JP2001127647A - Decoder for parallel consecutive code, decoding method, and storage medium recording decoding program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enhance the error correction capability in a decoder for parallel consecutive codes (turbo cods) in order to cope with an environment with a low S/N. SOLUTION: The decoder for parallel consecutive codes of this invention is provided with a soft output element decoding section 11 that receives a communication path organization series U, communication path redundant information series Y1 and external likelihood information Le2 (u) and outputs soft output likelihood value L1 (u) and with a likelihood update processing section 15 that compares the preceding soft output likelihood value L2 (u) with this soft output likelihood value L1 (u) to selectively update succeeding external likelihood information Le1 (u).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a decoder for a parallel concatenated code (turbo code) applied to an information communication system or an information recording system, and more particularly to a bit error rate characteristic in an environment having a low signal-to-noise ratio characteristic. The present invention relates to a parallel concatenated code decoder, a decoding method and a recording medium on which a decoding program is recorded.

[0002]

2. Description of the Related Art A parallel concatenated code known as a turbo code has been attracting attention as a code giving characteristics very close to the Shannon limit. FIGS. 8 and 9 are block diagrams showing configurations of a known turbo encoder and decoder. FIG. 8 shows a turbo coder with a coding rate r = 1/3, which includes component coder 101 and 103 and interleaver 105. The coded bit sequence d is branched and transmitted as a systematic bit sequence u, and is input to the elementary encoder 101 and the interleaver 105.

[0003] An element encoder 101 encodes an encoded bit sequence d by an error correction code (hereinafter, referred to as an element code) and outputs a redundant bit sequence y1. The interleaver 105 generally writes the coded bit sequence d once in a memory and reads it out in a different order,
The order of the data is mixed and sent to the element encoder 103.
The element encoder 103 encodes the interleaved bit sequence to be encoded by the element code to generate a redundant bit sequence y2
Is sent. The element encoders 101 and 103 include a recursive systematic convolutional encoder (RSC).
lutional Encoder) is usually used.

[0004] The turbo decoder shown in FIG. 9 includes soft output element decoders (Soft Output Decoders) 113 and 117, an interleaver 115, a deinterleaver 119, and a delay unit 3.
4 and 36 and a hard decision unit 121. The soft output element decoder 113 generates a channel organization information sequence U corresponding to the organization bit sequence u, a channel redundancy information sequence Y1 corresponding to the redundancy bit sequence y1, and external likelihood information Le2 (u) (initial value is 0
) And the soft output likelihood value L1 of the systematic bit sequence u
(u) (not shown) is calculated. The soft output likelihood value L1 (u) includes the decoding result with reliability information of the decoding result.

The soft output element decoder 113 further includes a soft output likelihood value L1 (u) of the systematic bit sequence u and a channel organization information sequence U
And external likelihood information Le2 (u) from the external likelihood information (Le1 (u) = L1
(u) -U-Le2 (u)) is output. The interleaver 115 inputs the external likelihood information Le1 (u) and the channel organization information sequence U via the delay unit 34, and outputs the interleaved external likelihood information Le1 '(u) and the channel organization information sequence U 'To the soft output element decoder 117.

The soft output element decoder 117 outputs a redundant bit sequence y2 via a delay unit 36 in addition to the interleaved external likelihood information Le1 '(u) and the channel organization information sequence U'.
, And outputs an interleaved soft output likelihood value L2 ′ (u),
From the interleaved soft output likelihood value L2 ′ (u) and channel organization information sequence U ′, and external likelihood information Le1 ′ (u), external likelihood information (Le2 ′ (u) = L2 ′ (u ) -U'-Le1 '(u)) and sends it to the deinterleaver 119.

[0007] The deinterleaver 119 is the interleaver 1
15 is output in the reverse order of the data exchange order. That is, the interleaved soft output likelihood value L2 ′ (u)
And the external likelihood information Le2 '(u) are returned to the order before the interleaving, and the soft output likelihood value L2 (u) and the external likelihood information Le2 (u)
Output as Further, hard decision unit 121 makes a hard decision on soft output likelihood value L2 (u) and outputs final decoded result d ＾.
The external likelihood information Le2 (u) is fed back to the soft output element decoder 113 for the next decoding process.

As described above, in the turbo decoder shown in FIG. 9, the decoding process is repeated while updating the soft output likelihood values L1 (u) and L2 (u) of the systematic bit sequence u, and a plurality of loops are performed. After that, the soft output likelihood value L2 from the soft output element decoder 117
(u) is hard determined. Also, the soft output element decoder 113,
The algorithm used in 117 includes a maximum posterior probability (MAP: Maximum A Posteriori Probabilit
y) MAX-lo with reduced amount of decoding and MAP decoding
g-MAP, soft output Viterbi algorithm (SOVA: So
ft Output Viterbi Algorithm) decoding and the like are known.

Further, in actual turbo code decoding, after a predetermined number of iterative decoding processes, if the characteristics of the iterative processes are not further improved, the decoding process is stopped from the viewpoint of the amount of calculation and the processing delay. You. As a criterion for stopping the decoding process, there is known a method of observing the hard decision result of the soft output element decoder 117 and stopping the decoding if there is no change.

[0010]

It is clear from various research results that a turbo code is a high-performance code exhibiting a characteristic very close to the Shannon limit, but it is apparent from a variety of research results so far that a turbo code is used in a mobile communication system or the like. When applied as a generalization scheme, a situation often occurs where the signal-to-noise ratio of the communication channel is not sufficient. For this reason, it is desired that the error correction capability of the turbo code decoder be further improved especially in an environment where the signal-to-noise ratio is low. Further, the turbo coding and decoding system may be applied to two-way communication of voice as an application of an information source, and it is desired to reduce a processing delay time.

In the turbo code and decoding method shown in FIGS. 8 and 9, as described above, the soft output likelihood values L1 (u) and L2 (u) of the systematic bit sequence u and the external likelihood information Le1 (u ) And Le2 (u) are updated and the decoding process is repeated. That is, the channel organization information sequence U and the channel redundancy information sequences Y1 and Y2 input to the soft output element decoders 113 and 117 are not updated during the iterative decoding process.

In an environment where the signal-to-noise ratio is low, the channel organization information sequence U, which is the input to the soft output element decoder 113,
Also, many noise components are included in the channel redundancy information sequences Y1 and Y2. The reliability of the channel organization information sequence U is low, and the soft output likelihood values L1 (u) and L2
If the sign of (u) is different, or if the signs of the soft output likelihood values L1 (u) and L2 (u) do not converge due to repeated inversion even if the iterative decoding process is performed, the bit error rate is significantly degraded. I do.

The causes of this deterioration are the soft output likelihood values L1 (u) and L2
If the sign of (u) is different, the likelihood value including the possibility of the error is output to the external likelihood information Le2 (u ) Or Le1 (u)
It is considered that the effect continues to exist in the process of the iterative decoding process. Further, when the reliability of the channel redundancy information series Y1 and Y2 is low,
In each decoding process, the soft output likelihood values L1 (u), L1
This causes suppression of the increase in the absolute value of 2 (u), and the characteristics are not improved even if the number of times of iterative decoding is increased.

The above problem is caused not only by the fact that the decoding characteristics deteriorate in an environment where the signal-to-noise ratio is low, but also in the case where the above-described decoding processing stop criterion is followed, the number of repetitions until the processing is stopped increases, and the processing delay increases Time increases. An object of the present invention is to provide a decoder, a decoding method, and a recording medium on which a decoding program is recorded, which improves decoding characteristics of a turbo code in an environment with a low signal-to-noise ratio.

Another object of the present invention is to provide a method for processing soft output likelihood values L1 (u) and L2 (u) and external likelihood information in a decoding process.
By appropriately updating the channel information U, Y1, Y2 input to the soft output element decoder in addition to Le1 (u), Le2 (u),
An object of the present invention is to provide a decoder, a decoding method, and a recording medium on which a decoding program is recorded, in which characteristics are improved by increasing the reliability of likelihood information handled in iterative decoding processing.

Still another object of the present invention is to provide a decoder, a decoding method, and a decoding program for shortening a processing delay time until completion of a decoding process by obtaining predetermined decoding characteristics by a small number of iterative decoding processes. To provide a recording medium on which is recorded. More specifically, by improving the reliability of the soft output likelihood value as described above, the probability that the sign of the soft output likelihood value L1 (u) and L2 (u) is different is reduced, and until the decoding stop criterion is met. The object of the present invention is to reduce the processing delay time while improving the convergence and improving the error rate.

[0017]

In order to achieve the above object, a parallel concatenated code decoder according to the present invention receives a channel organization information sequence, a channel redundancy information sequence and prior likelihood information, A soft output element decoding unit that outputs information, and a prior likelihood that selectively updates the next prior likelihood information by comparing the previous posterior likelihood information with the present posterior likelihood information this time. And an updating unit.

In another aspect, a parallel concatenated code decoder according to the present invention is a soft output element for receiving a channel organization information sequence, a channel redundancy information sequence, and prior likelihood information and outputting posterior likelihood information. A decoding unit; and an organization information sequence updating unit configured to compare the previous posterior likelihood information with the communication channel organization information sequence to selectively update the communication channel organization information sequence.

In another aspect, a parallel concatenated code decoder according to the present invention comprises a soft output element for receiving a channel organization information sequence, a channel redundancy information sequence, and prior likelihood information and outputting posterior likelihood information. A decoding unit, and a redundant information sequence updating unit that compares the previous posterior likelihood information with the communication channel redundant information sequence to selectively update the communication channel redundant information sequence.

The parallel concatenated code decoding method according to the present invention comprises a soft output element decoding step of inputting a channel organization information sequence, a channel redundancy information sequence, and prior likelihood information and outputting post-likelihood information; A prior likelihood updating step of comparing the posterior likelihood information with the present posterior likelihood information and selectively updating the next prior likelihood information.

In another aspect, a parallel concatenated code decoding method according to the present invention provides a soft output element for receiving a channel organization information sequence, a channel redundancy information sequence, and prior likelihood information and outputting posterior likelihood information. A decoding step; and an organization information sequence updating step of selectively updating the communication channel organization information sequence by comparing the previous posterior likelihood information with the communication channel organization information sequence.

In another aspect, a parallel concatenated code decoding method according to the present invention provides a soft output element for inputting a channel organization information sequence, a channel redundancy information sequence, and prior likelihood information and outputting posterior likelihood information. A decoding step, and a redundant information sequence updating step of comparing the previous posterior likelihood information with the channel redundant information sequence to selectively update the channel redundant information sequence.

According to the present invention, a soft output element decoding unit for inputting a channel organization information sequence, a channel redundancy information sequence and prior likelihood information and outputting posterior likelihood information is provided. And a prior likelihood updating unit for selectively updating the next prior likelihood information by comparing degree information and the present posterior likelihood information, and a function of a parallel concatenated code decoder comprising: It is a computer-readable recording medium on which a program is recorded.

In another aspect, the present invention provides a soft output element decoding unit which receives a channel organization information sequence, a channel redundancy information sequence, and prior likelihood information and outputs posterior likelihood information. And an organization information sequence updating unit that selectively updates the channel organization information sequence by comparing the previous posterior likelihood information with the channel organization information sequence, to function as a parallel concatenated code decoder. Computer-readable recording medium on which a program for recording is recorded.

Further, in another aspect, the present invention provides a soft output element decoding unit which inputs a channel organization information sequence, a channel redundancy information sequence, and prior likelihood information, and outputs posterior likelihood information. A redundant information sequence updating unit that compares the previous posterior likelihood information with the channel redundant information sequence and selectively updates the channel redundant information sequence, thereby functioning as a parallel concatenated code decoder. Computer-readable recording medium on which a program for recording is recorded.

[0026]

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings. FIG. 1 shows a first embodiment of the present invention.
FIG. 2 is a block diagram showing a configuration of a turbo code decoder according to the embodiment. The turbo code decoder shown in FIG. 1 has a configuration with a coding rate r = 1/3 as an example, and has two soft output element decoders 11 and 27.

The turbo code decoder according to the present embodiment includes a soft output element decoder (Soft Output Decoder) 11,
27, an interleaver 21, a deinterleaver 29,
Hard decision devices 13 and 31 and delay devices 23, 25, 33 and 3
5, and likelihood update processing units 15 and 37, and the likelihood update processing unit 15 (37) includes a decision value comparator (Decision Bit
Comparison) 17 (39) and the external likelihood update unit (Externa
l Likelihood Update Unit) 19 (41).

In this embodiment, the soft output element decoder 11
The hard decision result 75 of the soft output likelihood value L1 (u) from the soft output likelihood value L2 (u) from the soft output element decoder 27 and the hard decision result 91 of the soft output likelihood value L2 (u) are And external likelihood information Le1 (u) and Le2 (u) to be input to the next soft output element decoder are generated based on the comparison result.

The soft output element decoder 11 includes a channel organization information sequence U corresponding to the organization bit sequence u from the turbo encoder (see FIG. 8) and a channel redundancy information sequence Y1 corresponding to the redundancy bit sequence y1. In addition, external likelihood information Le2 (u) as a feedback component by the previous iterative decoding
And performs soft output decoding to output a soft output likelihood value L1 (u) of the systematic bit sequence u. This soft output likelihood value L1 (u) is
The input is directly input to the external likelihood update unit 19 of the likelihood update processing unit 15 and is also input to the hard decision unit 13 to perform a hard decision. As a hard decision result 75, the determination value comparator 17 of the likelihood update processing unit 15 is used. Is input to

The soft output likelihood values L1 (u) and L2 (u) are output from the soft output element decoders 11 and 27 as likelihood information of the systematic bit sequence u. After a predetermined number of iterative decoding processes, the code of the soft output likelihood value L1 (u) or L2 (u) becomes the final decoded output. Therefore, when the codes of the soft output likelihood values L1 (u) and L2 (u) are different for a certain channel organization information sequence U, one of the likelihood values contains an error. In order to avoid deterioration of decoding performance due to using an erroneous likelihood value as external likelihood information Le2 (u) in the next soft output element decoding, the decoder of the turbo code according to the present embodiment uses soft output element decoding. The likelihood update processing units 15 and 37 corresponding to the soft output likelihood values L1 (u) and L2 (u) are provided outside the units 11 and 27, respectively.

The judgment value comparator 17 of the likelihood update processing unit 15
Is a hard decision result 75 corresponding to the soft output likelihood value L1 (u),
A soft output likelihood value L2 (u) is input via the delay unit 25 to the hard decision result 91 corresponding to the soft output likelihood value L2 (u), and a comparison is made. The comparison result 77 is sent to the external likelihood updating unit 19.

The external likelihood updating unit 19 generates a soft decision likelihood value L1.
(u), L2 (u) and external likelihood information Le2 (u) are input, and a predetermined update process is executed based on the comparison result 77. For example, when the comparison result 77 indicating that the hard decision results 75 and 91 are equal is input, the soft output likelihood value L1 (u) output from the soft output element decoder 11 is used without updating.

When the comparison result 77 indicating that the hard decision results 75 and 91 are different is input, for example, the average of the soft output likelihood value L1 (u) and the soft output likelihood value L2 (u) is obtained. The value is obtained, updated to a new soft output likelihood value L1 (u) (not shown), and output. In order to more positively use the likelihood value, the average value may be obtained after weighting the likelihood value of each of the soft output likelihood values L1 (u) and L2 (u) with a predetermined weight. Updated soft output likelihood value
The absolute value of L1 (u) becomes smaller when the comparison result 77 indicating that the hard decision results 75 and 91 are different is output.

Further, the external likelihood update unit 19 uses the updated soft output likelihood value L1 (u) to generate external likelihood information (Le1 (u) = L1
(u) -U-Le2 (u)) is obtained and sent to the interleaver 21.
The interleaver 21 receives the channel organization information sequence U and the external likelihood information Le1 (u), reads out the data based on a predetermined reading order, agitates each data, and interleaves the channel organization information sequence U ′ and The external likelihood information Le1 '(u) is transmitted.

The soft output element decoder 27 outputs a channel organization information sequence U ′ corresponding to the interleaved organization bit sequence.
And external likelihood information Le1 ′ (u), a redundant bit sequence (y
2, see FIG. 8) to perform soft-output decoding and output an interleaved soft-output likelihood value L2 ′ (u).

The data of the interleaved soft output likelihood values L2 '(u) are in the order of being mixed by the interleaver 21, and are mixed by the deinterleaver 29 based on the reading order reverse to that of the interleaver 21. Is generated as the soft output likelihood value L2 (u). The soft output likelihood value L2 (u) is input to the hard decision unit 31 to make a hard decision, and the hard decision result 91 is input to the decision value comparator 39 of the likelihood update processing unit 37.

The judgment value comparator 39 of the likelihood update processing section 37
Is a hard decision result 91 corresponding to the soft output likelihood value L2 (u),
The hard decision result 75 corresponding to the previous soft output likelihood value L1 (u) is input via the delay unit 23 and compared, and the comparison result 9
3 is sent to the external likelihood updating unit 41.

The external likelihood updating unit 41 calculates the soft output likelihood value L2
(u) and the previous soft output likelihood value L1 (u) via the delay unit 33
And the channel organization information sequence U and the external likelihood information Le1 (u), and executes the same update processing as the above-described likelihood update processing unit 15 based on the comparison result 93. For example, hard decision result 7
When the comparison result 93 indicating that 5, 91 are equal is input, the soft output likelihood value L2 (u) output from the soft output element decoder 27 is used without updating.

When the comparison result 93 indicating that the hard decision results 75 and 91 are different is inputted, for example, the average value of the soft output likelihood value L1 (u) and the soft output likelihood value L2 (u) is obtained. And updates it as a new soft output likelihood value L2 (u). As described above, in order to use the likelihood value more positively, an average value may be obtained after weighting the likelihood value of each of the soft output likelihood values 73 and 89 with a predetermined weight. The absolute value of the updated soft output likelihood value L1 (u) becomes smaller when the comparison result 93 indicating that the hard decision results 75 and 91 are different is output.

Further, the external likelihood updating section 41 uses the updated soft output likelihood value L2 (u) to generate external likelihood information (Le2 (u) = L2
(u) -U-Le1 (u)) is obtained and fed back to the external likelihood updating unit 19 corresponding to the next soft output decoding. FIG. 2 is a block diagram showing a configuration of a turbo code decoder according to the second embodiment of the present invention. Components (or signals) having the same functions (or properties) as the components (or signals) in the above-described first embodiment are denoted by the same reference numerals, and description thereof is omitted.

As compared with the turbo code decoder according to the first embodiment, the turbo code decoder according to the present embodiment further includes organization information sequence updating units 43 and 44 having the same configuration. By these organization information sequence updating units 43 and 44,
The channel organization information sequence U corresponding to the organization bit sequence u (see FIG. 8), the interleaved channel organization information sequence U ′
Are appropriately updated, and the decoding characteristics of the turbo code are further improved in an environment where the signal-to-noise ratio is deteriorated.

Each of the organization information sequence updating units 43 and 44
It includes hard decision units 45 and 47, a decision value comparator 49, and an element decoder input update unit 51. The hard decision unit 45 makes a hard decision by inputting the channel organization information sequence U, and the hard decision unit 47 makes a hard decision by inputting the soft output likelihood value L2 (u).

The judgment value comparator 49 includes hard judgment units 45 and 47.
Are input and compared, and the comparison result is sent to the element decoder input update unit 51. The element decoder input update unit 51 performs predetermined processing on the channel organization information sequence U and the soft output likelihood value L2 (u) based on the comparison result from the decision value comparator 49, and U is selectively updated, and the soft output element decoder 1
Send to 1.

In the element decoder input updating unit 51, when the comparison results indicating that the hard decision results from the hard decision units 45 and 47 are equal to each other, for example, the channel organization information sequence U
Are used as they are without being updated.
When the comparison result indicating that the hard decision results are different is input, for example, an average value of the channel organization information sequence U and the soft output likelihood value L2 (u) is obtained, and a new channel organization information is obtained. Update to series U and output. In order to more positively use the likelihood value, the channel organization information sequence U and the soft output likelihood value L2
An average value may be obtained after multiplying (u) by a predetermined weighting coefficient.

The organization information sequence updating unit 44 is also configured in the same manner as the organization information sequence updating unit 43 (the contents are not shown), and the hard decision unit 45 inputs the above-mentioned communication channel organization information sequence U, and the hard decision unit 47 Inputs the soft output likelihood value L1 (u). The element decoder input update unit 51 performs predetermined processing on the channel organization information sequence U and the soft output likelihood value L1 (u) based on the comparison result of the decision value comparator 49, and Selectively update,
The information is sent to the interleaver 21 as the channel organization information sequence U1 and is also sent to the hard decision unit 45 of the organization information sequence update unit 43.

The channel organization information sequence U and the interleaved channel organization information sequence U ′ in the first embodiment described above are interleaved with the channel organization information sequence U by the organization information sequence updating units 43 and 44. It will be appropriately updated to the channel organization information sequence U1 '.

As described above, the soft output likelihood values L1 (u) and L2 (u)
Indicates the likelihood information of the organization bit sequence u, so if the reliability of the likelihood information is high, the codes of the soft output likelihoods L1 (u) and L2 (u) match the code of the channel organization information sequence U. Should do it. Therefore, when those codes are different from each other, it can be determined that an error is included in either information, and deterioration of decoding performance due to performing element decoding using an erroneous likelihood value or a channel value is avoided. You.

In the case where the organization information sequence updating units 43 and 44 in the present embodiment have different hard decision results between the channel organization information sequence U to be determined and the soft output likelihood value L2 (u) or L1 (u). The channel organization information sequence U or U is updated only when the hard decision results are equal, but the channel organization information sequence and the soft output likelihood value may be updated so that the absolute values thereof become large. .

FIG. 3 is a block diagram showing a configuration of a turbo code decoder according to the third embodiment of the present invention. Note that components (or signals) having the same functions (or properties) as the components (or signals) in the above-described second embodiment are given the same reference numerals and description thereof is omitted.

As compared with the turbo code decoder according to the second embodiment, the turbo code decoder according to the present embodiment further includes redundant information sequence updating units 53 and 54 having the same configuration. By the redundant information sequence updating units 53 and 54,
Each of the channel redundancy information sequences Y1 and Y2 corresponding to the redundant bit sequences y1 and y2 (see FIG. 8) is appropriately updated, and the decoding characteristics of the turbo code are further improved in an environment where the signal-to-noise ratio is deteriorated. . Redundant information sequence updating units 53 and 54
Are respectively a soft input / output encoder 55 and hard decision units 57 and 59
, A decision value comparator 61, and an element decoder input update unit 63.

Here, the likelihood values of the channel redundancy information sequences Y1 and Y2 of the redundant bit sequences y1 and y2 cannot be obtained in the decoder according to the above-described first or second embodiment. Therefore, in the present embodiment, the channel redundancy information sequence Y
1 and Y2, the soft output likelihoods L2 (u) and L2, respectively.
Values L (y1) and L (y2) (not shown) obtained by re-encoding 1 (u) are used. For the above re-encoding, a soft input / output encoder (Soft Output RSC Encoder) 55 that provides a soft input / soft output equivalent to that obtained when the encoded bit sequence d is encoded is used.

Here, at the time of encoding, the exclusive OR operation performed on the binary values (one bit of 0 or 1) a and b is as follows:
At the time of the re-encoding, the values to be handled are log likelihoods L (a) and L (b). For example, the following equation (-1) * sign [L (a)] * sign [L
(B)] * min [L (a), L (b)]. Hard decision unit 57
Performs a hard decision by inputting the channel redundancy information sequence Y1,
The hard decision unit 59 performs a hard decision by inputting a value L (y1) obtained by re-encoding the soft output likelihood value L2 (u).

The decision value comparator 61 includes hard decision devices 57 and 59
Are input and compared, and the comparison result is sent to the element decoder input update unit 63. Based on the comparison result from the decision value comparator 61, the element decoder input update unit 63 calculates the channel redundancy information sequence Y1 and the soft output likelihood value L2 (u).
And selectively updates the channel redundant information sequence Y1 to send to the soft output element decoder 11 as the channel redundant information sequence Y1 and to the hard decision unit 57 via the delay unit 24. I do.

In the element decoder input updating unit 63, when the comparison results indicating that the hard decision results from the respective hard decision units 57 and 59 are equal, for example, the channel redundancy information sequence Y
1 is used as it is without updating, and each hard decision unit 57,5
In the case where the comparison result indicating that the hard decision result is different from that of No. 9 is input, for example, an average value of the channel redundancy information sequence Y1 and the soft output likelihood value L2 (u) is obtained, and a new It is updated as the information sequence Y1. In order to use the likelihood value more positively, an average value may be obtained after weighting the channel redundancy information sequence Y1 and the soft output likelihood value L2 (u) with a predetermined weight.

The redundant information sequence updating unit 54 is also configured in the same manner as the redundant information sequence updating unit 53 (the contents are not shown), and the channel redundant information sequence Y2 and the re-encoded value L (y2) (shown in FIG. Z) is used. In the redundant information sequence updating unit 54, the hard decision unit 57 receives the channel redundancy information sequence Y2, and the hard decision unit 59 receives the value L (y2) obtained by re-encoding the soft output likelihood value L1 (u). I do. The element decoder input update unit 63 is provided with the judgment value comparator 6
1 to update the channel redundancy information sequence Y2 by performing predetermined processing on the channel redundancy information sequence Y2 and the soft output likelihood value L1 (u) based on the comparison result of 1. The signal is sent to the decoder 27 and sent to the redundant information sequence updating unit 54 via the delay unit 26.

The redundant information sequence updating sections 53 and 54 appropriately update the redundant channel information sequences Y1 and Y2 in the first or second embodiment to the redundant channel information sequences Y1 and Y2. . In the redundant information sequence updating units 53 and 54 in the present embodiment, the communication channel redundant information sequences Y1 and Y2 to be determined and the re-encoded value L
Although the channel redundancy information sequence Y1 or Y2 is updated only when the hard decision results of (y1) and L (y2) are different, the channel redundancy information sequence and the recoding are performed even when the hard decision results are equal. The value may be updated so that the absolute value of the converted value becomes large.

FIG. 4 is a diagram showing a signal-to-noise power ratio per bit-bit error rate characteristic of the turbo code decoder according to the first embodiment. FIG. 4 shows characteristics when the number of repetitions in the turbo code decoder according to the first embodiment is one (it = 1 (prop)) to eight (it = 8 (prop)). Similarly, the number of repetitions in the conventional turbo decoder is changed from 1 (it = 1 (conv)) to 8
The characteristics for four types of times (it = 8 (conv)) are plotted. According to the present embodiment, it is shown that, in an environment where the signal-to-noise ratio is low, the bit error rate characteristic is greatly improved as compared with the related art as the number of repetition decoding increases.

FIG. 5 is a diagram showing the number of times of iterative decoding-bit error rate characteristics of the turbo code decoder according to the first embodiment. FIG. 5 relates to FIG. 4 and shows that the signal to noise ratio in the turbo code decoder according to the first embodiment is 1 dB.
(1 dB (prop)) and 2 dB (2 dB (prop)) are plotted. Similarly, the signal-to-noise ratio of the conventional turbo decoder is 1 dB (1 dB (cond)).
v)) and 2 dB (2 dB (conv)). For example, when attention is paid to each plot where the signal-to-noise ratio in the related art and the present embodiment is 2 dB,
According to the present embodiment, the bit error rate characteristic is greatly improved by the effect of the iterative decoding in a region where the bit error rate characteristic is not significantly improved even if the number of repetition decodings is increased in the related art (for example, the number of repetition times is three or more). Is shown.

A turbo code decoder according to the fourth embodiment of the present invention will be described. In the above-described first to third embodiments, the updating process of the soft output likelihood values L1 (u) and L2 (u) is executed based on the comparison result of their signs. In order to synchronize sequences in the element decoder, a delay unit for the decoding processing time is required.

In this embodiment, the soft output likelihood values L1 (u), L1
By multiplying each of 2 (u) by a predetermined weighting coefficient α, the error rate of the soft output likelihood values L1 (u) and L2 (u) is reduced, the number of repetitions of the turbo code decoder is reduced, and the The convergence of decoding is improved. A turbo code decoder according to the fifth embodiment of the present invention will be described. By convention, if the signal-to-noise characteristics are sufficiently high, the external likelihood information Le1 (u) and Le2
It is known that the code of (u) substantially matches the final decoding result d ＾.

In the present embodiment, the external likelihood information Le1 (u), Le2 (Le2) is added to the configuration according to the first to fourth embodiments.
A configuration combining the control functions of the sign of (u) is adopted. Based on the comparison result in the judgment value comparator 17 (or 39) described above, external likelihood information Le1 (u) (or Le2 (u)) is obtained from the soft output likelihood value L1 (u) (or L2 (u)). Is obtained, in response to a change in the sign (from + to-or from + to +), the code of the external likelihood information Le1 (u) (or Le2 (u)) is forcibly output as 0. Control.

In the first to fifth embodiments described above, the delay units 23, 25, 33 and 3 having delay times corresponding to the processing times in the soft output element decoders 11 and 27 are described.
5, but if the hard decision results 75 and 91 corresponding to the soft output likelihood values L1 (u) and L2 (u) compared in the likelihood update processing units 15 and 37 are aligned in time, good. For example,
When a memory is used for the interleaver 21 and the deinterleaver 29, the memory performs storage processing equivalent to the delay unit. Therefore, the delay units 23, 25, 33, and 35 are omitted by controlling the storage time of the memory. be able to.

In the first to fifth embodiments described above, it is assumed that the interleaver 21 and the deinterleaver 29 are configured using a memory. However, the present invention is not limited to this, and data is stored in a predetermined order. Any configuration that allows stirring may be used. Further, in the above-described first to fifth embodiments,
The likelihood update processing units 15 and 37 update the soft output likelihood values L1 (u) and L2 (u) only when the hard decision results 75 and 91 of the likelihood values to be determined are different. Determination results 75, 9
Even when 1 is equal, it is possible to perform an update process in which the absolute value of the likelihood increases.

Further, in the first to fifth embodiments described above, the likelihood update processing units 15 and 37 are provided at the subsequent stage of the soft output element decoders 11 and 27, however, the circuit configuration is simplified. In consideration of the above, it is possible to reduce the number of sets of the soft output element decoder 11 (27) and the likelihood update processing unit 15 (37) (however, one or more sets are necessary).

FIG. 6 is a block diagram showing a configuration of a general turbo decoder having one soft output decoder. In addition,
Data having the same properties as the data shown in FIG. 8 are denoted by the same reference numerals, and description thereof will be omitted. This turbo decoder
Soft output element decoder 160 and interleavers 155, 15
7, a deinterleaver 159, a hard decision unit 122, and selectors 147, 149, 151, 153.

In the odd decoding process, the selector 1
Each of 47, 149 and 151 is the repetition control signal 1
45, the channel organization information sequence U, the channel redundancy information sequence Y1, and the external likelihood information Le2 (u) are sent to the soft output element decoder 160. The soft output element decoder 160 inputs the channel organization information sequence U, the channel redundancy information sequence Y1, and the external likelihood information Le2 (u), and outputs external likelihood information Le1 (u).
The selector 153 sends external likelihood information Le1 (u) to the interleaver 157 based on the repetition control signal 145.
The interleaver 157 mixes the external likelihood information Le1 (u) in a predetermined data order. Interleaved external likelihood information
Le1 '(u) is used for the next decoding process.

In the even-numbered decoding process, the selector 1
Each of 47, 149 and 151 is the repetition control signal 1
The interleaved channel organization information sequence U ′, the channel redundancy information sequence Y2, and the interleaved external likelihood information Le1 ′ (u) are sent to the soft output element decoder 160 based on 45. The soft output element decoder 160 receives the interleaved channel organization information sequence U ′, the channel redundancy information sequence Y2, and the interleaved external likelihood information Le1 ′ (u), and receives the interleaved soft output likelihood. The value L2 ′ (u) and the external likelihood information Le2 ′ (u) are output. The selector 153 converts the soft output likelihood value L2 ′ (u) interleaved based on the repetition control signal 145 and the external likelihood information into a deinterleaver 159.
To send to. Deinterleaver 159 is the interleaver 1
Stirring is performed in the reverse data order to 57, and external likelihood information Le2 (u)
Is transmitted for the next decryption process. The external likelihood information Le2 (u) is obtained by the hard decision
Output as ＾.

In the turbo decoder shown in FIG. 6, odd-numbered and even-numbered decoding processes are realized in response to the repetition control signal 145. When this configuration is applied to the turbo code decoder of the present invention shown in FIG.
3 and the interleaver 157, the likelihood update processing unit 15 is inserted between the selector 151 and the deinterleaver 159, and the likelihood update processing unit 37 is inserted between the selector 151 and the deinterleaver 159.
5 and 37 can be constituted by inputting predetermined data shown in FIG. The likelihood update processing unit 15 (3
7), the external likelihood information Le1 (u) from the selector 153 and the external likelihood information Le2 (u) from the deinterleaver 159 are selectively input according to the repetition control signal 145. 1 may be configured to be input.

When the configuration shown in FIG. 6 is applied to the turbo code decoder of the present invention shown in FIG. 2 or 3, for example, likelihood update processing section 15 is provided between selector 153 and interleaver 157. Then, the likelihood update processing unit 37 is inserted between the selector 151 and the deinterleaver 159. Further, the organization information sequence updating unit 43 (44) that inputs the channel organization information sequence U and the soft output likelihood value L2 (u) and outputs the channel organization information sequence U, or the channel redundancy information sequence Y1
And a soft output likelihood value L2 (u) or L1 (u), and a redundant information sequence updating unit 53 or 54 for outputting the channel redundant information sequence Y1 or Y2.

The turbo code decoder according to the present invention may be provided with one soft output element decoder 11, or may be provided with a plurality of soft output element decoders. In any case, the soft output element decoder 11 (27) outputs the channel organization information sequence U and the channel redundancy information sequences Y1 and Y2, and the external likelihood information (prior likelihood information) obtained by the previous decoding process. And outputs a soft output likelihood value (posterior likelihood information). Likelihood update processing unit 15 (3
7) selectively updates the next prior likelihood information according to a comparison result based on the current posterior likelihood information and the previous posterior likelihood information, and updates the updated prior likelihood information in the next time. Used for decryption processing.

The organization information sequence updating section 43 (44)
Is the previous posterior likelihood information and the channel organization information sequence U (U)
And selectively updates the channel organization information sequence U (U). The redundant information sequence update unit 53 (54)
Is the previous posterior likelihood information and the channel redundancy information sequence Y1 (Y
2) and selectively updates the channel redundancy information sequence Y1 (Y2). The present invention is not limited to the above configuration.

The turbo code decoder according to the present invention is also realized by a program for causing the turbo code decoder to function. This program is, for example, CD-
It is stored in a computer-readable recording medium such as a ROM. FIG. 7 is a block diagram illustrating a configuration example of a system that accommodates a recording medium that stores a program that executes an operation of a turbo code decoder according to the present invention. This system includes an input device 107, a CPU 108, a ROM 10
9, the RAM 111, the display device 139, and the CD-R
OM driver 143.

The recording medium on which the program for executing the operation of the turbo code decoder of the present invention is recorded may be the ROM 109 itself, or may be a C storage medium as an external storage device.
A program reading device such as a D-ROM driver 143 is provided, and a CD that can be read by being inserted therein is provided.
-It may be a ROM 141 or the like. Further, the recording medium may be a magnetic tape, a cassette tape, a floppy disk, a hard disk, an MO / MD / DVD, or the like, or a semiconductor memory.

[0074]

According to the present invention, the error correction capability of the turbo code decoder is improved in an environment where the signal-to-noise ratio of the communication channel is not sufficient. In particular, in the course of the turbo code decoding process, not only the soft output likelihood value and the external likelihood information but also the channel organization information sequence U input to the decoder.
Also, by appropriately updating the channel redundancy information sequences Y1 and Y2, the reliability of the likelihood value handled in the iterative decoding process can be increased, and the error rate characteristics can be improved.

In the decoding process, the reliability and convergence of the likelihood information can be improved, and the error rate can be improved, and the processing delay time according to the decoding process standard can be reduced. further,
Since the input to the decoder and the likelihood value for passing between the element decoders are updated, it can be widely implemented regardless of the algorithm of the element decoder to be selected.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a turbo code decoder according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a turbo code decoder according to a second embodiment of the present invention.

FIG. 3 is a block diagram showing a configuration of a turbo code decoder according to a third embodiment of the present invention.

FIG. 4 is a diagram illustrating a bit error rate characteristic with respect to a signal-to-noise power ratio per bit of a turbo code decoder according to the first embodiment of the present invention.

FIG. 5 is a diagram illustrating a bit error rate characteristic with respect to the number of times of iterative decoding of the turbo code decoder according to the first embodiment of the present invention.

FIG. 6 is a block diagram showing a configuration of a general turbo code decoder having one soft output decoder.

FIG. 7 is a block diagram illustrating a configuration example of a system accommodating a recording medium storing a program for executing an operation of a turbo code decoder according to the present invention.

FIG. 8 shows a known turbo encoder (encoding rate r = 1);
FIG. 3 is a block diagram showing a configuration of (/ 3).

FIG. 9 is a block diagram illustrating a configuration of a turbo decoder that is a known technique.

[Explanation of symbols]

U, U, U1: Channel organization information sequence Y1, Y1, Y2, Y2: Channel redundancy information sequence L1 (u), L2 (u): Soft output likelihood value Le1 (u), Le2 (u): External Likelihood information U ', U': Interleaved channel organization information sequence Le1 '(u), Le2' (u): Interleaved external likelihood information L2 '(u): Interleaved soft output likelihood value d: encoded bit sequence u: systematic bit sequence y1, y2: redundant bit sequence d ＾: final decoding result 11, 27, 101, 103, 160: soft output element decoder 13, 31, 45, 47, 57, 59, 121, 12
2: Hard decision unit 15, 37: Likelihood update processing unit 17, 39, 49, 61: Judgment value comparator 19, 41: External likelihood update unit 21, 105, 115, 155, 157: Interleaver 23, 24, 25, 26, 33, 34, 35, 36:
Delay units 29, 119, 159: Deinterleavers 43, 44: Organizational information sequence updating units 51, 63: Element encoder input updating units 53, 54: Redundant information sequence updating units 55: Soft input / output encoders 75, 91: Hard Judgment result 77, 93: Comparison result 107: Input device 108: CPU 109: ROM 111: RAM 139: Display device 141: CD-ROM 143: CD-ROM driver 113, 117, 160: Element decoder 145: Iterative control signal 147, 149, 151, 153: selector

Claims (9)

[Claims] A soft output element decoding unit which inputs a channel organization information sequence, a channel redundancy information sequence, and prior likelihood information and outputs posterior likelihood information; A prior art likelihood updating unit for comparing the posterior likelihood information with the posterior likelihood information to selectively update the next prior likelihood information. 2. A soft output element decoding unit which inputs a channel organization information sequence, a channel redundancy information sequence, and prior likelihood information and outputs posterior likelihood information; An organization information sequence updating unit for comparing the channel organization information sequence with the channel organization information sequence to selectively update the channel organization information sequence. 3. A soft output element decoding unit which inputs a channel organization information sequence, a channel redundancy information sequence, and prior likelihood information and outputs posterior likelihood information; And a redundant information sequence updating unit for selectively updating the communication channel redundant information sequence by comparing with the channel redundant information sequence. 4. A soft output element decoding step of inputting a channel organization information sequence, a channel redundancy information sequence, and prior likelihood information and outputting posterior likelihood information; A prior likelihood updating step of comparing the posterior likelihood information with the subsequent likelihood information to selectively update the next prior likelihood information. 5. A soft output element decoding step of inputting a channel organization information sequence, a channel redundancy information sequence, and prior likelihood information and outputting posterior likelihood information; An organization information sequence updating step of selectively updating the channel organization information sequence by comparing with a channel organization information sequence. 6. A soft output element decoding step of inputting a channel organization information sequence, a channel redundancy information sequence, and prior likelihood information and outputting posterior likelihood information; A redundant information sequence updating step of selectively updating the communication channel redundant information sequence by comparing with a channel redundant information sequence. 7. A computer, comprising: a communication channel organization information sequence;
A soft output element decoding unit that inputs a channel redundancy information sequence and prior likelihood information and outputs posterior likelihood information, and compares the previous posterior likelihood information with the present posterior likelihood information. A computer-readable recording medium on which is recorded a program for functioning as a parallel concatenated code decoder, comprising: a prior likelihood update unit for selectively updating the next prior likelihood information. 8. A computer, comprising: a communication channel organization information sequence;
A soft-output element decoding unit that inputs a channel redundancy information sequence and a priori likelihood information and outputs a posteriori likelihood information, and compares the previous posterior likelihood information with the channel organization information sequence to perform the communication. A computer-readable recording medium recording a program for functioning as a parallel concatenated code decoder, comprising: an organization information sequence updating unit for selectively updating a road organization information sequence. 9. A computer comprising: a communication channel organization information sequence;
A soft-output element decoding unit that inputs a channel redundancy information sequence and a priori likelihood information and outputs a posteriori likelihood information, and compares the previous posterior likelihood information with the channel redundancy information sequence to perform the communication. A computer-readable recording medium having recorded thereon a program for functioning as a parallel concatenated code decoder, comprising: a redundant information sequence updating unit for selectively updating a path redundant information sequence.