KR20030082699A - Turbo internal interleaving method and device based on golden sequence - Google Patents

Abstract

PURPOSE: A method for interleaving the turbo inside by using a golden sequence and a device for the same are provided to improve the performance in comparison with the block interleaving method known for an excellent performance at the short frame by creating an interleaving pattern. CONSTITUTION: A device for interleaving the turbo inside by using a golden sequence includes a memory(316), a golden column table(312), a memory controller(308) and a selector(314). The memory(316) stores the input data when the input enable is applied and outputs the input data when the output enable is applied. The golden column table(312) has a golden sequence generated by the method for interleaving the turbo inside by using the golden sequence. The memory controller(308) controls the memory(316). And, the selector(314) is set as a sequential memory assignment by receiving the control of the memory controller(308) and determines the order of the input and output of the memory(316).

Description

Turbo internal interleaving method using golden sequence and device therefor {TURBO INTERNAL INTERLEAVING METHOD AND DEVICE BASED ON GOLDEN SEQUENCE}

The present invention relates to an internal interleaver of a turbo encoder and a turbo decoder used in a wireless communication system. More particularly, the present invention relates to a golden sequence in a turbo code having a relatively short frame such as a voice data service. It relates to a method and an apparatus for interleaving using.

The turbo code published by Berrou et al. In 1993 is two or more recursive systematic convolutional codes, whose performance is known to approach Shannon's limit. Factors that determine the superior performance of the turbo code include the length and structure of the interleaver, a code rate, an iterative decoding algorithm, and the like.

Hereinafter, a turbo code according to the related art will be described with reference to the accompanying drawings.

1 is a view showing a turbo encoder according to the prior art. First, a data block 0 having a length of k bits is input to the turbo encoder, and Nk tail bits are added to the data block 0 input from the pad 2 to form a frame having N bits. Since a conventional conventional turbo coder is a systematic regression code system, an N bit frame made by the above-described process becomes an output x ₀ (8) after a predetermined delay, and at the same time, the N bit frame passes through M internal interleavers (4). The sequence of bit strings in each frame are mixed, and each of them becomes the outputs X ₁ , X ₂ , ..., X _M (10) having small correlations through the constituent encoder 6, respectively, according to the unit time as a whole. Symbols of x ₀ (8), X ₁ , X ₂ ,..., X _M (10) are generated to form a turbo codeword.

2 is a block diagram of a turbo decoder including a component decoder 1 108 and a component decoder 2 124 therein for decoding a 1/3 turbo code. As shown in the block diagram of FIG. 2, the turbo decoder consists of two or more internal component decoders, and the bit error rate is continuously improved through repeated decoding by exchanging extrinsic information between the component decoders. .

As shown in FIG. 1, an internal interleaver 4 is positioned in a turbo encoder, and the internal interleaver 4 has a low correlation code between parity bits formed by two or more internal component encoders 6. It plays a role in generating heat, and it can be said that the purpose is to suppress the occurrence of low weight codewords that generally cause performance degradation of error correction codes.

The turbo internal interleaver may form various interleaving patterns such as block interleaving, diagonal interleaving, and random interleaving. Random interleaver is basically used to minimize performance degradation of the turbo code due to the double interleaver. Since the structure of the random interleaver causes difficulties in the performance analysis and design of the interleaver, the interleaving method of the specific structure that is expected to perform better is actually selected. In particular, when the frame length of data is relatively short, such as a voice service, the performance of the interleaver in the form of block interleaving is known to be the best. However, a better performance interleaving method should be proposed for application efficiency in a short frame of the turbo code.

In order to solve the above problem, the present invention is known to provide excellent performance in a short frame by generating an interleaving pattern using a golden sequence as one of regular interleaving methods in the case of a short frame length such as a voice service. An object of the present invention is to provide a turbo internal interleaving method and apparatus which is superior to the block interleaving method.

1 is a block diagram showing the structure of a typical turbo encoder;

2 is a block diagram of a turbo decoder for decoding a 1/3 turbo code;

3 is a diagram illustrating an example in which the weight of a codeword does not change even after interleaving in a conventional internal interleaver.

4 is a block diagram showing the structure of a turbo internal interleaver according to the present invention;

5 is a graph showing a performance comparison of the turbo code when applying the turbo internal interleaver according to the present invention.

In order to achieve the above object, the present invention is a rearranged by multiplying the initial index by the golden rule and performing a modulo-1 operation by inputting the initial memory address and index, and rearranging the changed index in ascending order. It provides a turbo internal interleaving method using the golden sequence comprising the step of generating a golden sequence by using the memory address as an interleaving pattern.

Also, a memory for storing input data when an input enable is applied and a outputting of the stored input data when an output enable is applied, a golden column table for storing a golden sequence generated by the method, a memory controller for controlling the memory; Under the control of the memory controller, when the input data is input, the sequential memory is set, and when the input data is output, the selector is set so as to be sequentially stored in the golden column table. It provides a turbo internal interleaving apparatus using a golden sequence comprising a.

The above object and various advantages of the present invention will become more apparent from the embodiments of the present invention described below with reference to the accompanying drawings by those skilled in the art.

First, the technical idea applied to the present invention will be briefly described. In the related art, the purpose of using the turbo internal interleaver is similar to that of a channel interleaver used between a conventional channel and a channel encoder and used to distribute a concatenation error.

3 illustrates an example in which a self termination sequence occurs in an input string having a weight of 4. In general, low weight codewords have a large correlation between codes, which degrades the performance of channel codes. When the codeword is interleaved through the block interleaver as shown in FIG. 3, the bit string 210 and the bit string 212 read in the read direction 200 are positioned in the write direction 202 even after interleaving. Since the bit string 206 and the bit string 208 are formed, the actual self-terminating string does not change its position even after interleaving, so that a low weight codeword before interleaving is generated even after interleaving. The superior turbo interleaver can be said to obtain performance improvement by efficiently interleaving such input sequence to increase the hamming distance of codewords.

On the other hand, the block-type interleaver is known to have better performance than the random interleaver because the frame interleaver has a smaller probability of generating a self-termination string when the frame length is shorter. It does not have the disadvantage. There are active researches on the structural design by grasping the weight distribution of codewords, and the S-Random interleaver design method which reduces the influence of self-termination string by forcibly dropping the double interleaving index by distance S is the structural design of the random interleaver. It is widely known as a method. Similarly, in order to overcome the limitations of the block interleaver, the diagonal interleaver and the odd-even interleaver have been announced. Field) interleaving methods are being studied. Accordingly, an interleaver with good performance can be designed even with a constant interleaving pattern in a short frame. For this purpose, an interleaving method using a golden sequence is used as an interleaving method.

The golden sequence is generated using the golden rule, which divides length 1 into two parts, g and 1-g. The ratio of the long to full length of the two divided lengths is equal to the ratio of the length of the short and long parts of the two lengths. That is, g / 1 = (1-g) / g. In this case, g is about 0.618. To form an interleaving pattern, g is multiplied by the interleaving index g, modulo-1 operation is performed, and the entire index is rearranged.

The following describes the generation procedure of the golden sequence. This procedure is performed by the initial index and the golden rule, and describes, for example, an interleaving pattern generation procedure having a frame length of 16.

First we need to set the initial memory address and index

int1 = (00000 (1), 0000l (2), 00010 (3), 000ll (4), 00100 (5), 00101 (6), 00110 (7), 00111 (8), 01000 (9), 01001 ( 10), 01010 (11), 01011 (12), 01100 (13), 01101 (14), 01111 (15), 10000 (16)}

If you multiply each index by g and then perform a modulo-1 operation

int2 = (00000 (0.6180), 00001 (0.2361), 00010 (0.8541), 00011 (0.4721), 00100 (0.0902), 00101 (0.7082), 00110 (0.3262), 00111 (0.9443), 01000 (0.5623), 01001 ( 0.1803), 01010 (0.7984), 01011 (0.4164), 0ll00 (0.0344), 01101 (0.6525), 01111 (0.2705), 10000 (0.8885)}

Same as Finally, ascending each index

int3 = (01100 (0.0344), 00100 (0.0902), 01001 (0.1803), 00001 (0.2361), 01111 (0.2705), 00l10 (0.3262), 01011 (0.4164), 00011 (0.4721), 01000 (0.5623), 00000 ( 0.6180), 01101 (0.6525), 00101 (0.7082), 01010 (0.7984), 00010 (0.8541), 10000 (0.8885), 00111 (0.9443)}

At this time, the address becomes a new interleaving pattern.

Using the interleaving method, when the frame length is 192 bits, the interleaving pattern is obtained, and S1 and S2 are obtained from the viewpoint of S random interleaving, and S1 = 55 and S2 = 9, which is generally considered in S random interleaver design. , S2 = (N / 2) ^{l / 2} = (192/2) ^{l / 2} = 9.79. Here, S1 represents a distance difference from a neighboring index, and S2 represents that all indexes in S2 have a distance difference of S1 around the observation index.

Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. 4 is an embodiment of an interleaving apparatus using the interleaving method for next generation mobile communication such as 3GPP.

As shown in Fig. 4, the memory 316 address input is set to the sequential memory designation of the counter 310 by the control of the selector 314 of the memory controller 308 at the time of data input, whereby the input data 306 is stored in memory. And are sequentially stored at 316. When the input data 306 block is completed, the memory controller 308 sets the selector 314 to be sequentially assigned to the memory 316 address input in the golden column table 312. Therefore, the input data 306 which was sequentially stored in the memory 316 at the time of input is changed in the order interleaved using the golden sequence in synchronism with the clock 304 when the output enable 318 is applied. Is output to 320.

While the invention has been shown and described in connection with specific embodiments thereof, it is well known in the art that various modifications and changes can be made without departing from the spirit and scope of the invention as indicated by the claims. Anyone who owns it can easily find out.

As described above, when the golden sequence-based turbo internal interleaver according to the present invention is provided as an interleaver of turbo codes in a short frame, a turbo having a better performance than a block interleaver which generally shows excellent performance in a short frame such as a voice service is provided. You can design your code. In addition, it becomes possible to design an interleaver with a performance that cannot be obtained by a random interleaver in a short frame. That is, it is possible to design an interleaver that can replace the self-terminating sequence efficiently, which is superior to the block interleaver, but which can increase the minimum hamming distance of the codeword through an additional interleaving method.

5 is an exemplary diagram illustrating error performance of a block, random, and 3GPP standard prime interleaver according to the present invention and an interleaver using a golden sequence. When the frame size is 192 bits, bit error rates are compared.

In the simulation, an Additive White Gaussian Noise (AWGN) channel is assumed, and an encoder with a code rate of 1/3 and a generation polynomial of the constituent encoder is {13, 15} ₈ . In addition, the trellis of the two encoders was terminated and designed in the same manner as the turbo encoder and decoder of the 3GPP standard. The decoding algorithm of the decoder was a Log-MAP (Maximum A Posteriori) algorithm.

As shown in FIG. 5, it is known that the turbo code performance 404 and 412 of the block interleaver are superior to the turbo code performance 406 and 414 of the random interleaver. The turbocode performance 408 and 416 when the prime interleaver, which is a turbo interleaver of the 3GPP standard, is similar to the turbocode performance 404 and 412 when the block interleaver is applied. In addition, it can be seen that the performance (410, 418) of the turbo code is the best when the interleaver using the golden sequence according to the present invention.

Claims (4)

For application efficiency in a short frame of turbocode, multiplying the initial index by a golden ratio and performing a modulo-1 operation with an initial memory address and an index as input; And And generating a golden sequence using the relocated memory address as an interleaving pattern by aligning the changed indexes. The turbo internal interleaving according to claim 1, wherein the interleaving pattern is generated according to a frame length required based on a frame length output from an internal component decoder of the turbo decoder and an input enable signal. Way. For application efficiency in a short frame of turbo code, when an input enable is applied, the input data is stored in memory by synchronizing with the input enable signal, and when an output enable is applied, the input data is synchronized with the output enable signal. A turbo internal interleaving method using a golden sequence for outputting the stored input data sequentially in a golden sequence from a memory. The golden sequence is generated by multiplying the initial index by an initial memory address and an index, performing a modulo-1 operation, and then aligning the changed index as an interleaving pattern. Turbo internal interleaving method using golden sequence. A memory for storing input data when an input enable is applied and outputting the stored input data when an output enable is applied; A golden column table having a golden sequence generated by the method of claim 1; A memory controller controlling the memory; And Under the control of the memory controller, when the input data is input, the memory is set to a sequential memory designation, and when the input data is output, the memory controller is set to a sequential memory stored in the golden column table to determine the input / output order of the memory. Turbo internal interleaving apparatus using a golden sequence, characterized in that it comprises a selector.