduction rate calculation unit 120 may determine a threshold value having a relatively small size as the code length increases (

) Can be set.

According to one side, the expected reduction rate calculation unit 120 may increase the syndrome counter value by '1' each time a syndrome is detected in a half-iteration decoding process.

)만큼 수행되는 반-반복(half-iteration) 복호 과정에서 더블-에러 신드롬이 검출될 때 마다, 신드롬 카운터 값(

)을 '1' 씩 증가시킬 수 있다.Specifically, the expected reduction rate calculation unit 120 is the maximum number of half-repeated decoding (

Whenever a double-error syndrome is detected in the half-iteration decoding process performed by ), the syndrome counter value (

) Can be increased by '1'.

)을 이용하여 첫번째 반-반복 복호 과정을 통해 검출된 더블-에러 신드롬의 초기 검출 비율(

)을 산출할 수 있다. According to one side, the expected decrease rate calculation unit 120 is the syndrome counter value increased through the first half-repeated decoding process (

The initial detection rate of the double-error syndrome detected through the first half-repeated decoding process using) (

) Can be calculated.

)은

번째 반-반복 복호 과정을 통해 검출된 전체 신드롬들(노-에러 신드롬, 싱글-에러 신드롬 및 더블-에러 신드롬) 중에서 더블-에러 신드롬이 차지하는 비율을 의미할 수 있다. For example, the detection rate (

)silver

This may mean the ratio of the double-error syndrome among all syndromes (no-error syndrome, single-error syndrome, and double-error syndrome) detected through the second half-repeated decoding process.

)을 최대 반-반복 복호 수(

)로 나누어 기대 감소율(

)을 산출할 수 있다. According to one side, the expected reduction rate calculation unit 120 is the calculated initial detection rate (

) To the maximum number of half-repeated decodes (

) Divided by the expected reduction rate (

) Can be calculated.

)을 연산할 수 있다. Specifically, the expected decrease rate calculation unit 120 is the expected decrease rate for the double-error syndrome through Equation 1 below (

) Can be calculated.

[Equation 1]

는 기대 감소율,

은 초기 검출 비율,

는 최대 반-반복 복호 수,

는

번째 반-반복 복호 과정에 대응되는 검출 비율,

는 신드롬 카운터 값, n은 블록 터보 부호 각 행 또는 열 방향에서의 총 코드워드 개수를 나타낸다.here,

Is the expected rate of decline,

Is the initial detection rate,

Is the maximum number of half-repeated decodes,

Is

Detection ratio corresponding to the second half-repeated decoding process,

Denotes the syndrome counter value, and n denotes the total number of codewords in each row or column direction of the block turbo code.

More specifically, the reason for counting the number of double-error syndromes in the early termination apparatus 100 is that the factor that most adversely affects error correction is the double-error syndrome.

)의 크기만큼 더블-에러 신드롬의 검출 비율이 감소한다면, 결과적으로 반-반복 카운터가 최대 반-반복 복호 수(

)에 도달함에 따라 그 검출 비율이 0으로 수렴할 것으로 기대할 수 있다. That is, in the early termination device 100, as the half-repeated decoding process of the block turbo code proceeds, the expected reduction rate (

If the detection rate of double-error syndrome decreases by the amount of ), as a result, the half-repeated counter is the maximum number of half-repeated decodes (

), it can be expected that the detection rate will converge to zero.

The early termination determination unit 130 according to an embodiment calculates a difference in the detection rate of the syndrome in the half-repeated decoding process, and whether the half-repeated decoding process is prematurely terminated based on the calculated detection rate difference and the calculated expected reduction rate. Can judge.

)을 이용하여 m-1번째 검출 비율(

)을 산출하고, m번째 반-반복 복호 과정을 통해 증가된 신드롬 카운터 값(

)을 이용하여 m번째 검출 비율(

)을 산출하며, m-1번째 검출 비율(

)에서 m번째 검출 비율(

)을 감산하여 검출 비율차(

)를 산출할 수 있다. According to one side, the early termination determination unit 130 is the m-1th (here, m is a positive integer equal to or greater than 2) syndrome counter value (

) Using the m-1th detection rate (

) Is calculated, and the syndrome counter value is increased through the m-th half-repeated decoding process (

) Using the m-th detection rate (

) Is calculated, and the m-1th detection rate (

) To the m-th detection ratio (

) By subtracting the detection rate difference (

) Can be calculated.

)가 산출된 기대 감소율(

) 보다 작으면, 공차 카운터 값(S)을 증가시킬 수 있으며, 증가된 공차 카운터 값(S)이 기설정된 판단 임계값(

)과 일치하게 되면, 반-반복 복호 과정을 조기 종료시킬 수 있다. According to one side, the early termination determination unit 130 is the calculated detection ratio difference (

) Is the calculated expected reduction rate (

), the tolerance counter value (S) can be increased, and the increased tolerance counter value (S) is a predetermined judgment threshold (

If it matches ), the half-repeated decoding process can be terminated early.

)를 통해 m번째 반-반복 복호 과정 동안 오류정정으로 인해 더블-에러 신드롬의 검출 비율이 얼마나 줄어들었는지를 판단할 수 있다. That is, the early termination determination unit 130 is the calculated detection ratio difference (

), it is possible to determine how much the detection rate of the double-error syndrome has decreased due to error correction during the m-th half-repeated decoding process.

)를 통해 오류 정정이 얼마나 원활하게 수행되었는지를 확인할 수 있으며, 산출된 검출 비율차(

)를 기대 감소율(

)과 비교하여 검출 비율차(

)가 기대 감소율(

) 보다 작을 때마다 공차 카운터 값(S)을 증가시킬 수 있다. In other words, the early termination determination unit 130 is the calculated detection ratio difference (

), you can check how smoothly the error correction was performed, and the calculated detection ratio difference (

) To the expected reduction rate (

) And the detection rate difference (

) Is the expected reduction rate (

Whenever it is smaller than ), the tolerance counter value (S) can be increased.

)보다 작다는 것을 의미하며, 결국 예상되는 기대 감소율(

) 보다 오류 정정이 원활하게 이루어지지 않았음을 의미한다. This means that the actual reduction in the detection rate of double-error syndrome in the continuous half-repeated decoding process is the expected reduction rate (

), which means that the expected decrease rate (

), it means that error correction was not made smoothly.

)에 도달하게 되면, 오류정정 과정을 신속하게 종료시킬 수 있다. Accordingly, the early termination determination unit 130 may increase the tolerance counter value S by '1' each time the above-described situation is repeated in the semi-repetitive decoding process, and the tolerance counter value increased by '1' ( S) is a predetermined judgment threshold (

) Is reached, the error correction process can be quickly terminated.

)을 설정하는 이유는 블록 터보 부호가 부호 길이에 따라, 그리고 채널 환경에 따른 오류 분포에 따라 복호 수렴 속도가 달라질 수 있기 때문이다. In the early termination apparatus 100 according to an embodiment, the tolerance counter value (S) and the determination threshold value (

) Is set because the decoding convergence speed may vary according to the code length of the block turbo code and the error distribution according to the channel environment.

For example, when the length of the block turbo code is relatively small, the decoding convergence speed may be relatively slow.

In other words, the shorter the length of the constituent code used for code formation, the slower the effect of error correction appears in the initial half-repeated decoding process, and then the convergence speed increases sharply in the subsequent half-repeated decoding process, resulting in the possibility of success in error correction. exist.

)가 기대 감소율(

) 보다 작을 때마다 조기 종료를 결정하면, 조기 종료 장치(100)의 오류정정 성능의 저하를 야기할 수 있다. Therefore, the detection ratio difference (

) Is the expected reduction rate (

If it is determined that the early termination is smaller than ), the error correction performance of the early termination apparatus 100 may be deteriorated.

)을 설정하며, 공차 카운터 값(S)의 증가 이벤트가 여러 번 발생되어 공차 카운터 값(S)이 판단 임계값(

)까지 도달하게 되면 오류정정 성능의 저하 없이 조기 종료를 수행할 수 있다. Thus, the early termination device 100 is the tolerance counter value (S) and the determination threshold (

), and the tolerance counter value (S) is judged by the threshold value (

), early termination can be performed without deteriorating the error correction performance.

That is, the early termination apparatus 100 may perform effective early termination by adding only a very simple arithmetic operation using syndrome detection information already included in the existing Chase-Pyndiah algorithm.

In other words, the early termination device 100 can effectively perform early termination with only a simple arithmetic operation based on the syndrome detection information already included in the existing Chase-Pyndiah algorithm, thereby minimizing the amount of additional computation for early termination. The complexity of decoding calculation can be greatly reduced.

FIG. 2 is a diagram illustrating a detection rate of each syndrome according to an increase in the number of half-repeated decoding in a process of performing error correction using the existing Chase-Pyndiah algorithm.

Referring to FIG. 2, (a) to (d) of FIG. 2 show the detection results of each syndrome when the signal to noise ratio (SNR) is 1.0 dB, 2.75 dB, 3.25 dB, and 3.5 dB, respectively. In this case, 1.0 dB is the area where the bit error ratio (BER) is 10 ^-1 , 2.75 dB is the area where the bit error rate is 10 ^-2 , 3.25 dB is the area where the bit error rate is 10 ^-4 , and 3.5 dB is It may correspond to a region having a bit error rate of 10 ^-6 .

In addition, in FIGS. 2A to 2D, the x-axis represents the half-interation number, and the y-axis represents the syndrome detection ratio of each syndrome.

In addition, the simulation process for calculating the results of FIGS. 2A to 2D consisted of block turbo codes using the most commonly used extended Hamming codes, and the codes used were (64,57). ,4) It is a ^2- block turbo code, and the maximum number of half-repeated decoding is set to 8 times and the minimum number of reliability bits is set to 4.

Meanwhile, since the extended Hamming codes are used as constituent codes in FIGS. 2A to 2D, the types of syndromes detectable during the decoding process are no-error syndrome and single-error. It can be classified into syndrome and double-error syndrome.

In general, block turbo codes maximize their effect by repeatedly performing the same error correction process like other error correction codes.

The maximum number of repetitive decoding in the field of error correction can be set differently depending on the type of code, code length, and code rate, and generally 4 to 6 times in the block turbo code. Here, one repetitive decoding may have the same meaning as two semi-repetitive decoding.

The reason for defining the maximum number of repetitive decoding is to prevent an excessive increase in decoding computation complexity.

In other words, if the initial error rate of the block turbo codeword is high, and if the maximum number of repetitive decoding is not determined, the decoding process continues without ending and the computational complexity may be continuously increased, which is known in the prior art. This means that early termination techniques are unlikely to be effectively applied.

According to (a) to (d) of FIG. 2, it can be seen that in the first half-repetition process, the single-error syndrome and the double-error syndrome occupy most of the total ratio. In particular, in the low SNR region, the degree is close to 100%, which means that error bits exist in most of the constituent code codewords.

Subsequently, while continuously performing the iterative decoding process, the ratio of single-error syndrome and double-error syndrome gradually decreases, and the ratio of no-error syndrome increases accordingly.

Intuitively, this can be thought to be because the number of errors in the code gradually decreases due to the execution of the Chase-Pyndiah algorithm, but the degree of decrease and increase of the ratios differs depending on the SNR area.

For example, it can be seen that the slope of the reduction ratio of single-error syndrome and double-error syndrome is very steep in the relatively high SNR region of 3.25 dB and 3.5 dB region. Therefore, it can be expected that error correction will succeed with a very high probability before reaching the maximum number of half-repeated decoding.

In addition, if conventional early termination techniques known during the decoding process are applied, error correction is terminated before the maximum number of half-repeated decoding is reached, thereby reducing the complexity of the entire decoding calculation.

However, in the relatively low SNR region and the intermediate SNR region of 1.0 dB and 2.75 dB region, it can be seen that although the error syndromes are somewhat reduced, the degree of reduction is not sufficient.

In other words, even if error correction is sufficiently performed up to the maximum number of half-repeated decoding, it can be confirmed that the detection rate of the no-error syndrome does not reach 100%, which means that there is a high possibility that error correction will fail as a result. do.

In such a case, it is meaningless to perform error correction as many as the maximum number of half-repeated decodings previously defined, and if the possibility of error correction failure is predicted in advance before the maximum number of half-repeated decodings is reached, it is a little earlier. It may be more reasonable to terminate the error correction process at the timing and request retransmission to the transmitter.

Of course, in general, when constructing a communication system, the data transmission power is sufficiently considered in consideration of the channel environment, but the initial error rate in the error correction code codeword is excessively large due to factors such as the fading effect of the wireless communication channel. There are plenty of possibilities. In other words, it may cause an unintended increase in complexity of decoding calculations, depending on the communication channel environment characteristics.

On the other hand, the early termination apparatus according to an embodiment considers the possibility of continuous error correction failure by using the syndrome detection result of each constituent code codeword during the half-repeated decoding process of the block turbo code, and the possibility of error correction failure is very high. If it is determined that it is high, the decoding process can be terminated in advance, and through this, the total decoding computation complexity can be effectively reduced by reducing the average number of half-repetitive decoding.

3 is a diagram for explaining error correction performance according to a determination threshold setting in an early termination apparatus according to an embodiment.

3, (a) to (c) of FIG. 3 show the bit error rates when the block turbo codes are (64,57,4) ² , (128,120,4) ² and (256,247,4) ² , respectively. bit error ratio, BER).

Specifically, in FIGS. 3A to 3C, the x-axis represents a signal to noise ratio (SNR), and the y-axis represents the bit error rate.

In addition, in (a) to (c) of Fig. 3,'CPA' represents a simulation result based on the known Chase-Pyndiah algorithm, and'S_max_2','S_max_3' and'S_max_4' are early termination according to an embodiment. The simulation results when the judgment threshold values of the device are '2', '3' and '4', respectively are shown.

Meanwhile, in all simulation processes according to FIGS. 3A to 3C, the maximum number of half-repeated decoding is set to 8 times and the minimum number of reliability bits is set to 4 in common.

According to FIGS. 3A to 3C, it can be seen that the determination threshold should be set to different values according to signs.

Specifically, in the case of the block turbo code of (64,57,4) ² , it can be confirmed that the size of the determination threshold must be set to at least '4' or more to maintain error correction performance without deterioration.

If a judgment threshold of less than '4' is used in the block turbo code of (64,57,4) ² , it may be possible to reduce the average number of half-repeated decoding, but the error correction performance deteriorates accordingly. Can be.

On the other hand, in the case of the block turbo code of (256,247,4) ² , it can be seen that a smaller decision threshold can be used. Specifically, it can be seen that the block turbo code of (256,247,4) ² does not affect the error correction performance even if the determination threshold is set to '2'.

As a result, according to the simulation result, the determination threshold value according to an embodiment may set a determination threshold value of a relatively small size as the length of the block turbo code increases, and the early termination apparatus according to an embodiment sets the determination threshold value. As a result, it can be confirmed that the deterioration of the error correction performance can be completely prevented and the computational complexity can be reduced.

FIG. 4 is a diagram illustrating a measurement result of an average number of half-repeated decoding according to a determination threshold setting in an early termination apparatus according to an exemplary embodiment.

Referring to FIG. 4, in (a) to (c) of FIG. 4, the average half when the block turbo codes are (64,57,4) ² , (128,120,4) ² and (256,247,4) ^{2 respectively-} Represents the number of repetitive decoding.

Specifically, in (a) to (c) of FIG. 4, the x-axis represents a signal to noise ratio (SNR), and the y-axis represents the average number of half-repeat decoding.

In addition, in (a) to (c) of Fig. 4,'CPA' represents a simulation result based on the known Chase-Pyndiah algorithm, and'S_max_2','S_max_3' and'S_max_4' are early termination according to an embodiment. The simulation results when the judgment threshold values of the device are '2', '3' and '4', respectively are shown.

Meanwhile, in all simulation processes according to FIGS. 4A to 4C, the maximum number of half-repeated decoding is set to 8 times and the minimum number of reliability bits is set to 4 in common.

According to (a) to (c) of FIG. 4, it can be seen that the size of the average number of half-repeated decoding varies according to the size setting of the determination threshold in the low and intermediate SNR regions, and the size of the value is It can be seen that the smaller it is, the greater the complexity reduction effect.

Based on the error correction performance result described with reference to FIG. 2, the read threshold applicable to each block turbo code is (64,57,4) ² to '4', (128,120,4) ² to '3', (256,247) ,4) can be from ² to '2'.

으로 표현될 수 있으며, 여기서

는 사전에 정의되어 사용하는 최대 반-반복 복호의 총 횟수, n은 코드워드 길이, p는 소프트 값들 중 크기가 상대적으로 작은 비트를 나타낸다. If error correction is performed based on the known Chase-Pyndiah algorithm, the complexity of the entire decoding calculation is

Can be expressed as, where

Denotes the total number of maximum half-repeated decoding that is predefined and used, n denotes a codeword length, and p denotes a bit having a relatively small size among soft values.

의 크기가 변화하게 되고, 결국 일실시예에 따른 조기 종료 장치에 따른 조기 종료 기법의 적용 전후를 기준으로 하는 복호 계산의 복잡도 감소 효과는 (64,57,4)²에서 최대 약 25.75%, (128,120,4)²에서 최대 약 45.5%, (256,247,4)²에서 최대 약 66.125% 인 것을 확인할 수 있다.Therefore, the early termination device according to an embodiment is not only in the high SNR area but also in other areas

The size of is changed, and as a result, the effect of reducing the complexity of the decoding calculation based on before and after the application of the early termination technique according to the early termination device according to an embodiment is (64,57,4) ² to a maximum of about 25.75%, ( 128,120,4) can be confirmed to be in ² up to about 45.5%, (256,247,4) from ² up to about 66.125%.

5 is a diagram for describing an early termination method according to an embodiment.

In other words, FIG. 5 is a diagram illustrating a method of operating an early termination device according to an embodiment described with reference to FIGS. 1 to 4, and overlaps with the content described with reference to FIGS. 1 to 4 among the contents described with reference to FIG. Description will be omitted.

Referring to FIG. 5, in step 510, in the early termination method according to an embodiment, a codeword receiver may receive a constituent code codeword input to a decoder.

Next, in step 520, in the early termination method according to an embodiment, in the expected reduction rate calculation unit, the syndrome of the configuration protection codeword received in the half-iteration decoding process performed a predetermined number of times. It is possible to detect, count the number of detected syndromes, and calculate an expected reduction rate of the syndrome based on the counted number of syndromes.

Next, in step 530, in the early termination method according to an embodiment, the early termination determination unit calculates a difference in detection rate of the syndrome in the semi-repetitive decoding process, and based on the calculated detection rate difference and the calculated expected reduction rate, -It is possible to determine whether the iterative decoding process is terminated early.

Consequently, using the present invention, it is possible to determine whether or not to terminate the decoding process early by using the syndrome information of the constituent code codeword input to the block turbo code decoder.

In addition, the present invention can effectively reduce the average number of half-repeated decoding without deteriorating the error correction ability by using the degree of reduction in the charge ratio of the double-error syndrome as the half-repeated decoding process proceeds.

As described above, although the embodiments have been described by the limited drawings, various modifications and variations are possible from the above description to those of ordinary skill in the art. For example, the described techniques are performed in a different order from the described method, and/or components such as a system, structure, device, circuit, etc. described are combined or combined in a form different from the described method, or other components Alternatively, even if substituted or substituted by an equivalent, an appropriate result can be achieved.

Therefore, other implementations, other embodiments, and claims and equivalents fall within the scope of the claims to be described later.

100: early termination device 110: codeword receiver
120: expected reduction rate calculation unit 130: early termination determination unit

Claims (11)

A codeword receiving unit for receiving a constituent code codeword input to a decoder;
No-error syndrome, single-error syndrome, and double-error of the received constituent code code word in the half-iteration decoding process performed a predetermined number of times Detect at least one syndrome among double-error syndromes, count the number of the detected at least one syndrome, and the double-error in the detected at least one syndrome based on the counted number of syndromes. An expected reduction rate calculation unit that calculates a detection ratio, which is a ratio occupied by a syndrome, and calculates an expected reduction ratio of the double-error syndrome based on the calculated detection ratio; and
m-1th (where m is a positive integer greater than or equal to 2) detection by subtracting the m-th detection ratio calculated through the m-th half-repeated decoding process from the m-1th detection ratio calculated through the half-repeated decoding process An early termination determination unit that calculates a ratio difference and determines whether or not the half-repeated decoding process is prematurely terminated based on the calculated detection ratio difference and the calculated expected reduction rate
Early termination device comprising a. The method of claim 1,
The constituent code codeword is an error correction code word based on a block turbo code (BTC).
Early termination device. delete The method of claim 1,
The expected reduction rate calculation unit,
Initializing a tolerance counter value, a syndrome counter value, and a half-iteration counter variable for determining whether to terminate the early stage before performing the half-iteration decoding process
Early termination device. The method of claim 1,
The expected reduction rate calculation unit,
Setting a determination threshold corresponding to at least one of a type of a constituent code, a code length, and a code rate of the received constituent code codeword
Early termination device. The method of claim 4,
The expected reduction rate calculation unit,
Increasing the syndrome counter value by '1' whenever the at least one syndrome is detected in the half-iteration decoding process
Early termination device. The method of claim 6,
The expected reduction rate calculation unit,
The initial detection rate is calculated using the syndrome counter value increased through the first half-repeated decoding process, and the expected reduction rate is calculated by dividing the calculated initial detection rate by the maximum number of half-repeated decoding.
Early termination device. The method of claim 6,
The early termination determination unit,
The m-1th detection rate is calculated using the syndrome counter value increased through the m-1th half-repeated decoding process, and the syndrome counter value increased through the m-th half-repeated decoding process is used. to calculate the m-th detection rate
Early termination device. The method of claim 4,
The early termination determination unit,
If the calculated detection ratio difference is less than the calculated expected reduction rate, increasing the tolerance counter value
Early termination device. The method of claim 9,
The early termination determination unit,
If the increased tolerance counter value coincides with a preset determination threshold, early termination of the half-repeated decoding process
Early termination device. Receiving, at a codeword receiving unit, a constituent code codeword input to a decoder;
In the expected reduction rate calculation unit, a no-error syndrome and a single-error syndrome of the received constituent code codeword in a half-iteration decoding process performed a predetermined number of times. syndrome) and double-error syndrome, counting the number of the detected at least one syndrome, and the detected at least one syndrome based on the counted number of syndromes. Calculating a detection ratio, which is a ratio occupied by the double-error syndrome in syndrome, and calculating an expected reduction rate of the double-error syndrome based on the calculated detection ratio; and
In the early termination determination unit, the m-th (where m is a positive integer greater than or equal to 2) the m-th detection rate calculated through the half-repeated decoding process, the m-th calculated through the m-th half-repeated decoding process Calculating a detection ratio difference by subtracting the detection ratio, and determining whether or not the half-repetitive decoding process is prematurely terminated based on the calculated detection ratio difference and the calculated expected reduction rate
Early termination method comprising a.

Images