CN101841340B - Difference upper limit acquiring method and device and viterbi decoding method and device - Google Patents

Abstract

The embodiment of the invention relates to difference upper limit acquiring method and device and viterbi decoding method and device. The difference upper limit acquiring method comprises the following steps of: acquiring a first difference upper limit between accumulated metrics of a first optimal path and a second optimal path from different statuses, wherein at least one of the first optimal path and the second optimal path is a non-optimal path; acquiring a second difference upper limit between accumulated metrics of a third optimal path and a fourth optimal path from different statuses, wherein both the third optimal path and the fourth optimal path are optimal paths from respective status; and obtaining a larger value in the first difference upper limit and the second difference upper limit by comparison. The embodiment of the invention provides the difference upper limits of the accumulated metrics among optimal paths in different statuses based on a four-PLVA (Parallel List Viterbi Algorithm) under the condition of decoding tail-biting and convoluting convolutional codes.

Description

Difference upper limit acquisition methods and device, viterbi coding method and device

Technical field

The embodiment of the invention relates to communication technical field, relates in particular to a kind of difference upper limit acquisition methods and device, viterbi coding method and device.

Background technology

Convolution code can be used Viterbi decoding algorithm, and (Viterbi Decoding Algorithm is called for short: VA) decipher.VA is the optimal solution code calculation of maximum-likelihood sequence estimation, its main thought is that the burst that will receive carries out relevant with all possible emission codeword sequence, then therefrom choose the sequence of likelihood value maximum, VA finishes forward error correction by effectively search for the global optimum path in grid.Yet, in a lot of practical applications, if not only know optimal path, but also know overall sub-optimal path, even know again preferred path etc. of the overall situation, will obtain comparatively significantly improvement in performance.Parallel Concatenated Convolutional Code sub-optimal path Viterbi decoding algorithm (Parallel List Viterbi Decoding Algorithm, be called for short: PLVA) be exactly by seeking simultaneously many globally-optimal selection paths (comprise global optimum path, overall sub-optimal path and the overall situation again preferred path etc.), utilizing the information in many globally-optimal selection paths to improve performance.

In the PLVA algorithm, need to carry out size to the cumulative metric value in different paths and relatively come realizing route to select, the cumulative metric value is that the continuous accumulation calculating by branch metric obtains.The absolute size of cumulative metric value may be very large, but the just relative size between the cumulative metrics value that really works.Therefore when algorithm is realized, do not need to set bit wide according to the absolute size of cumulative metric value, only need to set bit wide according to the relative size of cumulative metric value, can reduce saving resource consumption.Concrete grammar is as follows:

When two fixed-point number A and B do subtraction, if satisfy:

|A-B|＜2 ^n-1 (1)

Then have:

A-B＝((Amod2 ⁿ)-(Bmod2 ⁿ))mod2 ⁿ (2)

Can see, if can determine the difference upper limit of fixed-point number A and B, so just can be to its delivery in the computational process of A and B, and the operation of delivery herein actual be exactly the above data from overflow in its n position of leaving.Therefore, if can estimate in advance the difference upper limit Δ of cumulative metric value, so just the bit wide of cumulative metric value can be set as ceil (log ₂(Δ))+1, its high position of then leaving is overflowed, and can not affect performance.The fixed point bit wide of obtaining by this method is the minimum bit wide under the guaranteed performance prerequisite.

The acquisition methods of the difference upper limit of two kinds of cumulative metric values is provided in the prior art, has been described as follows:

Prior art one has provided the acquisition methods of the difference upper limit between the cumulative metric value in each state optimization path of any time among the VA, its conclusion is: in the base four VA fixed-point algorithms of convolution code, any time, there was the upper bound in the difference of cumulative metric value of optimal path of any two states, and this upper bound is

Wherein the constraint length of convolution coder is M+1, λ _MaxBe the difference upper limit of basic four algorithm branches tolerance, the effect of ceil () function is to ask the smallest positive integral that is not less than given real number. B is the fixed point bit wide of the soft information of channel, and R is the code check of convolution code.But prior art one has only provided the difference upper limit between the cumulative metric value of different conditions optimal path among the VA, and the fixed point bit wide that can only be used for the VA path metric is chosen.

Prior art two has provided the acquisition methods of PLVA difference upper limit of the cumulative metric value of free routing when making zero convolution code decoding, its conclusion is: in basic four LVA fixed-point algorithms, there is the upper bound in the difference that enters the cumulative metric value between the front L bar preferred path of same state, and this upper bound is

Wherein L is the preferred path number among the LVA, obvious L 〉=2, and M+1 is the constraint length of convolutional encoding, λ _MaxBe the difference upper limit of basic four branch metrics,

B is the fixed point bit wide of the soft information of channel, and R is the code check of convolution code.But prior art two has only provided in the situation that the convolution code that makes zero, the difference upper limit of the cumulative metric value among the PLVA between the different preferred paths of same state, and the fixed point bit wide of the path metric of the PLVA of the convolution code that therefore can only be used for making zero is chosen.

For the situation of tail-biting convolutional code among the PLVA, also need to obtain the difference upper limit of the cumulative metric value between the preferred path of different conditions, but the conclusion of prior art one and prior art two can not be applied to all in the basic four PLVA algorithms.

Summary of the invention

The embodiment of the invention provides a kind of difference upper limit acquisition methods and device, viterbi coding method and device, in order to the difference upper limit of the cumulative metric value of the preferred path that obtains different conditions, and viterbi coding method.

The embodiment of the invention provides a kind of difference upper limit acquisition methods, comprising:

Obtain from the first difference upper limit between the cumulative metric value of the first preferred path of different conditions and the second preferred path, at least one is non-optimal path in described the first preferred path and the second preferred path;

Obtain from the second difference upper limit between the cumulative metric value of the 3rd preferred path of different conditions and the 4th preferred path, described the 3rd preferred path and the 4th preferred path are the separately optimal path of state;

Relatively obtain the higher value in the first difference upper limit and the second difference upper limit.

The embodiment of the invention provides a kind of viterbi coding method, comprising:

Obtain the difference upper limit λ of branch metric among the Parallel Concatenated Convolutional Code sub-optimal path Viterbi decoding algorithm PLVA _Max

Obtain the constraint length M+1 of preferred path number L and convolution coder;

According to described difference upper limit λ _Max, described preferred path number L and described convolution coder constraint length M+1, obtain difference upper limit and be

Determine the fixed point bit wide according to described difference upper limit, carry out decoded operation according to described fixed point bit wide.

The embodiment of the invention provides a kind of difference upper limit deriving means, comprising:

The first acquisition module is used for obtaining from the first difference upper limit between the cumulative metric value of the first preferred path of different conditions and the second preferred path, and at least one is non-optimal path in described the first preferred path and the second preferred path;

The second acquisition module is used for obtaining from the second difference upper limit between the cumulative metric value of the 3rd preferred path of different conditions and the 4th preferred path, and described the 3rd preferred path and the 4th preferred path are the separately optimal path of state;

Comparison module is for the higher value that relatively obtains the first difference upper limit and the second difference upper limit.

The embodiment of the invention provides a kind of Viterbi decoding device, comprising:

The 3rd acquisition module is for the difference upper limit λ that obtains Parallel Concatenated Convolutional Code sub-optimal path Viterbi decoding algorithm PLVA branch metric _Max

The 4th acquisition module is for the constraint length M+1 that obtains preferred path number L and convolution coder;

The 5th acquisition module is used for according to described difference upper limit λ _Max, described preferred path number L and described convolution coder constraint length M+1, obtain difference upper limit and be

The 6th acquisition module is used for according to described difference upper limit, obtains the fixed point bit wide;

Decoding module is used for according to described fixed point bit wide, carries out decoded operation.

The difference upper limit acquisition methods that the embodiment of the invention provides and device, viterbi coding method and device, can in basic four PLVA algorithms, obtain the difference upper limit of cumulative metric value of the preferred path of different conditions, and computational process is simple, has simplified algorithm complex, has saved hardware resource.

Description of drawings

In order to be illustrated more clearly in the technical scheme in the embodiment of the invention, the below will do to introduce simply to the accompanying drawing of required use among the embodiment, apparently, accompanying drawing in the following describes only is some embodiments of the present invention, for those of ordinary skills, under the prerequisite of not paying creative work, can also obtain according to these accompanying drawings other accompanying drawing.

The schematic flow sheet of the difference upper limit acquisition methods that Fig. 1 provides for the embodiment of the invention one;

The schematic flow sheet of the difference upper limit acquisition methods that Fig. 2 provides for the embodiment of the invention two;

The schematic flow sheet of the viterbi coding method that Fig. 3 provides for the embodiment of the invention three;

The structural representation of the difference upper limit deriving means that Fig. 4 provides for the embodiment of the invention four;

The structural representation of the Viterbi decoding device that Fig. 5 provides for the embodiment of the invention five.

Embodiment

Below in conjunction with the accompanying drawing in the embodiment of the invention, the technical scheme in the embodiment of the invention is clearly and completely described, obviously, described embodiment only is the present invention's part embodiment, rather than whole embodiment.Based on the embodiment among the present invention, those of ordinary skills belong to the scope of protection of the invention not making the every other embodiment that obtains under the creative work prerequisite.

The flow chart of the difference upper limit acquisition methods that Fig. 1 provides for the embodiment of the invention one.As shown in Figure 1, the present embodiment comprises:

Step 101, obtain from the first difference upper limit between the cumulative metric value of the first preferred path of different conditions and the second preferred path, at least one is non-optimal path in this first preferred path and the second preferred path;

Step 102, obtain from the second difference upper limit between the cumulative metric value of the 3rd preferred path of different conditions and the 4th preferred path, the 3rd preferred path and the 4th preferred path are the separately optimal path of state;

Step 103, relatively obtain higher value in the first difference upper limit and the second difference upper limit.

Take basic four PLVA algorithms as example, in embodiments of the present invention, basic four PLVA algorithms have three in the situation that tail-biting convolutional code decoding needs to carry out cumulative metric value size scene relatively, are respectively the first scene, the second scene and the 3rd scene.

1) the first scene is the scene that compares between the cumulative metric value of all preferred paths of arbitrary moment entering arbitrary state.

For basic four PLVA algorithms, arbitrary moment will have the preferred path of 4 states to enter arbitrary state, if the number of the preferred path of a certain state is L, then to enter the number of all preferred paths of arbitrary state be 4L arbitrary moment, needs in the first scene the cumulative metric value of this 4L preferred path is compared.

2) the second scene is before recalling, the scene that compares between the cumulative metric value of institute's stateful optimal path.

3) the 3rd scene is for selecting L preferred condition, the scene that compares between the cumulative metric value of all preferred paths of a described L preferred condition in the second scene.

After the cumulative metric value of completeer institute stateful optimal path, according to comparative result, select the less state of the cumulative metric value of a front L optimal path as L preferred condition, need in the 3rd scene cumulative metric value of the L*L bar preferred path of this L preferred condition is compared.

Because in above-mentioned three kinds of scenes, what participate in comparing all is the cumulative metric value, so the embodiment of the invention need to decide according to the maximum of difference upper limit in these three kinds of scenes the fixed point bit wide of cumulative metric.Wherein, the first scene and the 3rd scene can adopt method shown in Figure 1 to obtain difference upper limit under the scene separately.

In above-mentioned the first scene, the difference upper limit acquisition methods that can adopt the embodiment of the invention to provide obtains the difference upper limit of the first scene.Be described as follows.

In the first scene, need to carry out size relatively to the cumulative metric value that arbitrary moment enters the 4L bar preferred path of arbitrary state, select the less preferred path of L bar cumulative metric value as the preferred path of current state.This 4L bar preferred path is established it and is divided into 4 groups respectively from 4 different states, and the preferred path of each group is from same state, and for each group, the L bar preferred path that belongs to this group is arranged sequentially good according to from small to large.Owing to only need to therefrom select the less preferred path of L bar cumulative metric value, therefore participating in preferred path number reality relatively less than the 4L bar.Be described further below by a concrete example.

Suppose L=4, then the magnitude relationship between 4L bar preferred path and the cumulative metric value thereof is respectively:

PM _A[1]≤PM _A[2]≤PM _A[3]≤PM _A[4]

PM _B[1]≤PM _B[2]≤PM _B[3]≤PM _B[4]

PM _C[1]≤PM _C[2]≤PM _C[3]≤PM _C[4]

PM _D[1]≤PM _D[2]≤PM _D[3]≤PM _D[4](3)

Wherein, A, B, C and D represent 4 states, PM _A[1] expression is from the cumulative metric value of the optimal path of state A, PM _A[2] expression is from the cumulative metric value in the less preferred path of state A, PM _A[3] expression is from the cumulative metric value of the again preferred path of state A, PM _A[4] expression is from the cumulative metric value in the again less preferred path of state A, and other states similarly.

The comparison procedure of the first scene is as follows:

i＝1；j＝1；k＝1；l＝1；

for(m＝0；m＜L；m++)

{

x ₁＝PM _A[i]；x ₂＝PM _B[j]；x ₃＝PM _C[k]；x ₄＝PM _D[l]；

if(min(x ₁，x ₂，x ₃，x ₄)＝＝x ₁)

i++；

else if(min(x ₁，x ₂，x ₃，x ₄)＝＝x ₂)

j++；

else if(min(x ₁，x ₂，x ₃，x ₄)＝＝x ₃)

k++；

else if(min(x ₁，x ₂，x ₃，x ₄)＝＝x ₄)

l++；

}

Above-mentioned comparison procedure is one section C programmer, can obtain the less preferred path of L bar cumulative metric value by carrying out this section C programmer.From comparison procedure, can find following features:

1, when any one a group preferred path because the cumulative metric value is minimum after selected, the cumulative metric value certainty of picking out the preferred path that participation compares next time from this group is greater than the cumulative metric value of current selected preferred path.

2, participate in for the first time i relatively, j, k, the value of four labels of l must be { 1,1,1,1}, and the last i relatively that participates in, j, k, the value of four labels of l might be for { 1,1,1, L}, this maximum that difference between label of selected preferred path also has been described is L-1.

Always at first select little cumulative metric value (the cumulative metric value of namely picking out is always little than the cumulative metric value of not selecting) in above-mentioned first characteristics explanation comparison procedure at every turn, then the larger cumulative metric value of cumulative metric value of picking out with a ratio is come and was selected remaining continuation last time relatively, and this cumulative metric value of newly picking out is less with the difference of the cumulative metric value of being left with the cumulative metric value that its fullsized of difference of the cumulative metric value of being left is picked out at the beginning.

Based on above-mentioned analysis, in conjunction with the conclusion of prior art one and prior art two, the embodiment of the invention two provides the acquisition methods of the difference upper limit of the first scene.

The flow chart of the difference upper limit acquisition methods that Fig. 2 provides for the embodiment of the invention two.As shown in Figure 2, the method comprises:

Step 201, obtain from the 3rd difference upper limit between the cumulative metric value of the preferred path of same state.

According to the conclusion of prior art two, there is the upper bound from same state and the difference upper limit that enters the cumulative metric value between the front m bar preferred path of same state, this upper bound (i.e. the 3rd difference upper limit) is:

Take from same state A as example, then have:

${\mathrm{PM}}_A\left[m\right]-{\mathrm{PM}}_A[m-1]\≤(m-1)(\mathrm{ceil}\left(\frac{M}{2}\right)+1){\mathrm{\λ}}_{\max }---\left(4\right)$

Wherein, the constraint length of convolution coder is M+1, λ _MaxDifference upper limit for basic four algorithm branches tolerance.

Step 202, according to the 3rd difference upper limit, obtain from the first difference upper limit between the cumulative metric value of the first preferred path of different conditions and the second preferred path, wherein at least one is non-optimal path in the first preferred path and the second preferred path.

If the first preferred path is the m preferred path from state A, the second preferred path is the n preferred path from state B, in m preferred path and the n preferred path at least one be not its from the optimal path of state, that is to say to have one among m and the n at least greater than 1.

Obviously, in comparison procedure before, must once satisfy PM _A[m-1]＜PM _B[n] or PM _A[n-1]＜PM _B[m]; Might as well establish PM _A[m-1]＜PM _B[n] then has according to formula (4):

${\mathrm{PM}}_A\left[m\right]-{\mathrm{PM}}_B\left[n\right]<{\mathrm{PM}}_A\left[m\right]-{\mathrm{PM}}_A[m-1]\&le;(m-1)(\mathrm{ceil}\left(\frac{M}{2}\right)+1){\mathrm{\&lambda;}}_{\max }---\left(5\right)$

Consider and have PM _A[n-1]＜PM _BThe situation of [m], can obtain from the first difference upper limit between the cumulative metric value of the first preferred path of different conditions and the second preferred path be:

$(\max (m,n)-1)(\mathrm{ceil}\left(\frac{M}{2}\right)+1){\mathrm{\&lambda;}}_{\max }.$

Step 203, obtain from the second difference upper limit between the cumulative metric value of the 3rd preferred path of different conditions and the 4th preferred path, wherein the 3rd preferred path and the 4th preferred path are the separately optimal path of state.

Because the 3rd preferred path and the 4th preferred path are the separately optimal path of state, conclusion according to prior art one, there is the upper bound in the difference of the cumulative metric value of the optimal path of the arbitrary moment two different conditions, and this upper bound (i.e. the second difference upper limit) is:

Above-mentioned the first preferred path, the second preferred path, the 3rd preferred path or the 4th preferred path be in the basic four PLVA algorithms arbitrary moment enter one of all preferred paths of arbitrary state, namely be the preferred path in the first scene.

Step 204, relatively obtain higher value in the first difference upper limit and the second difference upper limit, obtain thus the difference upper limit of the first scene.

More above-mentioned the first difference upper limit and the second difference upper limit as can be known, the first difference upper limit is higher value.Further, in the first scene, the cumulative metric value of actual participation comparison is cumulative metric value and the branch metric sum of preferred path, and max (m, n)≤L is arranged again, and the difference upper limit that can obtain the first scene should be:

$(L-1)\&times;(\mathrm{ceil}\left(\frac{M}{2}\right)+1){\mathrm{\&lambda;}}_{\max }+{\mathrm{\&lambda;}}_{\max }.$

The fixed point bit wide of the cumulative path metrics that the difference upper limit of above-mentioned the first scene is corresponding should be

Wherein

B is the fixed point bit wide of the soft information of channel, and R is the code check of convolution code.

In the second scene, only need the difference upper limit of the cumulative metric value of the optimal path between the consideration different conditions, can directly utilize the conclusion of prior art one, the difference upper limit that draws the second scene is

Wherein, the constraint length of convolution coder is M+1, λ _MaxBe the difference upper limit of basic four algorithm branches tolerance, the bit wide of the cumulative path metrics that it is corresponding should be

The situation of the situation of the 3rd scene and the first scene is similar, it all is the cumulative metric value that needs to compare from the preferred path of different conditions, the acquisition methods of the difference upper limit of the 3rd scene can referring to description corresponding to Fig. 2, also can directly utilize the analysis result of the first scene.

If utilize the analysis result of the first scene, what it should be noted that the first scene participation comparison is cumulative metric value and the branch metric sum of preferred path, and the 3rd scene participate in relatively be the cumulative metric value of preferred path, therefore the difference upper limit of the 3rd scene should be

Wherein, the constraint length of convolution coder is M+1, λ _MaxBe the difference upper limit of basic four PLVA algorithm branches tolerance, the fixed point bit wide of the cumulative path metrics that it is corresponding should be Wherein B is the fixed point bit wide of the soft information of channel, and R is the code check of convolution code.。

The analysis result of comprehensive above-mentioned the first scene, the second scene and the 3rd scene, the difference upper limit that can find the first scene is maximum, therefore basic four PLVA algorithms in the situation that the difference upper limit of tail-biting convolutional code decoding cumulative metric value should be:

The embodiment of the invention has provided basic four PLVA algorithms in the situation that tail-biting convolutional code decoding needs to carry out cumulative metric value size scene relatively, obtain respectively the difference upper limit of three scenes, and then obtain the maximum of the difference upper limit of three scenes, with the difference upper limit of this maximum as cumulative metric value between the preferred path of different conditions.

Further, obtaining the fixed point bit wide is the most important condition of Viterbi decoding, and the algorithm calculated difference upper limit and the fixed point bit wide that adopt the present embodiment to provide in viterbi decoder also can be saved hardware resource, thereby be simplified the process of Viterbi decoding.

The flow chart of the viterbi coding method that Fig. 3 provides for the embodiment of the invention three.As shown in Figure 3, the present embodiment comprises:

Step 301, obtain the difference upper limit λ of branch metric in the PLVA algorithm _Max

Step 302, obtain the constraint length M+1 of preferred path number L and convolution coder;

Step 303, according to difference upper limit λ _Max, preferred path number L and convolution coder constraint length M+1, obtain difference upper limit and be

Step 304, determine the fixed point bit wide according to difference upper limit, carry out decoded operation according to the fixed point bit wide.Further, the fixed point bit wide of determining according to above-mentioned difference upper limit is specially:

$\mathrm{ceil}\left({\log }_2(((L-1)\&times;(\mathrm{ceil}\left(\frac{M}{2}\right)+1)+1)\&times;{\mathrm{\&lambda;}}_{\max })\right)+1.$

The present embodiment is applied to above-mentioned analysis result in the decode procedure of VA.Specifically, according to above-mentioned analysis result, at first obtain the difference upper limit λ of branch metric in the PLVA algorithm _Max, preferred path number L and convolution coder constraint length M+1; Then, the above-mentioned parameter substitution formula that obtains is obtained difference upper limit; Determine the fixed point bit wide according to this difference upper limit; This fixed point bit wide has determined that all data should adopt the bit sequence of how much length to represent in the VA decode procedure, for example: represent 3 with 011, or represent 3 with 0011, definite according to the fixed point bit wide; Therefore, obtained after the fixed point bit wide, just can begin to carry out decoded operation.

In the decode procedure of viterbi decoder, the method that employing the present embodiment provides is according to the difference upper limit λ of branch metric in the PLVA algorithm _Max, preferred path number L and convolution coder constraint length M+1, the above-mentioned formula of substitution can obtain difference upper limit, obtains the fixed point bit wide according to difference upper limit.The present embodiment utilizes above-mentioned formula, and computational process is simple, has simplified algorithm complex, has saved hardware resource.Above-mentioned fixed point bit wide can be under the prerequisite that guaranteed performance is not lost, resource consumption is reduced to minimum, thereby saved the cost of product.

The structural representation of the difference upper limit deriving means that Fig. 4 provides for the embodiment of the invention four.As shown in Figure 4, the present embodiment comprises: the first acquisition module 11, the second acquisition module 12 and comparison module 13, wherein:

The first acquisition module 11 is used for obtaining from the first difference upper limit between the cumulative metric value of the first preferred path of different conditions and the second preferred path, and at least one is non-optimal path in this first preferred path and the second preferred path;

The second acquisition module 12 is used for obtaining from the second difference upper limit between the cumulative metric value of the 3rd preferred path of different conditions and the 4th preferred path, and the 3rd preferred path and the 4th preferred path are the separately optimal path of state;

Comparison module 13 is used for relatively obtaining the higher value of the first difference upper limit and the second difference upper limit.

Further, the first acquisition module 11 is concrete to be used for according to from the 3rd difference upper limit between the cumulative metric value of the preferred path of same state, obtains from the first difference upper limit between the cumulative metric value of the first preferred path of different conditions and the second preferred path.

According to the associated description of said method embodiment, need to carry out cumulative metrics value size scene relatively in the basic four PLVA algorithms and have three, be respectively above-mentioned the first scene, the second scene and the 3rd scene.Wherein, the first scene and the 3rd scene can adopt device shown in Figure 4 to obtain difference upper limit under the scene separately.

Specifically, in the first scene, need to carry out size relatively to the cumulative metric value that arbitrary moment enters the 4L bar preferred path of arbitrary state, select the less preferred path of L bar cumulative metric value as the preferred path of current state.The first acquisition module 11 is according to from the 3rd difference upper limit between the cumulative metric value of the preferred path of same state, obtains from the first difference upper limit between the cumulative metric value of the first preferred path of different conditions and the second preferred path.Wherein the 3rd difference upper limit can obtain according to the conclusion of prior art two, is specially

The first acquisition module 11 obtains the first difference upper limit according to formula (5)

The second acquisition module 12 is according to the conclusion of prior art one, obtains the second difference upper limit to be

The higher value that comparison module 13 relatively obtains in the first difference upper limit and the second difference upper limit is the first difference upper limit.The difference upper limit that can obtain the first scene according to the first difference upper limit is

The situation of the situation of the 3rd scene and the first scene is similar, all is the cumulative metric value that needs to compare from the preferred path of different conditions.If utilize the analysis result of the first scene, the difference upper limit that obtains the 3rd scene should be

The device that the present embodiment provides can be used for realizing said method embodiment that the implementation procedure of each module concrete function of difference upper limit deriving means can with reference to the associated description of said method embodiment, repeat no more.

The difference upper limit deriving means that the present embodiment provides can obtain the difference upper limit of the first scene and the 3rd scene, thereby make it possible to obtain the maximum of the difference upper limit of three scenes, with the difference upper limit of this maximum as cumulative metric value between the preferred path of different conditions.

The structural representation of the Viterbi decoding device that Fig. 5 provides for the embodiment of the invention five.As shown in Figure 5, the present embodiment comprises: the 3rd acquisition module 21, the 4th acquisition module 22, the 5th acquisition module 23, the 6th acquisition module 24 and decoding module 25, wherein:

The 3rd acquisition module 21 is used for obtaining the difference upper limit λ of PLVA branch metric _Max

The 4th acquisition module 22 is used for obtaining the constraint length M+1 of preferred path number L and convolution coder;

The 5th acquisition module 23 is used for according to difference upper limit λ _Max, preferred path number L and convolution coder constraint length M+1, obtain difference upper limit and be

The 6th acquisition module 24 is used for according to difference upper limit, obtains the fixed point bit wide;

Decoding module 25 is used for carrying out decoded operation according to the fixed point bit wide.

Further, the 6th acquisition module 24 specifically is used for according to difference upper limit, obtains the fixed point bit wide and is

$\mathrm{ceil}\left({\log }_2(((L-1)\&times;(\mathrm{ceil}\left(\frac{M}{2}\right)+1)+1)\&times;{\mathrm{\&lambda;}}_{\max })\right)+1.$

The present embodiment is applied to the analysis result of embodiment of the method in the decode procedure of VA.Specifically, according to above-mentioned analysis result, the 3rd acquisition module 21 obtains the difference upper limit λ of branch metric among the PLVA _Max, the 4th acquisition module 22 obtains the constraint length M+1 of preferred path number L and convolution coder; The 5th acquisition module 23 obtains difference upper limit with the above-mentioned parameter substitution formula that obtains; The 6th acquisition module 24 is used for obtaining the fixed point bit wide according to difference upper limit, this fixed point bit wide has determined that all data should adopt the bit sequence of how much length to represent in the VA decode procedure, for example: represent 3 with 011, or represent 3 with 0011, definite according to the fixed point bit wide; Decoding module 25 carries out decoded operation according to the fixed point bit wide.

In the decode procedure of VA, the device that employing the present embodiment provides is according to the difference upper limit λ of branch metric in the PLVA algorithm _Max, preferred path number L and convolution coder constraint length M+1, the above-mentioned formula of substitution can obtain difference upper limit, obtains the fixed point bit wide according to difference upper limit.The present embodiment utilizes above-mentioned formula, and computational process is simple, has simplified algorithm complex, has saved hardware resource.Above-mentioned fixed point bit wide can be under the prerequisite that guaranteed performance is not lost, resource consumption is reduced to minimum, thereby saved the cost of product.

One of ordinary skill in the art will appreciate that: all or part of step that realizes said method embodiment can be finished by the relevant hardware of program command, aforesaid program can be stored in the computer read/write memory medium, this program is when carrying out, execution comprises the step of said method embodiment, and aforesaid storage medium comprises: the various media that can be program code stored such as ROM, RAM, magnetic disc or CD.

It should be noted that at last: above embodiment only in order to the technical scheme of the embodiment of the invention to be described, is not intended to limit; Although with reference to previous embodiment the embodiment of the invention is had been described in detail, those of ordinary skill in the art is to be understood that: it still can be made amendment to the technical scheme that aforementioned each embodiment puts down in writing, and perhaps part technical characterictic wherein is equal to replacement; And these modifications or replacement do not make the essence of appropriate technical solution break away from the spirit and scope of each embodiment technical scheme of the embodiment of the invention.

Claims (6)

1. a difference upper limit acquisition methods is characterized in that, comprising:

According to from the 3rd difference upper limit between the cumulative metric value of the preferred path of same state, obtain from the first difference upper limit between the cumulative metric value of the first preferred path of different conditions and the second preferred path, at least one is non-optimal path in described the first preferred path and the second preferred path;

Obtain from the second difference upper limit between the cumulative metric value of the 3rd preferred path of different conditions and the 4th preferred path, described the 3rd preferred path and the 4th preferred path are the separately optimal path of state;

Relatively obtain the higher value in the first difference upper limit and the second difference upper limit.

2. difference upper limit acquisition methods according to claim 1, it is characterized in that, described the first preferred path, the second preferred path, the 3rd preferred path or the 4th preferred path be among the basic four Parallel Concatenated Convolutional Code sub-optimal path Viterbi decoding algorithm PLVA arbitrary moment enter one of all preferred paths of arbitrary state

；

The higher value of described the first difference upper limit and the second difference upper limit is

Wherein, the constraint length of convolution coder is M+1, λ _MaxBe the difference upper limit of branch metric among basic four PLVA, L is the preferred path number, and the effect of ceil function is to ask the smallest positive integral that is not less than given real number.

3. difference upper limit acquisition methods according to claim 1, it is characterized in that, described the first preferred path, the second preferred path, the 3rd preferred path or the 4th preferred path are one of all preferred paths of L preferred condition selecting in the second scene; Described the second scene be among the basic four Parallel Concatenated Convolutional Code sub-optimal path Viterbi decoding algorithm PLVA before recalling, the scene that compares between the cumulative metric value of institute's stateful the first preferred path;

The higher value of described the first difference upper limit and the second difference upper limit is

Wherein, the constraint length of convolution coder is M+1, λ _MaxBe the difference upper limit of branch metric among the basic four Parallel Concatenated Convolutional Code sub-optimal path Viterbi decoding algorithm PLVA, the effect of ceil function is to ask the smallest positive integral that is not less than given real number, and L is the preferred path number.

4. a viterbi coding method is characterized in that, comprising:

Obtain the difference upper limit λ of branch metric among the Parallel Concatenated Convolutional Code sub-optimal path Viterbi decoding algorithm PLVA _Max

Obtain the constraint length M+1 of preferred path number L and convolution coder;

According to described difference upper limit λ _Max, described preferred path number L and described convolution coder constraint length M+1, obtain difference upper limit and be

Determine the fixed point bit wide according to described difference upper limit, carry out decoded operation according to described fixed point bit wide;

Described fixed point bit wide is

The effect of ceil function is to ask the smallest positive integral that is not less than given real number.

5. a difference upper limit deriving means is characterized in that, comprising:

The first acquisition module, be used for basis from the 3rd difference upper limit between the cumulative metric value of the preferred path of same state, obtain from the first difference upper limit between the cumulative metric value of the first preferred path of different conditions and the second preferred path, at least one is non-optimal path in described the first preferred path and the second preferred path;

The second acquisition module is used for obtaining from the second difference upper limit between the cumulative metric value of the 3rd preferred path of different conditions and the 4th preferred path, and described the 3rd preferred path and the 4th preferred path are the separately optimal path of state;

Comparison module is for the higher value that relatively obtains the first difference upper limit and the second difference upper limit.

6. a Viterbi decoding device is characterized in that, comprising:

The 3rd acquisition module is for the difference upper limit λ that obtains Parallel Concatenated Convolutional Code sub-optimal path Viterbi decoding algorithm PLVA branch metric _Max

The 4th acquisition module is for the constraint length M+1 that obtains preferred path number L and convolution coder;

The 5th acquisition module is used for according to described difference upper limit λ _Max, described preferred path number L and described convolution coder constraint length M+1, obtain difference upper limit and be

The 6th acquisition module is used for according to described difference upper limit, obtains the fixed point bit wide to be The effect of ceil function is to ask the smallest positive integral that is not less than given real number.

Images