CN103888148A - LDPC hard decision decoding method for dynamic threshold value bit-flipping - Google Patents

Abstract

The invention discloses an LDPC hard decision decoding method for dynamic threshold value bit-flipping. An initial flipping threshold value T is determined, a dynamic threshold value, namely, a dynamic flipping threshold value is adopted, the flipping threshold value is changed into the dynamic threshold value so that the misflipping probability can be reduced, the number of times of iterations of decoding is reduced, and bit-flipping decoding of an LDPC is efficient. Meanwhile, the error correcting capability of the LDPC is improved. The error correcting capability of the LDPC hard decision decoding method is better than that of a standard bit-flipping method.

Description

A kind of LDPC code Hard decision decoding method of dynamic threshold bit reversal

Technical field

The present invention relates to the bit reversal building method of a kind of LDPC (Low Density Parity Check, low density parity check code) Hard decision decoding device.

Background technology

Need to carry out in the application of signal transmission various, often can adopt error correcting code (ECC, Error Correcting Code), can make signal transmission errors time, righted the wrong and obtain correct signal at receiving terminal.

Error correcting code can be applied in many systems, in communication system, may be subject to the interference of channel effect and interchannel noise, thereby cause the data of storing in flash memory device incorrect when signal transmission.The data of storing in flash memory device are the data after error correcting code device code, and for flash memory storage control device, error correcting code is an essential functional unit.

Along with the technique of memory is more and more advanced, memory cell volume is more and more less, and the data that memory cell is stored are also increasing gradually, causes flash memories constantly to raise reading the error probability producing in process.Therefore, the error correcting code error correcting capability of flash memories control device is to determine the whether qualified key factor of storage control device.But, in flash memory storage control device, adopt stronger error correcting code to need higher operand and longer operation time, thereby the application of increase flash memories control device is very restricted.

LDPC code is a kind of packeting error-correcting code with sparse check matrix that Robert Gallager proposed in thesis for the doctorate in 1962.Almost be applicable to all channels, its performance is approached shannon limit, and describes and realize simple, and the simple and practicable parallel work-flow of decoding, is applicable to hardware and realizes.

LDPC code has huge application potential, in deep space communication, optical fiber communication, satellite communication, satellite digital video, digital watermarking, magnetic/light/Hologram Storage, movement and fixed radio communication, cable modulating/demodulating and digital subscriber, is used widely.

More and more advanced according to the technique of flash memory device, the error correcting capability of the error correcting code in flash memories control device also needs to strengthen.In current flash memories control device, main error correcting code is BCH code, along with increasing of error probability, and BCH code space requirement and operational capability are also increased gradually; Along with the raising of flash memory storage technique, the error correcting capability of BCH code has been not suitable for the development of flash technology gradually, so need the error correcting code that error correcting capability is stronger.It is more appropriate selecting LDPC code to replace BCH code.

In the Hard decision decoding device of LDPC code, the setting of bit reversal threshold value T, plays vital effect to the Hard decision decoding of LDPC code, and it affects the error correcting capability of LDPC code and its decoding rate.

Increase the iterations of decoding if what bit reversal threshold value T set crosses conference, thereby affect decoding rate.

If the too small meeting that bit reversal threshold value T sets causes mistake upset just to increase like this iterations of decoding, thereby affects decoding rate.

For irregular check matrix, choose fixing bit reversal threshold value T, thereby can cause like that the very large mistake probability that overturns greatly to increase the iterations of decoding, serious in the situation that, cause decoding failure.

When the error rate of code element in sequence Z reaches certain value, use fixing bit reversal threshold value T to carry out decoding, decoder iterations is much can successfully not carry out decoding.

Summary of the invention

Based on dynamic threshold, the object of the present invention is to provide a kind of LDPC code Hard decision decoding method of dynamic threshold bit reversal, improve decoding rate and error correcting capability.

The present invention is by the following technical solutions:

A LDPC code Hard decision decoding method for dynamic threshold bit reversal, determines an initial turn threshold T, comprises the following steps:

1) read codeword sequence, according to codeword sequence computing syndrome;

2) judge whether syndrome is zero, if zero, iteration stopped, output codons sequence; Otherwise enter step 3)

3) each symbol bits to each code word in codeword sequence, calculates the number that does not meet check equations that this symbol bits participates in;

4) calculate turn threshold: along with increasing of iterations, turn threshold T successively decreases, and obtains current turn threshold;

5) if the number of ungratified check equations is greater than current turn threshold, corresponding symbol bits is overturn, and calculates the syndrome of the rear corresponding codeword sequence of upset;

6) repeating step 2), 3), 4) and 5), until successfully decoded, or while reaching maximum iteration time, output decoding failure.

The LDPC code Hard decision decoding method of above-mentioned dynamic threshold bit reversal, current turn threshold T:

T＝j(b-c·m)/a；

In formula: it is that m represents current for which time iteration, j is that symbol bits is corresponding when prostatitis column weight, and a, b, c are experience normal number.

The LDPC code Hard decision decoding method of above-mentioned dynamic threshold bit reversal, determines an initial turn threshold T, and the step of overturning in iteration is as follows:

A) according to hard decision sequence Z=[z ₀, z ₁..., z _n-1] computing syndrome S=[s ₀, s ₁..., s _m-1]:

$s_m={\mathrm{\Σ}}_{n=0}^{N-1}{H}_{\mathrm{mn}}{z_n}^T\mathrm{mod}2,m=\mathrm{0,1},...,M-1;$

Wherein H _mntest matrix, if syndrome S=0 stops iteration, output hard decision sequence Z also shows successfully decodedly, otherwise enters b);

B) to each symbol bits z _n, calculate the number f of the ungratified check equations of its participation _n:

$f_n={\mathrm{\Σ}}_{m=0}^{M-1}{s}_m{H}_{\mathrm{mn},n=\mathrm{0,1},...,N-1}$

If f _n> T, z overturns _n, obtain new hard decision sequence Z;

C) repeat a) and b) until successfully decoded, or reach maximum iteration time and show decoding failure;

Hard decision sequence is step 5) codeword sequence that obtains after upset.

According to the present invention, adopt dynamic threshold, namely dynamic turn threshold, becomes dynamic threshold by turn threshold and can reduce the probability that mistake is overturn, and has also reduced the iterations of decoding, makes the bit-flipping decoding of LDPC code more efficient; Also the error correcting capability that has improved LDPC code, the present invention strengthens to some extent than the error correcting capability of the bit reversal method of standard simultaneously.

Accompanying drawing explanation

Fig. 1 is the theory diagram of flash memories memory control device.

Fig. 2 is LDPC symbol bits upset decoding flow chart.

Embodiment

LDPC code is the one of linear block codes, and it has all characteristics of linear block codes.LDPC code can be divided into rule (regular-LDPC) and irregular (irregular-LDPC) two large class hypothesis check matrix H ₀for m × n rank matrix, regular LDPC code can be remembered and is (n, j, k), wherein n is code length, j is weight (i.e. 1 the number in row of the every row of check matrix, be called for short column weight (column weight), k is weight (1 the number at once of the every row of check matrix, be called for short row heavy (row weight)), and generally have j>2,1 number of the check matrix each row and column of k>j and irregular LDPC codes is incomplete same.

The interative encode method of LDPC code is broadly divided into two kinds: one is hard-decision method, and one is soft decision method.What hard-decision bits method for turning transmitted in iterative process is binary system hard information, and soft decision method transmits in iterative process is and the soft information of real number of probability correlation.Hard-decision method is simple to operate, be easy to hardware and realize, but error-correcting performance is general; Soft decision method better performances, but implementation complexity is higher.This programme improves mainly for Hard decision decoding device and improves its error correction energy ability.The detailed process of the hard-decision bits upset interpretation method of LDPC code is introduced as follows.

Bit reversal method is taken turns in iteration every, first according to the value of last round of hard decision sequence computing syndrome.If all syndrome is 0, stop iteration, and show successfully decoded, otherwise the number of the check equations that the syndrome of calculating each bit participation is 1, the bit reversal that check equations number is greater than certain predetermined threshold threshold value T will be participated in not meeting, obtain a new hard decision sequence, then enter next round iteration, until successfully decoded or reach maximum iteration time and show decoding failure.Suitably select the size of threshold value T, can reach best decoding performance.

Hard decision sequence is the sequence after the sequence that needs decoding of decoder reception is overturn through bit reversal method.

Fig. 1 is the functional block diagram after an embodiment of flash memory storage control device simplifies.In flash memory storage controller, comprise LDPC code decoder.Flash memory storage controller is mainly responsible for the read-write of data and the storage of data and other function.

Flash memory storage controller obtains that data are encoded computing through LDPC code decoder and the data that produce store into flash memories from main frame.If the data that main frame goes in flash memories need flash memory storage controller to read out from flash memories, decode computing and produce data and input to main frame through LDPC code decoder.

Fig. 2 is the LDPC code coding method flow chart of dynamic threshold (in literary composition, dynamic threshold, turn threshold, dynamic turn threshold and bit reversal threshold value represent the title of same threshold value under different environments for use, those skilled in the art will readily understand) bit reversal.Obtain codeword sequence by flash memory storage controller from flash memories, ldpc code decoder reads this codeword sequence.The whole decode procedure of LDPC code as can be seen from Figure 2.Determine an initial turn threshold T, thereby its process is as follows:

(1) read after codeword sequence, according to codeword sequence computing syndrome;

(2) judge whether syndrome S is zero, and zero stops iteration if, output codons sequence also shows successfully decodedly, otherwise enters (3);

(3) each symbol bits to each code word in each codeword sequence, calculates the number of the ungratified check equations of its participation;

(4) adopt the mode of successively decreasing to determine new turn threshold;

(5), if ungratified check equations number is greater than new turn threshold, corresponding code word is overturn;

(6) repeat (2), (3), (4) and (5) until successfully decoded, or reach maximum iteration time and show decoding failure.

Code element: usually represent a binary digit with identical symbol of the time interval in digital communication, the signal in such time interval is called (binary system) code element.And this interval is called as Baud Length.Code word (code word) is made up of some code elements.

The Essential Terms that are this area about used term, it is not explained in detail, as syndrome, according to the general explanation of Principle of Communication, be Linear codes (n, k) syndrome of code, if the error correcting capability of this yard is t, all error patterns that weight is less than or equal to t so have unique syndrome (syndrome) corresponding with it.

About turn threshold, adopt given formula to successively decrease, amount of calculation is smaller, but representative bad, can not effectively represent the present situation of decoding.

On the method basis of the bit-flipping decoding of LDPC code, the LDPC code coding method of a kind of dynamic threshold bit reversal of proposition.Principle is, LDPC code is in the time carrying out bit-flipping decoding, and its turn threshold is a fixed value, the upset easily making the mistake in decode procedure, thus the iterations that has increased decoder makes decoding time long.For the more effective bit-flipping decoding that utilizes LDPC code, this case has proposed a kind of Hard decision decoding method of LDPC code of dynamic threshold bit reversal.Idiographic flow as shown in Figure 2.

In the Hard decision decoding device of LDPC code, the setting of bit reversal threshold value T, plays vital effect to the Hard decision decoding of LDPC code, and it affects the error correcting capability of LDPC code and its decoding rate.

Increase the iterations of decoding if what bit reversal threshold value T set crosses conference, thereby affect decoding rate.

If the too small meeting that bit reversal threshold value T sets causes mistake upset just to increase like this iterations of decoding, thereby affects decoding rate.

For irregular check matrix, choose fixing bit reversal threshold value T, thereby can cause like that the very large mistake probability that overturns greatly to increase the iterations of decoding, serious in the situation that, cause decoding failure.

When the error rate of code element in sequence Z reaches certain value, use fixing bit reversal threshold value T to carry out decoding, decoder iterations is much can successfully not carry out decoding.

The Hard decision decoding method of the LDPC code of a kind of dynamic threshold bit reversal that this programme proposes can effectively be avoided above-mentioned problem.The interpretation method of this patent can reduce the iterations of decoding, thereby has increased decoding rate; Also error correcting capability can be improved, more mistake in sequence Z can be entangled.

The formula of dynamic threshold: T=j (b-cm)/a.Wherein T is threshold value, and which time iteration m is, j is column weight, and a, b, c are experience normal number (technique of constant and flash memories and the mistake that need to entangle are relevant).

Wherein a, b, c are empirical, for constant, in formula, do not affect dynamic threshold " dynamically " characteristic, and as initial threshold can be thought the fixed threshold under normal condition, the fixed threshold that those skilled in the art determines accordingly, by obtaining the dynamic threshold directly related with iterations and column weight with the correlation of column weight and iterations, thereby can effectively reduce amount of calculation, improve decoding efficiency.Because experience Changshu is on " dynamically " not impact, therefore, do not repeat them here.

The column weight j of dynamic threshold T and check matrix, decoder are that number of times, the empirical of which time iteration is closely bound up.The column weight of check matrix is the limiting value of dynamic threshold, and it is less that the larger dynamic threshold T of decoder iterations changes space, the erroneous decision that empirical need to be corrected by flash memories technique and sequence Z.The iteration of decoder each time, the code element in sequence Z can judge and determine whether overturning according to the dynamic threshold of each code element, carries out multiple code elements at every turn and overturns.

Syndrome is check matrix H _mnwith sequence Z transposition product the results are shown in following formula s _m.

Bit reversal method concrete steps are as follows:

(a) according to hard decision sequence Z=[z ₀, z ₁..., z _n-1] computing syndrome S=[s ₀, s ₁..., s _m-1]:

$s_m={\mathrm{\Σ}}_{n=0}^{N-1}{H}_{\mathrm{mn}}{z_n}^T\mathrm{mod}2,m=\mathrm{0,1},...,M-1$

Wherein H _mntest matrix, if syndrome S=0 stops iteration, output hard decision sequence Z also shows successfully decodedly, otherwise enters (b).

(b) to each symbol bits z _n, calculate the number f of the ungratified check equations of its participation _n:

$f_n={\mathrm{\Σ}}_{m=0}^{M-1}{s}_m{H}_{\mathrm{mn}},n=\mathrm{0,1},...,N-1$

If f _n> T, z overturns _n, obtain new hard decision sequence Z.

(c) repeat (a) and (b) until successfully decoded, or reach maximum iteration time and show decoding failure.

Can find out, bit reversal method is simple hard-decision method, only needs logical operation, realizes very simple.

Claims (3)

1. a LDPC code Hard decision decoding method for dynamic threshold bit reversal, determines an initial turn threshold T, it is characterized in that, comprises the following steps:

1) read codeword sequence, according to codeword sequence computing syndrome;

2) judge whether syndrome is zero, if zero, iteration stopped, output codons sequence; Otherwise enter step 3)

3) each symbol bits to each code word in codeword sequence, calculates the number that does not meet check equations that this symbol bits participates in;

4) calculate turn threshold: along with increasing of iterations, turn threshold T successively decreases, and obtains current turn threshold; ;

5) if the number of ungratified check equations is greater than current turn threshold, corresponding symbol bits is overturn, and calculates the syndrome of the rear corresponding codeword sequence of upset;

6) repeating step 2), 3), 4) and 5), until successfully decoded, or while reaching maximum iteration time, output decoding failure.

2. the LDPC code Hard decision decoding method of dynamic threshold bit reversal according to claim 1, is characterized in that current turn threshold T:

T＝j(b-c·m)/a；

In formula: it is that m represents current for which time iteration, j is that symbol bits is corresponding when prostatitis column weight, and a, b, c are experience normal number.

3. the LDPC code Hard decision decoding method of dynamic threshold bit reversal according to claim 1 and 2, determines an initial turn threshold T, it is characterized in that, the step of overturning in iteration is as follows:

A) according to hard decision sequence Z=[z ₀, z ₁..., z _n-1] computing syndrome S=[s ₀, s ₁..., s _m-1]:

$s_m={\mathrm{\Σ}}_{n=0}^{N-1}{H}_{\mathrm{mn}}{z_n}^T\mathrm{mod}2,m=\mathrm{0,1},...,M-1;$

Wherein H _mntest matrix, if syndrome S=0 stops iteration, output hard decision sequence Z also shows successfully decodedly, otherwise enters b);

B) to each symbol bits z _n, calculate the number f of the ungratified check equations of its participation _n:

$f_n={\mathrm{\Σ}}_{m=0}^{M-1}{s}_m{H}_{\mathrm{mn}},n=\mathrm{0,1},...,N-1$

If f _n> T, z overturns _n, obtain new hard decision sequence Z;

C) repeat a) and b) until successfully decoded, or reach maximum iteration time and show decoding failure;

Hard decision sequence is step 5) codeword sequence that obtains after upset.

Images