Novel Turbo code decoding state value normalization method

Abstract

The invention discloses a novel Turbo code decoding state value normalization method, which comprises the following steps of: calculating N state values for a first bit of an input component decoder to form a state value sequence X; the state value sequence X calculated in the first step is compared with the maximum value in the positive direction and the minimum value in the negative direction one by one, the state values larger than the maximum value form a mark sequence Fmax, and the state values smaller than the minimum value form a mark sequence Fmin; judging the flag sequences Fmax and Fmin obtained in the step 2 to obtain a new state value sequence X '; taking the new state value sequence X'as input, and calculating N state values of the next bit; and repeating the step 2 to the step 4 to complete the calculation of the forward recursion values and the backward recursion values of all bits so as to complete the whole decoding process. According to the method, only fixed two-stage running water required by the state value needs to be calculated, and the time delay is greatly reduced.

Description

Technical Field

[0001] The invention belongs to the technical field of Turbo coding and error correction coding, and relates to a novel Turbo code decoding state value normalization method. Background Art

[0002] In wireless communication, coding is usually used to achieve reliable data transmission under low signal-to-noise ratio. Turbo coding is a commonly used forward error correction (FEC) technology that combines convolutional codes with random interleavers. While achieving random coding, it uses interleavers to construct long codes from short codes, and uses soft output iterative decoding to approximate maximum likelihood decoding. Turbo codes make full use of the basic conditions of Shannon's channel coding theorem and obtain performance close to the Shannon limit, so they have been widely used.

[0003] Compared with the coding structure of Turbo code, the decoding process of Turbo code is more complicated. Figure 1 As shown. The decoding process of Turbo code is a cyclic iterative decoding process. In the forward recursive and backward recursive operations in the component decoder, in order to prevent data overflow, it is necessary to normalize the several state values ​​of each bit, that is, find the maximum value among these state values, and then subtract this maximum value from each state value. This step requires several levels of pipeline to complete, so the delay of the component decoder is very large. In addition to multiple iterative decoding, this high-latency decoding scheme is not applicable in some specific systems. Summary of the invention

[0004] (I) Purpose of the invention

[0005] The purpose of the present invention is to solve the problem of long delay in traditional Turbo code decoding, and propose a new Turbo code decoding state value normalization method. Based on the idea of ​​maximum directional overflow, the pipeline calculation of traditional normalization processing is omitted, and the decoding delay is greatly reduced without damaging the decoding performance.

[0006] (II) Technical solution

[0007] In order to solve the above technical problems, the present invention proposes a new Turbo code decoding state value normalization method to solve the defect that the traditional Turbo code decoding state value normalization processing pipeline leads to a large overall decoding delay. In the Turbo code decoding process, each component decoder needs to calculate the forward recursive value and the backward recursive value, which adopts the same normalization processing method; the processing process starts with the component decoder of the Turbo code decoding calculating N state values ​​for a certain bit, and referring to Figure 1As shown, the method of the present invention comprises the following steps:

[0008] Step 1: Calculate N state values ​​X = {X1, X2, ..., X n}, n = 1, 2, ... N;

[0009] Step 2: Compare the state value sequence X calculated in step 1 with the maximum value X in the positive direction one by one. max and the minimum value X in the negative direction min Compare: If it is greater than the maximum value, the flag F max_n Set to 1, otherwise set to 0, thus obtaining the flag sequence F max ={F max_1 ,F max_2 ,…,F max_n}; If it is less than the minimum value, the flag F min_n Set to 1, otherwise set to 0, thus obtaining the flag sequence F min ={F min_1 ,F min_2 ,…,F min_n};

[0010] Step 3: For the flag sequence F obtained in step 2 max and F min Make a judgment: If all are 0, then the state value X is not processed, and the new state value sequence X'=X; if F max If it is not all 0, then each state variable of state value X is subtracted from X. max , and obtain a new state value sequence X'={X1',X2',...,X n '}; if F min If it is not all 0, then each state variable of state value X is subtracted from X. min , and obtain a new state value sequence X'={X1',X2',...,X n '};

[0011] Step 4: Take the new state value sequence X' obtained in step 3 as input and calculate the N state values ​​of the next bit;

[0012] Step 5: Repeat steps 2 to 4 to complete the calculation of the forward recursive value and the backward recursive value of all bits, thereby completing the entire decoding process.

[0013] in:

[0014] (1) Maximum value X in the positive direction max and the minimum value X in the negative direction min The setting can be calculated based on the quantization bit width of the data during the decoding process. Generally, it can be set to half of the maximum value and the minimum value that the quantization bit width can represent.

[0015] (2) Step 2 can be calculated synchronously when calculating the state value, without the need for additional pipeline delay;

[0016] (3) Step 3 can be implemented using combinational logic without pipeline delay.

[0017] (III) Beneficial effects

[0018] The novel Turbo code decoding state value normalization method provided by the above technical solution has the following beneficial effects compared with the traditional state value normalization processing operation:

[0019] (1) Greatly reduced decoding delay: In the traditional normalization operation, calculating all the state values ​​of a bit requires two fixed pipelines, then finding the maximum value requires several pipelines, and finally subtracting the maximum value requires one pipeline; while the method of the present invention only requires two fixed pipelines required to calculate the state value, and the delay is greatly reduced;

[0020] (2) While the decoding delay is reduced, the decoding performance does not degrade and is slightly better than the traditional operation method, such as Figure 2 shown. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A schematic diagram of Turbo decoding of a method for normalizing Turbo code decoding state values ​​according to an embodiment of the present invention;

[0022] Figure 2 This is a curve diagram of the bit error rate changing with Eb / N0 under different signal-to-noise ratios according to the new Turbo code decoding state value normalization method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, content and advantages of the present invention more clear, the specific implementation methods of the present invention are further described in detail below in conjunction with the drawings and examples.

[0024] This embodiment describes the application of the new Turbo code decoding state value normalization method of the present invention to a 1 / 3 code rate Turbo coding example in which the encoder uses 3 registers. At this time, the number of state values ​​calculated for each bit is N=2^3=8, the information sequence length is 1024, and the encoded length is 3072.

[0025] Step 1: The demodulated signal is expressed as [-104, 12, 20, -18, -83, 115, 35, -124, 53, -16…] with a total of 3072 symbols sent to the decoder. The calculation process is uniformly expressed with 18-bit quantization width, so the maximum value X in the positive direction is maxCan be set to 2^16, the minimum value X in the negative direction min It can be set to -2^16. The 8 state value sequence X calculated by the first bit of the input is [1515, -1412, 877, 979, -1292, -1900, -784, 1486];

[0026] Step 2: Compare the state value sequence X obtained in step 1 with the maximum value X in the positive direction one by one max and the minimum value X in the negative direction min By comparison, it is found that it is neither greater than the maximum value nor less than the minimum value, so the flag sequence F max is [0,0,0,0,0,0,0,0],F min Also [0,0,0,0,0,0,0,0];

[0027] Step 3: Because F max and F min If all are 0, the state value X is not processed, and the new state value sequence X' is [1515, -1412, 877, 979, -1292, -1900, -784, 1486];

[0028] Step 4: Take the new state value sequence X' obtained in step 3 as input and calculate the 8 state values ​​of the next bit [-2616, -3653, 2154, 1770, 2407, 3303, -2756, -3984];

[0029] Step 5: Repeat steps 2 to 4 to complete the calculation of the forward recursive value and the backward recursive value of all bits, thereby completing the entire decoding process.

[0030] In this embodiment, there are the following notable features:

[0031] (1) The 8 state values ​​calculated for the 103rd bit are [56550, -17831, -9960, 113866, 34586, 52404, -24116, -62253]. After step 2, we get: max is [0,0,0,1,0,0,0,0],F min is [0,0,0,0,0,0,0,0]; because F max If it is not all 0, then each state variable of the state value X is uniformly subtracted by 2^16 to obtain a new state value sequence X' of [-8986, -83367, -75496, 48330, -30950, -13132, -89652, -127789], which is used as input to calculate the 8 state values ​​of the next bit;

[0032] (2) The 8 state values ​​calculated for the 187th bit are [22150, 61710, -102270, -107171, 40154, 38842, -48305, 24058]. After step 2, we get: max is [0,0,0,0,0,0,0,0],F min is [0,0,1,1,0,0,0,0]; because F min If it is not all 0, then -2^16 is subtracted from each state variable of the state value X to obtain a new state value sequence X' of [87686, 127246, -36734, -41635, 105690, 104378, 17231, 89594], which is used as input to calculate the 8 state values ​​of the next bit;

[0033] (3) The delay of a single component decoder of the inventive method is 1024×2=2048 clks, while the delay of a single component decoder of the conventional method is 1024×6=6144, which is reduced by 2 / 3.

[0034] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A novel method for normalizing Turbo code decoding state values, characterized in that: The normalization process starts after the component decoder of the Turbo code decoder calculates N state values ​​for a certain bit, and includes the following steps: Step 1: Calculate N state values ​​for the first bit of the input component decoder to form a state value sequence X; Step 2: Compare the state value sequence X calculated in step 1 with the maximum value in the positive direction and the minimum value in the negative direction one by one. The state values ​​greater than the maximum value form the flag sequence F. max , the state value less than the minimum value forms the flag sequence F min ; Step 3: For the flag sequence F obtained in step 2 max and F min Make a judgment and obtain a new state value sequence X'; Step 4: Take the new state value sequence X' obtained in step 3 as input and calculate the N state values ​​of the next bit; Step 5: Repeat steps 2 to 4 to complete the calculation of the forward recursive value and the backward recursive value of all bits, thereby completing the entire decoding process.

2. The novel Turbo code decoding state value normalization method as claimed in claim 1, characterized in that: During the Turbo code decoding process, each component decoder calculates the forward recursive value and the backward recursive value, using the same normalization processing method.

3. The novel Turbo code decoding state value normalization method as claimed in claim 2, characterized in that: In step 1, the state value sequence is recorded as: X = {X1, X2, ..., X n },n=1,2,…N.

4. The novel Turbo code decoding state value normalization method as claimed in claim 3, characterized in that: In step 2, the state value sequence X is compared with the maximum value X in the positive direction one by one. max Do a comparison, if it is greater than the maximum value, then the flag F max_n Set to 1, otherwise set to 0, thus obtaining the flag sequence F max ={F max_1 ,F max_2 ,…,F max_n }.

5. The novel Turbo code decoding state value normalization method as claimed in claim 4, characterized in that: In step 2, the state value sequence X is compared with the minimum value X in the negative direction one by one. min Make a comparison. If it is less than the minimum value, the flag F min_n Set to 1, otherwise set to 0, thus obtaining the flag sequence F min ={F min_1 ,F min_2 ,…,F min_n }.

6. The novel Turbo code decoding state value normalization method as claimed in claim 5, characterized in that: In step 3, the flag sequence F max and F min When all are 0, the state value X is not processed, and the new state value sequence X'=X.

7. The novel Turbo code decoding state value normalization method as claimed in claim 6, characterized in that: In step 3, if F max If it is not all 0, then each state variable of state value X is subtracted from X. max , and obtain a new state value sequence X'={X1',X2',…,X n '}.

8. The novel Turbo code decoding state value normalization method as claimed in claim 7, characterized in that: In step 3, if F min If it is not all 0, then each state variable of state value X is subtracted from X. min , and obtain a new state value sequence X'={X1',X2',…,X n '}.

9. The novel Turbo code decoding state value normalization method as claimed in claim 8, characterized in that: Maximum value X in the positive direction max and the minimum value X in the negative direction min The setting is calculated based on the quantization bit width of the data during the decoding process.

10. The novel Turbo code decoding state value normalization method as claimed in claim 9, characterized in that: Maximum value X in the positive direction max and the minimum value X in the negative direction min Set to half of the maximum and minimum values ​​that can be represented by the quantization bit width respectively.

Images