KR20060129538A - Local erasure map decoder - Google Patents

Abstract

The present invention relates to a method for decoding at least one codeword, the at least one codeword having been generated by an encoder comprising a structure providing a code representable by a set of branch transitions in a trellis diagram. Further, the present invention provides a respective decoder, as well as a mobile station and a base station in a communication network employing the decoder. Moreover a communication system comprising the base stations and mobile stations is provided. To reduce the influence of wrong information in a decoding process the present invention suggests using only a subset of reliable information in the forward and/or backward recursion of a Maximum A-Posteriori (MAP) Algorithm or Max-Log-MAP Algorithm.

Description

Local erasure map decoder {LOCAL ERASURE MAP DECODER}

The present invention decodes at least one codeword generated by an encoder comprising a structure that provides a code that can be described as a set of branch transitions in a trellis diagram. It is about a method. The present invention also provides each decoder and a mobile station and base station in a communication network using the decoder. There is also provided a communication system including the base station and the mobile station.

The convolutional code and its associated code may be generated by one or more cascaded or concatenated shift registers. For simplicity, the binary shift register is considered below. The binary shift register can take the value of binary zero or binary one. When a shift occurs, the contents of each register are sent to subsequent registers to create new content. Typically the input to this encoder is used as the new contents of the first register.

The output of the binary shift register encoder is typically obtained by modulo binary addition of multiple shift register contents before being shifted. A simple binary shift register encoder is shown in FIG. 1, where the number of shift registers r = 2 and the number of states M = 4. Each shift register is denoted by D, and each modulo double addition unit is denoted by "+". Two output bits are obtained from one input bit, the first output bit is the same as the input bit (upper branch), and the second output bit is obtained by modulo two addition of the shift register state and the input bit (lower branch). .

2 shows a state transition diagram of the encoder of FIG. 1. Each state is represented by a value in the shift register. Each transition is indicated by an arrow. Transitions made by zero input bits are indicated by dashed lines, and transitions made by one input bit are indicated by solid lines. Each line is marked with an output bit corresponding to the input bit. Another notation of state transition diagram is trellis, which consists of trellis elements as shown in FIG. 3. Other details related to shift register encoding (also known as convolutional encoding) are disclosed, for example, in Lin et al., "Error Control Coding: Fundamentals and Applications," Prentice-Hall Inc., chapter 10.

Shift registers are typically used for convolutional codes. Recently, this has also been used for 'turbo codes' which have achieved very low error rates, making turbo codes very useful for communication systems.

More common decoding algorithms for shift register codes are the Viterbi algorithm and the maximum a-posteriori (MAP) algorithm. The former is often used for conventional convolutional codes, but the latter is very useful for decoding turbo codes due to its soft inductive probability output.

Maximum Inductive Algorithm

The maximum inductive algorithm is briefly described below. For simplicity, the case of binary is described in detail. Expanding to a non-binary case would be no problem for those skilled in the art. In general, non-binary event probabilities cannot be expressed by log-likelihood ratios. Instead, several (if possible login) absolute probability measures can be used. Obviously, all subsequent equations containing log probability ratios should be changed while maintaining the absolute probability measures mentioned above.

The simple characteristic of the binary case is that since only two events are possible, the event probability can be expressed in terms of log probability ratio (LLR), which is usually a natural log of the ratio of the probability

Where χ is one of two possible events.

In this specification, the following symbols are used.

This algorithm typically has two components: forward and reverse recursion. More specifically, two distributions α _k and β _k are updated as they cycle. The quantity α _k (s _k ) represents the probability measure that exists in the state S _k = m for the information bit k given the received sequence y ₁ ... y _k , and β _k (S _k ) is Given a received sequence y _k ... y _K , it represents a probability measure that exists in state S _k = m for information bit k.

Of the two cycle is the so-called branch transition probabilities r _k, is defined on the basis of _{i (y k, m ',} m "). This is assuming that the information that caused the transition bit d _k = i, the received codeword y _k Given the observations, it represents the probability of transition between states m 'and m ". The branch transition probability can be calculated as follows.

q | value of (d _k = i S _{k -1,} S _k) is a 1 or a 0, which is determined according to whether they are associated with a transition to state S _k from the i-bit state S _{k -1.} Pr {S _k | S _{k −1} } is a prior estimation probability of the information bit d _k . In terms of turbo decoding, this probability may be external information obtained from other decoders. Other terms will be apparent to those skilled in the art. For example, this probability is set to zero when no preliminary estimation information is available.

Assuming that the r value exists only for these transitions, Equation 2 can be simplified by excluding the index i, where q (d _k = i | S _{k −1} , S _k ) = 1. Using this assumption, this equation can be rewritten as

가 존재하는 경우를 생각해 보면, 식 3은 더 간략화될 수 있으며, 이로써 코드 레이트는 1/2이 된다. 또한, 2진의 경우를 생각해 보면, 식 3은 로그 표현을 사용해서 더 간략화될 수 있다. For each information bit d _k of each encoder input, two coded bits at the output of the encoder

Considering the case in which Equation 3 exists, Equation 3 can be further simplified, so that the code rate is 1/2. Also, considering the binary case, Equation 3 can be simplified further using a logarithmic representation.

For binary shift register codes, the number of states M is

It can be calculated as

reset

For each branch transition starting at state S _{k -1} and ending at state S _k , the branch transition probability for Binary Phase Shift Keying (BPSK) Additive White Gaussian Noise (AWGN) is

Where k is from 1 to K.

Since the last term is often used as

Can be represented again as

Is to use

L _c is the channel scaling factor that can be obtained from the signal-to-noise ratio (SNR), in which case

Where σ ² represents the channel noise variance.

The initial values of α _k and β _k can be initialized according to the system parameters. For code starting and ending in state m = 0, the initialization is

And

Will be.

Forward circulation

For each state S _k , k is from 1 to K and α _k is

Is calculated accordingly.

Reverse circulation

For each state S _k , k is from K-1 to 0 and β _k is

It can be calculated as

decoding

The entire decoding process consists of applying forward and backward cycling. After these cycles, the soft output decision (ie, inductive probability) of each information bit can be updated.

에 대응하는 순서쌍(m',m")의 세트이다. 유사하게 S^-는 d_k=0인 경우에 대해 정의된다. In the above equation, S ⁺ is any state transition caused by data input d _k = 1

Is a set of ordered pairs (m ', m ") corresponding to. Similarly, S ⁻ is defined for the case where d _k = 0.

Using equation 15, the value of the kth transmitted bit is

Can be estimated as

Note that the extraneous amount L ^e obtained in equation 14 can be used as the unique information of the subsequent decoder. Similarly, the amount L ⁱ of equation 15 could be obtained as unique information from foreign information of another decoder.

Those skilled in the art will appreciate that these two quantities may be set to appropriate values when no information is available from other decoders. For a detailed description of whether this algorithm can be applied to turbo code, unique information and foreign information, see Berrou et al., "Near Shannon Limit Error-Correcting Coding and Decoding: Turbo Codes (1)", Proc. IEEE Int. Conf. On Communications, pp. 1064-1070, May 1993.

Max Log MAP Algorithm

To simplify complex calculations, equations 12 and 13 can be simplified and replaced by:

And

Similarly, the judgment variable

Can be obtained by modifying

However, this approximation can degrade the decoding performance.

As can be seen from the equations of the forward and backward cycles, information is ultimately included from a number of values derived from the received vector corresponding to the transmitted codeword. In a noisy channel environment, multiple reception values are likely to carry false information, which means that they can be derived from these values and enter the decoding iteration process.

Accordingly, it is an object of the present invention to reduce the effects of this misinformation.

This object is solved by the subject matter of the independent claims. Advantageous embodiments of the invention are the subject matter of the dependent claims.

According to one aspect of the present invention, neither the forward and / or reverse circulation information required by the prior art equations is processed. Rather, according to this embodiment of the invention, some of these terms are excluded. The determination of which term is excluded can be determined, for example, according to its reliability. That is, the term used to determine the forward and / or reverse cycle to degrade decoding performance is omitted in each equation.

In one of the other embodiments of the present invention, a method of decoding at least one code word in a decoder is provided, wherein the at least one code word provides a structure that provides a code that can be represented as a set of branch transitions in a trellis diagram. It was generated by the containing encoder.

According to this embodiment, the method comprises the steps of initializing a set of branch transition probabilities of the decoder based on the received codeword and encoder structure and according to an initial state of the encoder used to encode the at least one codeword. Initializing a first probability distribution and a second probability distribution, recalculating a value of the first probability distribution using a cyclic algorithm based on a set of initial values of the first probability distribution and a branch transition probability, and a second Recalculating, using a cyclic algorithm, the value of the second probability distribution based on the initial value of the probability distribution and the set of branch transition probabilities, and the set of received codeword and branch transition probabilities, the first and second probability distributions. Reconstructing the decoded codeword based on the foreign probability measure computed based on.

In one or both of the steps of calculating the value of the first or second probability distribution, a subset and branch transition probability of the initial value of each of the first probability distribution or the second probability distribution in recalculating each probability distribution. A subset of the set of can be used. In addition, only a subset of values that meet a predetermined reliability criterion are used.

The encoder may also be described by a shift register structure that includes at least one of a feed-forward arithmetic operation and a feed-back arithmetic operation.

Also, in another embodiment of the present invention, the code is suitable for decoding using a maximum inductive algorithm.

In another embodiment of the present invention, the method further comprises initializing a set of branch transition probabilities using inherent probability measures.

Another embodiment of the present invention further includes the step of reconstructing the decoded codeword using the unique probability measure.

In another variation of this embodiment, a decoder, which can be described as two separate decoder instances, is used to decode at least one codeword in the first decoding step, the method being used to determine the foreign probability measure of the first decoder instance. And using as a unique probability measure of the two decoder instances.

In another variation of this embodiment, the method further includes performing a second decoding iteration at the first decoder instance, wherein the decoder instance uses the foreign probability measure of the second decoder instance as an intrinsic probability measure.

According to another embodiment of the invention, the reliability criterion is based on the channel estimate of the radio channel in which at least one codeword was received, the absolute value of the elements of the first and / or second probability distribution, the number of decoding steps performed and the random process. It can be made based on at least one of. In another variant, if the signal-to-noise ratio of this element and / or its absolute value is below a predetermined threshold, the reliability criterion cannot be satisfied by the element of the first or second probability distribution.

The invention also provides, in another embodiment, a decoder for decoding at least one codeword, wherein the at least one codeword comprises a structure for providing a code which can be described by a set of branch transitions in a trellis diagram. Is generated by the encoder.

The decoder initializes a set of branch transition probabilities of the decoder based on the received codeword and encoder structure, and according to the initial state of the encoder used to encode the at least one codeword, the first probability distribution and the second probability distribution. Initialize the first probability distribution based on the initial value of the first probability distribution and the branch transition probability and recalculate the value of the first probability distribution using a cyclic algorithm, Recalculate a value of the second probability distribution based on a cyclic algorithm, and decode the code based on a set of received codeword and branch transition probabilities, and a foreign probability measure calculated based on the first and second probability distributions. Processing means for reconstructing the word.

In addition, the processing means may, in one or both of recalculating the values of the first or second probability distributions, recalculate each of the probability distributions to serve as the sub-values of the initial values of each of the first probability distribution or the second probability distribution. A subset of the set of set and branch transition probabilities can be used, where only values of the subset that meet the predetermined confidence criterion are used.

In another embodiment of the present invention, a decoder is provided that includes means suitable for performing one of the decoding methods described above.

Further, another embodiment of the present invention relates to a mobile terminal of a mobile communication system, wherein the mobile terminal comprises: receiving means for receiving at least one codeword, demodulation means for demodulating at least one received codeword; And a decoder according to one of the embodiments of the present invention.

In another embodiment, the mobile terminal may further comprise encoding means for encoding data of at least one codeword and transmitting means for transmitting at least one codeword, wherein the at least one transmitted codeword is described above. It is suitable for decoding according to the decoding method.

In another embodiment of the present invention, there is provided a base station of a mobile communication system, the base station comprising: receiving means for receiving at least one codeword, demodulation means for demodulating at least one received codeword, and implementation of the invention It may include a decoder according to one of the examples.

In another embodiment, the base station may comprise encoding means for encoding data of at least one codeword and transmission means for transmitting at least one codeword, wherein the at least one transmitted codeword is described above. It is suitable for decoding according to the decoding method.

In addition, according to another embodiment, there is provided a mobile communication system comprising at least one base station according to an embodiment of the present invention and at least one mobile terminal according to an embodiment of the present invention.

In the following, the invention will be described in more detail with reference to the accompanying drawings. Similar or corresponding parts in the drawings are indicated by the same reference numerals.

1 illustrates an exemplary arrangement of a shift register encoder for systematic encoding;

2 is a state transition diagram of the encoder shown in FIG.

3 is a trellis partial view of the encoder shown in FIG. 1;

4 shows a trellis portion showing the parameters of the forward circulation, FIG.

5 is a diagram showing a trellis portion showing a variable of reverse circulation;

6 is a diagram showing a trellis portion showing a change in determination;

7 is a flowchart of a decoding process according to an embodiment of the present invention;

8 and 9 are flowcharts of a decoding process using the turbo principle according to another embodiment of the present invention;

10 illustrates a transmitter and receiver unit according to an embodiment of the present invention;

11 illustrates a mobile terminal according to an embodiment of the present invention, including the transmitter and receiver shown in FIG. 10;

12 shows a base station according to an embodiment of the present invention, including the transmitter and receiver shown in FIG. 10;

FIG. 13 is a schematic diagram illustrating a communication system according to an embodiment of the present invention including the mobile terminal shown in FIG. 11 and the base station (Node B) shown in FIG. 12;

은 "x는 세트 B가 없는 세트 A의 요소이다"라는 뜻으로 "x는 세트 A의 요소이지 세트 B의 요소는 아니다"라는 뜻이다. In the following description, the expression

Means "x is an element of set A without set B" and "x is an element of set A, not an element of set B".

As mentioned in the above description, the equations may be used in the decision phase of initialization, forward iteration, backward iteration and maximum inductive algorithm.

In general, such an equation includes the following terms.

The expression for initialization contains terms that contain the y value.

The equation of the forward cycle contains a term containing Γ and the determined α value.

The equation of the reverse cycle contains a term containing Γ and the determined β value.

The molecule in equation 12 of the forward cycle can be interpreted as the sum of the values of the state transitions starting at state S _{k -1} and ending at S _k = m. Thus, the next "forward set" can be defined.

T _{k, m} is the set of states S _{k -1,} S _{k -1,} where the state transition from S to _k is available by the information bit d _k.

therefore,

Similarly, the molecule of equation 13 in the reverse cycle can be interpreted as the sum of the values of the state transitions starting at state S _{k +1} and ending at S _k = m. Thus, the next "reverse set" can be defined.

U _{k, m} is a set of state S _{k +1,} where transition to the S _{k +1} from S _k is available by the information bit d _k.

therefore,

to be.

According to the invention, the exclusion sets Δ _{k, m} and Ω _{k, m} can be further defined for the forward and / or reverse circulation.

The exclusion set Δ _{k, m} represents an element in the forward set T _{k , m} that does not satisfy a certain reliability criterion and thus cannot be used in the forward circulation step. Similarly, the exclusion set Ω _{k, m} represents a reverse set element that does not satisfy a certain reliability criterion in U _{k , m} and thus cannot be used in the reverse cycle step.

By using the exclusion sets Δ _{k, m} and Ω _{k, m} , this equation can be modified as follows.

New forward rotation

Or in other ways

Can be simplified.

New reverse rotation

Or in other ways

Can be simplified.

If the sets Δ _{k, m} and Ω _{k, m} are both empty, they are the same as before. If the exclusion set Δ _{k, m} contains elements such as the forward set T _{k ,} m, the value of α _k (S _k = m) cannot be determined from the circulation.

In this case, it can be used to set the corresponding α _k (S _k = m) = − ∞. Similarly, if the exclusion set Ω _{k, m} includes elements such as the reverse set U _{k , m} , then β _k (S _k = m) = − ∞ may be set.

For the value of a particular k, if the exclusion set is equal to the forward set for all m = 1 ... M, then a _k (m) can be set to -lnM, which means that all states S _k = 1 .. Means that .M is the same. This also applies to the reverse set.

In general, the exclusion set may depend, for example, on the state index m in which the expression is used, the information bit index k in which the expression is used and / or the number of iterations of the decoding process (eg in turbo decoding).

Definition of exclusion set Δ _{k, m} and Ω _{k, m}

As described above, the exclusion sets Δ _{k, m} and Ω _{k, m} can be defined to exclude data that is assumed to be wrong or likely to be wrong in the equation (or decoding process). If such data is included, the generated output is likely to be wrong. Therefore, the present invention proposes to overcome this adverse effect on the decoding output by ignoring these values in the equation.

As described above, the exclusion set for the new forward step (see Equation 26 or Eq. 27) and the reverse recursion step (see Eqs. 28 or 29) is defined such that untrusted messages are excluded from the calculation. In another embodiment of the invention, the exclusion sets can be defined independently of one another, ie, the elements of the exclusion set Δ _{k, m} need not necessarily be elements of the exclusion set Ω _{k, m} .

Similarly, in another embodiment of the present invention, the exclusion sets Δ _{k, m} and Ω _{k, m} can be set independently in the decoding iteration. When increasing the number of repetitions, the overall reliability of the message sent can be increased to a fairly good transmission state. This can be applied, for example, to the decoding of turbo codes, where foreign information exchanged between decoding entities typically increases in reliability as the number of decoding iterations increases.

Thus, as the number of iterations increases, the number of elements in the exclusion set is reduced so that the step of decoding the exclusion set late (in terms of repetition) may be empty.

In another embodiment of the present invention, this exclusion set may, for example, vary depending on the number of iterations processed so far as well as the maximum number of decoding iterations, which may be a parameter given by the communication system. This allows the elements of the exclusion set to be gradually reduced as the iteration step proceeds.

Exemplary lists of possible criteria that can be used independently or in combination to define the exclusion set are channel estimation (signal-to-noise ratio), absolute LLR values, number of repetitions (in turbo decoding) and / or random processes.

For example, through the channel evaluation criteria, the exclusion set may be defined according to the quality of the received data to be identified. The channel assessment has the advantage of evaluating the reliability of the received coding information by providing a series of independent side information known at the decoder. However, the granularity of channel estimation is limited to segments of multiple bits to define the exclusion set, which cannot be applied alone in all situations.

The absolute LLR value criterion allows for reliable evaluation of fine precision. According to the definition of the LLR value, a large absolute value indicates a high accuracy. On the contrary, a small absolute value indicates low accuracy. Thus, the magnitude of an absolute LLR value can be used to define the minimum value of a given expression that becomes part of the exclusion set. For example, the LLR value criterion may be used alone or in combination with other criteria to determine the elements of the exclusion set.

Another possible criterion may be a random access criterion. This criterion can be used alone or in combination with other criteria to determine the members of the exclusion set. For example, by channel assessment, it may be assumed that 10% of the received information is not reliable. Thus, for some of each piece of information, the probability of becoming an exclusion set can be 10%.

In the following, another embodiment of the present invention will be described with reference to FIGS. 7, 8 and 9.

7 is a flowchart of a decoding process according to an embodiment of the present invention. Upon receiving the step codeword over the air interface in step 701, the decoder in step 702 may generate the exclusion sets Δ _{k, m} and Ω _{k, m} .

In order to generate the exclusion set, a number of different decision parameters may be used to determine which elements should be excluded from the calculation of the forward and / or reverse cycle steps 704 and 705. For example, the receiving means may provide information about the codeword or the channel quality of receiving the individual bits of the codeword, or may provide the decoder with the exclusion sets Δ _{k, m} and Ω _{k, m} .

Further, based on the knowledge of the encoder structure and the received codeword y _k in step 703, the branch transition probability Γ (y _k , S _{k −1} , S _k ) may be initialized. Further, in step 704, probability distributions alpha _k and beta _k are initialized. This can be done, for example, using knowledge of the encoder structure used to generate the received codeword y _k .

By properly initializing the decoder, the forward and backward cycles defined in equations 26-29 can be performed in steps 705 and 706. In this cycle, the exclusion sets Δ _{k, m} and Ω _{k, m} are taken into account, ie only a subset of the values of the distributions α _k and β _k and / or Γ (y _k , S _{k -1} , S _k ) Can be used to perform the steps.

및 디코딩된 코드워드의 개개의 비 트

에 대한 평가 기준 L(d_k)을 생성하는 단계를 포함할 수 있다. When recalculating new values of α _k and β _k, the codeword may be reconstructed by the decoder. For example, this step is foreign

And individual bits of the decoded codeword

Generating an evaluation criterion L (d _k ) for.

또는 평가 기준 L(d_k)을 후속하는 코드워드의 다음 디코딩 처리의 분기 전이 확률 Γ(y_k,S_{k -1},S_k)의 초기화를 위한 파라미터로서 다시 사용할 수 있다. 그러나, 이는 이전 코드워드의 에러가 다음 코드워드로 전달되는 것을 쉽게 할 수 있다. In another embodiment, outpatient

Alternatively, the evaluation criteria L (d _k ) can be used again as a parameter for the initialization of the branch transition probability Γ (y _k , S _{k -1} , S _k ) of the next decoding process of the subsequent codeword. However, this can make it easier for an error in the previous codeword to be passed to the next codeword.

8 and 9 are flowcharts of decoding processing using the turbo principle according to another embodiment of the present invention. In this embodiment, multiple decoder instances are used at the decoder. For example, such a structure can be applied to using a turbo encoder / decoder.

8 and 9 show the operation of the first decoder instance, and the right branch shows the operation of the second decoder instance. To further distinguish the parameters of the two decoder instances, 1 and 2 are superscripted or subscripted.

Basically, the steps performed by the two decoder instances are similar to the respective steps described with reference to FIG. Thus, the description of FIGS. 8 and 9 focuses on the changes applied to the decoding process.

를 생성하거나 획득할 때(단계(702) 참 조), 분기 전이 확률

및

의 값이 초기화될 수 있다(단계(703) 및 (704) 참조). 다음으로, 순방향 순환 단계(705) 및 역방향 순환 단계(706)가 수행된다. In FIG. 8, the receiving means in step 801 may receive the codeword y _k and provide it to the first decoder instance. Exclusion set using, for example, a reception quality indicator for each bit of the receiving means.

(See step 702), the branch transition probability

And

Can be initialized (see steps 703 and 704). Next, a forward circulation step 705 and a reverse circulation step 706 are performed.

를 재구성하는 대신 외래

(또는 이에 기초한 평가 기준 L₁(d_k))을 생성할 수 있다. 생성된 외래

(또는 평가 기준 L₁(d_k))은 그 디코딩 처리에 사용하기 위한 제 2 디코더 인스턴스로 송신될 수 있다. According to this embodiment of the present invention, the first decoder instance is codewordd at step 802.

Instead of reorganizing

(Or evaluation criteria L ₁ (d _k ) based thereon). Generated foreign

(Or evaluation criteria L ₁ (d _k )) may be sent to a second decoder instance for use in its decoding process.

를 생성할 수도 있고, 이를 제공받을 수도 있다. 다른 방안으로, 예컨대 점선 화살표로 표시된 바와 같이 제 1 디코더 인스턴스의 결과를 사용할 때, 단계(803)에서 제외 세트

가 생성될 것이다. 단계(803)에서 제 1 디코더 인스턴스의 처리 결과를 고려하는 것은 선택 사항이라는 점에 주의한다. In step 803, the second decoder instance receives the codeword y _k from the receiving means. Next, exclusion set

May be generated or provided. Alternatively, when using the result of the first decoder instance as indicated by a dashed arrow, for example, the exclusion set at step 803.

Will be generated. Note that it is optional to consider the processing result of the first decoder instance in step 803.

을 초기화할 수 있다. 외래

또는 평가 기준 L₁(d_k)이 제 2 디코더 인스턴스의 초기화에서 고유

로서 사용될 수 있다. 또한,

의 값은 단계(703, 704)에서 설명된 것과 유사한 방식으로 초기화된다. Next, in step 804, the second decoder instance has a branch transition probability.

Can be initialized. Outpatient

Or evaluation criteria L ₁ (d _k ) is unique in the initialization of the second decoder instance

Can be used as. Also,

The value of is initialized in a manner similar to that described in steps 703 and 704.

를 초기화할 때, 순방향 순환 단계(806) 및 역방향 순환 단계(807)는 도 7의 단계(705, 706)를 참조로 설명된 바와 같은 방식으로 수행된다.

When initializing, the forward cycling step 806 and the reverse cycling step 807 are performed in the same manner as described with reference to steps 705 and 706 of FIG.

를 재계산한 이후에, 코드워드

가 재구성될 수 있다. 도 8의 실시예에 따라서, 다음 단계(808)에서 외래

가 생성되어서, 이 값들에 기초해서 단계(809)에서 코드워드

가 재구성될 수 있다. Probability distribution

After recalculating, codeword

Can be reconstructed. According to the embodiment of FIG. 8, in the next step 808 the outpatient

Is generated and based on these values, the codeword in step 809

Can be reconstructed.

As will be appreciated, the second decoder instance can be operated with a delay associated with the first decoder instance so that the result of the first decoder instance can be used in the decoding process of the second decoder instance. Note that in another embodiment, the first decoder instance may reconstruct the decoded codeword, which may be compared with the decoded codeword obtained from the second decoder instance. In this case, the second decoder may or may not operate with a delay relative to the first decoder instance. This process will be described in more detail below with reference to FIG.

9 is a flowchart of a decoding process using the turbo principle according to another embodiment of the present invention. The decoding process of the two decoder instances shown in the left and right branches of FIG. 9 are almost identical. The first decoding iteration of the first decoder instance is similar to that described with reference to FIG. 8, that is, in the first decoding iteration, steps 901 and 902 are similar to steps 702 and 703 of FIGS. 7 and 9.

를 생성한다. 또한, 제 1 디코딩 인스턴스는 디코드 코드워드

를 구성한다. Upon initialization and calculation of the forward and backward cycles (see steps 704, 705, 706), the first decoder instance is a foreign provided to the first encoder instance.

Create In addition, the first decoding instance is a decode codeword.

Configure

Since the result of the first decoder instance can be used simultaneously or with a certain delay in step 804, the second decoder instance may perform the same decoding as the first decoder instance (steps 803-807, 809 and 904). ), Iterative decoding may be performed as described with reference to the second decoder instance of FIG. 8.

를 생성한다. 단계(905)에서, 2개의 생성된 코드워드

가 비교되어서, 같다고 나온 경우에, 단계(906)에서 디코딩 과정이 종료된다. When the first decoding iteration is over, the second decoding instance is reconstructed codeword

Create In step 905, the two generated codewords

Are compared and found to be the same, the decoding process ends at step 906.

를 제 1 디코더 인스턴스로 제공할 수 있다(단계(904)). 제 2 디코더 인스턴스와 유사하게, 제 1 디코더 인스턴스는 디코딩 반복에서 고유 정보 예컨대, 고유

로서 외래 정보를 사용할 수 있다. 즉, 제 2 디코더 인스턴스의 정보는 단계(902)에서 새롭게 초기화된 분기 전이 확률

의 세트를 획득하고, 선택적으로는 단계(901)에서 새로운 제외 세트

를 결정하는 데 사용될 수 있다.However, if a negative decision result is obtained at step 905, decoding iteration may be further performed. In this case, the second decoder instance is foreign as shown by the dotted line.

May be provided to the first decoder instance (step 904). Similar to the second decoder instance, the first decoder instance is unique in the decoding iteration, eg unique

Foreign information can be used as That is, the information of the second decoder instance is the branch transition probability newly initialized in step 902.

Obtain a set of s, optionally in step 901 a new set of exclusions

Can be used to determine.

를 획득하기까지 반복을 여러 번 수행할 수 있다. 또한, 미리 정해진 수의 반복 이후에도 재구성된 코드워드

가 매칭되지 않는 경우에는, 디코 딩 과정은 중단될 수도 있고, 다음 처리 인스턴스에 디코딩 에러 신호가 송신될 수 있다. This allows the decoder to similarly reconstruct the codeword, which will terminate the received codeword y _k .

Iteration can be performed several times until obtaining. Also reconstructed codewords after a predetermined number of iterations

If is not matched, the decoding process may be aborted and a decoding error signal may be sent to the next processing instance.

Although the example decoding process of FIG. 9 has been described as a decoder instance reconstructing and comparing codewords, the proposed process in the embodiment of FIG. 8 may be used in conjunction with performing multiple decoding iterations before reconstructing a codeword. Note that you can.

Next, Fig. 10 will be described in detail. 10 illustrates a transmitter and receiver unit in accordance with an embodiment of the present invention. The transmitter 1001 includes an encoder 1002 and a transmitting means 1003. The transmitting means comprises a demodulator for demodulating the signal encoded by the encoder 1002. As shown by the dashed arrows, the encoder 1002 may encode the input data into codewords suitable for decoding in accordance with embodiments of the various decoding processes described above. The modulated data can be transmitted by the transmitting means 1003 using the illustrated antenna.

The receiver 1004 receiving the encoded signal may comprise receiving means 1006, which may comprise a demodulator to demodulate the received signal. When extracting parameters such as the y _k value and the transmission quality or reliability criterion of each bit of the codeword y _k received at the receiving means 1006, these data may be provided to the decoder 1005, which is described above. Consider data for initiating the decoding process as described.

The decoder 1005 may comprise processing means 1007 suitable for decoding the received data to generate a reconstructed codeword according to the method described above.

11 and 12 illustrate a mobile terminal (UE) 1101 and a base station (node B), respectively, according to another embodiment of the invention. The mobile terminal 1101 and the base station may each include a transmitter 1001 and a receiver 1004 as shown in FIG. 10 to perform communication.

FIG. 13 shows a schematic diagram of a communication system according to an embodiment of the present invention, which includes the mobile terminal 1101 shown in FIG. 11 and the base station (Node B: 1201) shown in FIG.

This schematic diagram includes a core network (CN: 1303) and a UMTS terrestrial radio access network (UTRAN: 1302). Mobile terminal 1101 may be connected to UTRAN 1302 via a wireless link to Node B 1201. The base station of the UTRAN 1302 may be further connected to a radio network controller (RNC: 1304). CN 1303 may include a (gateway) mobile switching center (MSC) that connects itself to a public switched telephone network (PSTN). Home Location Register (HLR) and Visitor Location Register (VLR) can be used to store user-related information. In addition, the core network may provide access to an IP-based network through a Serving GPRS Support Node (SGSN) and a Gateway GPRS Support Node (GGSN).

Although reference has been made to mobile communication systems by way of example above, those skilled in the art will appreciate that the present invention can be applied for use in wireless (data) networks such as digital video broadcasts such as IEEE 802.11, DVB or digital audio broadcasts such as DAB or DRM, for example. I will understand.

Claims (16)

A method for decoding at the decoder at least one codeword generated by an encoder comprising a structure that provides a code that can be described as a set of branch transitions in a trellis diagram. To a) initializing a set of branch transition probabilities of the decoder based on the received codeword and the encoder structure; b) initializing a first probability distribution and a second probability distribution in accordance with an initial state of the encoder used to encode the at least one codeword; c) recalculating, using a recursive algorithm, the value of the first probability distribution based on the initial value of the first probability distribution and the set of branch transition probabilities; d) recalculating, using a cyclic algorithm, the value of the second probability distribution based on the initial value of the second probability distribution and the set of branch transition probabilities; e) reconstructing a decoded codeword based on an extrinsic probability measure calculated based on the set of received codewords and the branch transition probability, the first and second probability distributions. But A subset of the initial value of each of the first probability distribution or the second probability distribution and a subset of the set of branch transition probabilities for recalculating each probability distribution in one or both of steps c) and d) Using sets, The value of the subset satisfies a predetermined reliability criterion. Decoding method. The method of claim 1, The encoder may be described by a shift register structure comprising at least one of a feed-forward arithmetic operation and a feed-back arithmetic operation. Decoding method. The method according to claim 1 or 2, The code is suitable for decoding using a maximum a-posteriori algorithm. Decoding method. The method according to any one of claims 1 to 3, Initiating the set of branch transition probabilities using an inherent probability measure in step a). Decoding method. The method according to any one of claims 1 to 4, Reconstructing the decoded codeword using an inherent probability measure in step e). Decoding method. The method according to claim 4 or 5, Decode the at least one codeword in a first decoding step using a decoder that can be described as two separate decoder instances, Using the foreign probability measure of the first decoder instance as an inherent probability measure of the second decoder instance Decoding method. The method of claim 6, Performing at the first decoder instance a second decoding iteration comprising the steps a) to e), The first decoder instance uses the foreign probability measure of the second decoder instance as the inherent probability measure. Decoding method. The method according to any one of claims 1 to 7, The reliability criterion is based on at least one of a channel estimate of the radio channel in which the at least one codeword was received, an absolute value of the elements of the first and / or second probability distributions, the number of decoding steps performed and a random process. Decoding method. The method of claim 8, If the signal-to-noise ratio of the component and / or the absolute value of the component is less than or equal to a predetermined threshold, the reliability criterion is not satisfied by the component of the first or second probability distribution. Decoding method. A decoder for decoding at least one codeword generated by an encoder comprising a structure that provides a code that can be described as a set of branch transitions in a trellis diagram. a) means for initializing a set of branch transition probabilities of the decoder based on the received codeword and the encoder structure; b) means for initializing a first probability distribution and a second probability distribution in accordance with an initial state of the encoder used to encode the at least one codeword; c) means for recalculating, using a cyclic algorithm, the value of the first probability distribution based on the initial value of the first probability distribution and the set of branch transition probabilities; d) means for recalculating, using a cyclic algorithm, the value of the second probability distribution based on the set of initial values of the second probability distribution and the branch transition probability; e) means for reconstructing a decoded codeword based on the set of received codewords and the branch transition probability, the foreign probability measure calculated based on the first and second probability distributions, The means recalculate each probability distribution in one or both of c) and d), a subset of the initial value of each of the first probability distribution or the second probability distribution and a subset of the set of branch transition probabilities. Suitable for using The value of the subset satisfies a predetermined reliability criterion. Decoder. The method of claim 10, Means suitable for carrying out the method according to any one of claims 1 to 9 Decoder. In a mobile terminal of a mobile communication system, Receiving means for receiving at least one codeword, Demodulation means for demodulating the received at least one codeword; Decoder according to claim 10 or 11 Mobile terminal comprising a. The method of claim 12, Encoding means for encoding data of at least one codeword, Transmitting means for transmitting the at least one codeword More, The transmitted at least one codeword is suitable for decoding according to the method of any one of claims 1 to 9. Mobile terminal. In a base station of a mobile communication system, Receiving means for receiving at least one codeword, Demodulation means for demodulating the received at least one codeword; Decoder according to claim 10 or 11 Base station comprising a. The method of claim 14, Encoding means for encoding data of at least one codeword, Transmitting means for transmitting the at least one codeword More, The at least one transmitted codeword is suitable for decoding according to the method of any one of claims 1-9. Base station. A mobile communication system comprising at least one base station according to claim 14 or 15 and at least one mobile terminal according to claim 12 or 13.

Images