KR20080086316A - Method and apparatus for generating training sequence codes in communication system - Google Patents

Abstract

A method and an apparatus for generating TSC(Training Sequence Codes) in a communication system are provided to apply a TSC to a 16-QAM(Quadrature Amplitude Modulation) and a 32-QAM extendedly employed for a GERAN(GSM(Global System for Mobile Communication)/EDGE(Enhanced Data Rates for GSM Evolution)) RAN(Radio Access Network). A sequence 'A' and a sequence 'B', a pair of sequences having non-periodical autocorrelation characteristics, are generated. The last 'L' number of symbols of the sequence 'A' are duplicated to generate a protection sequence 'A'', and the last 'L' number of symbols of the sequence 'B' are duplicated to generate a protection sequence 'B''. Optimization is performed by evaluating degradation of an SNR(Signal-to-Noise Ratio) of TSC candidates to search TSCs as many as required and store them.

Description

Method and apparatus for generating training sequence codes in communication system {Method and Apparatus for Generating Training Sequence Codes in Communication System}

1 is a diagram illustrating a downlink transmitter structure of a general GERAN system.

2 is a diagram illustrating a receiver structure in a general GERAN system.

3 is a diagram illustrating a normal burst structure used in a conventional GERAN system.

4 illustrates a TSC structure according to an embodiment of the present invention.

5 illustrates a binary TSC set generated from a Golay complementary sequence according to the first embodiment of the present invention.

6 illustrates a binary TSC set generated from a periodic complementary sequence according to a second embodiment of the present invention.

7 is a flowchart illustrating a training sequence code generation procedure according to the first and second embodiments of the present invention.

8 illustrates a TSC structure according to an embodiment of the present invention.

9 illustrates a binary TSC set having a length of 31 symbols according to the first embodiment of the present invention.

10 illustrates a binary TSC set having a length of 31 symbols according to the second embodiment of the present invention.

The present invention relates to a method and apparatus for generating a training sequence code in a communication system, and more particularly, to a Global System for Mobile Communication (GSM) / Enhanced Data Rates for GSM Evolution (Evolution) Evolution Radio Access Network (RAN) (hereinafter, GERAN). A method and apparatus for generating a training sequence code in a system.

Current 3GPP (3 ^rd Generation Partnership Project) in TSG-GERAN meeting standard pushing the GERAN evolution to improve performance, such as data rate (Data Transmission Rate) and the spectral efficiency (Spectral Efficiency). In order to improve downlink and uplink performance, 16-QAM and 32-QAM, which are higher order Quadrature Amplitude Modulation (QAM) modulation schemes, are applied to conventional modulation schemes such as Gaussian Minimum Shift Keying (GMSK) and 8-Phase Shift Keying (PSK). Is added.

In addition, to increase the data rate and spectral efficiency, the symbol rate is improved from the existing rate of 270.833 ksymbols / s to the new rate of 325 symbols / s. The new symbol rate, which is 1.2 times higher than the existing symbol rate, will be applied to both uplink and downlink and will be reflected in the GERAN standard in the second half of 2007.

As described above, in the conventional GERAN system, there are GMSK and 8-PSK modulation. In the GMSK method, the binary data is passed through a Gaussian low pass filter to limit the bandwidth, and the spectrum is continuously changed between two frequencies by frequency modulation at a constant ratio. Excellent concentration and high out-of-band spectrum suppression. The 8-PSK scheme may increase frequency efficiency by modulating data to correspond to a phase shifted code of a carrier. As the encoding scheme used in the EDGE / EGPRS system, nine techniques for packet data traffic channels (PDTCH) are defined. The nine techniques are nine schemes from Modulation and Coding Schemes (MCSs) MCS-1 to MCS-9 for EDGE / EGPRS. In actual communication, one of various combinations of the modulation schemes and the encoding schemes is selected and used. MCS-1 to MCS-4 use GMSK modulation and MCS-5 to MCS-9 use 8-PSK modulation. The MCS scheme used for transmission depends on the quality of the measured channel. The structure of the transmitter / receiver of the GERAN system will be described with reference to the drawings.

1 illustrates a downlink transmitter structure of a general GERAN system.

Referring to FIG. 1, a radio link control (RLC) packet data block (RLC Block) is sent to a channel encoder (110), encoded by a convolutional code, and defined by a convolutional code. It is punctured according to the processing pattern and then sent to the interleaver 120. Data interleaved in the interleaver 120 is sent to a multiplexer 140 to allocate data to a physical channel. In addition, the RLC / MAC header information, the Uplink State Flag (hereinafter referred to as USF), and the Code Identifier bit 130 are also sent to the multiplexer 140. The multiplexer 140 distributes the collected data into four normal bursts and allocates each burst to a time slot of a TDMA frame. The data of each burst is modulated through a modulator 150, and a training sequence code (TSC) is added by the training sequence rotation unit 160, and a phase rotation is performed with respect to the TSC. Is then sent to the transmitter 170. Here, a detailed description of an apparatus, for example, a digital-to-analog converter, etc., which is additionally required for transmitting a modulated signal is omitted.

2 illustrates a receiver structure of a general GERAN system.

Referring to FIG. 2, the transmitted bursts are received in a time slot unit at a radio front end terminal 210 through a receiving antenna. The received data is buffered and reverse rotated in the buffering and reverse rotation unit 260. The modulation scheme detection and channel estimator 270 estimates the modulation scheme detection and channel information using these data. The equalization and demodulation are performed in the equalizer block 230 based on the detected modulation scheme and the estimated channel information, and then transmitted to the deinterleaver 240. The data is transmitted to the channel decoder 250 after the deinterleaving is performed, and the data transmitted through the channel decoder 250 is restored.

3 illustrates a normal burst structure used in a conventional GERAN system.

As shown in FIG. 3, in transmitting data in a conventional GERAN system, a training sequence code (hereinafter referred to as TSC) composed of 26 symbols is located at the center of a normal burst structure. do. Eight types of TSCs are defined in the standard and used in actual GSM networks and terminals, and one identical TSC is allocated in one cell. TSC is used in an equalizer for estimating radio channel state information at a receiver to remove noise and interference included in a transmission signal. In addition, link quality control (hereinafter referred to as LQC) may be determined by measuring and reporting channel quality or link quality of a receiver from the TSC.

The conventional TSC is composed of codes having excellent autocorrelation properties. Therefore, the conventional TSC has a good characteristic when performing channel estimation on one channel without considering the interchannel interference. However, when designing a cell structure in a cellular system, carrier frequencies are reused at a sufficient distance in consideration of co-channel interference (hereinafter referred to as CCI). However, as the reuse frequency of the carrier frequency increases, the CCI increases, and the increase of the CCI has a significant effect on the channel estimation and signal detection performance. Therefore, in a cellular system such as GSM, it is desirable to estimate an accurate channel using a joint channel estimation method when severe CCI exists. In this case, cross-correlation properties between TSCs have a great influence on the performance of the joint channel estimation method. However, currently used TSCs of GERAN adopt a design method that does not consider any cross-correlation characteristics, which not only degrades the performance of the system in the CCI environment but also high-order modulation such as 16-QAM and 32-QAM used in the GERAN evolution system. Extending the conventional TSC to the scheme may result in system degradation.

The technical problem to be achieved by the present invention is to propose a method for generating a new TSC and a TSC arrangement in order to solve the shortcomings of the TSC commonly used in the conventional GERAN system.

Another object of the present invention is to provide a transmission and reception apparatus for efficient data transmission and reception in a GERAN system using a TSC according to the present invention.

Another object of the present invention is to provide a method of extending and applying a newly generated TSC to 16-QAM and 32-QAM employed in a GERAN system.

According to an embodiment of the present invention, there is provided a method of generating a training sequence code in a communication system, the method comprising: generating a sequence A and a sequence B, which are sequence pairs having aperiodic autocorrelation, and generating the last L symbols of the sequence A; A protection sequence A 'is generated by copying, and the last L symbols of the sequence B are copied to generate a protection sequence B', and the optimization is performed through signal-to-noise ratio degradation evaluation of the training sequence code candidates. It includes the process of retrieving and storing the training sequence code.

In addition, according to an embodiment of the present invention, in a method for generating a training sequence code in a communication system, sequence A'A, which is a sequence pair having aperiodic autocorrelation characteristics, and sequence A'A, which has a periodic autocorrelation characteristic from sequence B, and Optimizing by generating a sequence B'B, searching all sequences for the sequence A'A and sequence B'B to construct training sequence code candidates, and evaluating signal-to-noise ratio degradation of the training sequence code candidates Performing search to retrieve and store the required number of training sequence codes, wherein the sequence A 'is generated by copying the last L symbols of the sequence A, and the sequence B' is the last L of the sequence B. Create two copies of the symbol.

Hereinafter, with reference to the accompanying drawings will be described in detail the operating principle of the preferred embodiment of the present invention. In the following description of the present invention, detailed descriptions of well-known functions or configurations will be omitted if it is determined that the detailed description of the present invention may unnecessarily obscure the subject matter of the present invention. Terms to be described later are terms defined in consideration of functions in the present invention, and may be changed according to intentions or customs of users or operators. Therefore, the definition should be made based on the contents throughout the specification.

In the present invention, in designing a GERAN system and a TSC to be applied to the GERAN system, both autocorrelation and cross-correlation characteristics are considered, and Golay complementary sequences are used to find a suitable TSC. In addition, to evaluate the interference characteristics between sequences, signal-to-noise ratio (hereinafter referred to as SNR) degradation is introduced as a criterion. In addition, the Mini-Max optimization method is introduced to find TSCs that have excellent cross-correlation characteristics of binary TSCs. For the minimum-maximum optimization method, refer to the contents described in Korean Patent Application No. 10-2007-0012983.

First, the TSC configuration according to the embodiment of the present invention will be described.

The present invention designs the TSC using Golay complementary sequences and quasi-complementary sequences. The TSC proposed in the present invention is generated to satisfy the TSC structure shown in FIG. 4 by using a Golay or a sequence having similar complementary sequences A and B in pairs.

In FIG. 4, the guard sequences A 'and B' having a length of z are placed at the most significant positions of A and B, respectively. The protection sequences A 'and B' should be set as short as possible while at the same time an inter-symbol interference between data symbols in the TSC within the same Timeslot or in the sequences A and B It must be long enough to eliminate ISI. That is, the number of symbols in one protection sequence must satisfy the condition z? L-1.

, N'=2(N+L-1)일 때 CCI를 고려하지 아니한 경우, 수신기에서의 수신 신호 샘플은 다음 수학식 5로 표현할 수 있다. A sequence X having a length N 'is defined as a TSC designed using a complementary sequence pair A and B of length N having a structure as shown in FIG. In other words,

When the CCI is not considered when N '= 2 (N + L-1), the received signal sample at the receiver may be expressed by Equation 5 below.

채널 메모리로 인하여 한 개의 버스트 내에서 데이터 심볼과 TSC 사이의 간섭을 제거하기 위한 z(이하에서 z = L-1을 사용함) 개의 길이를 갖는 보호심볼이다. 벡터 형태로 표시하면

와 같다. 여기에서 잡음벡터는

이고, X는

차원을 갖는 행렬로서 다음 수학식 2와 같다. Note that

This is a guard symbol having a length of z (hereinafter, z = L-1) for removing interference between a data symbol and a TSC in one burst due to channel memory. In vector form

Same as Where the noise vector is

And X is

A matrix having dimensions, as shown in Equation 2 below.

Meanwhile, the channel estimate using a well-known Least-Squares Error Estimate is as shown in Equation 3 below.

In Equation 3, X 'is a conjugate transpose matrix of X.

In GSM / EDGE, the channel tap length of the channel model is 6. Therefore, when the case of L = 6 will be described with an example, the received signal sample in the receiver can be expressed by the following equation (4).

In consideration of Equation 4, the TSC matrix is defined as Equation 5 below.

에서 y와 n 은 각각 (2N+1) 크기 를 갖는 벡터이다. 따라서 수학식 5에 정의된 X를 사용하여 수학식 3으로부터 최소제곱오차추정을 계산할 수 있고, X^tX는 (L × L)크기를 갖는 비주기적 자기상관 행렬이된다. 시퀀스 A와 B는 Golay 보완 시퀀스 또는 유사 보완 시퀀스이기 때문에,φ=X^tX는 대각성분 값이 2N인 대각행렬이 된다. Considering Equation 5,

Where y and n are vectors of size (2N + 1), respectively. Therefore, the least square error estimate can be calculated from Equation 3 using X defined in Equation 5, and X ^t X becomes an aperiodic autocorrelation matrix having a size of (L × L). Since sequences A and B are Golay complementary sequences or pseudo complementary sequences, φ = X ^t X is a diagonal matrix with a diagonal component value of 2N.

이라 정의하면, 두 동일 채널(Co-Channel) 신호에 대한 채널 임펄스 응답은

로 정의할 수 있다. 이 채널 임펄스 응답과 수학식 5를 이용하여 TSC 행렬을 다시 정의하면

과 같다. 따라서 CCI를 고려한 수신 신호는

가 되고, 결과적으로 최소제곱 채널 추정치는 수학식 6과 같이 계산된다. Assuming that there is one interference for each cell in the cellular communication system, the mutual cross-correlation characteristic between TSCs should be optimized for Joint Channel Estimation. The channel impulse response of the carrier and edge signals

In this case, the channel impulse response for two co-channel signals is

Can be defined as By redefining the TSC matrix using this channel impulse response and

Is the same as Therefore, the received signal considering CCI

As a result, the least-squares channel estimate is calculated as shown in Equation 6.

One criterion for evaluating the mean square error for a received signal may be SNR degradation (SNR Degradation in dB, hereinafter referred to as SNR-d). SNR-d is used to evaluate the cross-correlation property of TSCs, and is defined as in Equation 7 below.

In Equation 7, tr (φ ⁻¹ ) is the sum of the main diagonal components of the matrix φ ⁻¹ . The smaller the SNR-d value is, the better it is to evaluate the cross-correlation property of TSC.

Next, a method of finding a complementary sequence pair for A and B will be described with reference to Korean Patent Application No. 10-2007-0012983.

First, the Golay complementary sequence will be described. A well-known Golay complementary sequence can be found through computer search for a whole set of sequences having a relatively small length. That is, an Even-Shift Orthogonal Sequence with 2N lengths is uniquely determined from Golay complementary sequence pairs A and B with N lengths.

The present invention relates to a new TSC having two level signals. When generating TSCs, instead of considering aperiodic or non-periodic autocorrelation characteristics and cross-correlation characteristics, periodic correlation characteristics may be considered.

As shown in FIG. 4, the newly proposed TSC structure is generated based on a periodic complementary sequence. A method for finding an optimized TSC will be described through the following specific examples. The following two embodiments consist of eight different TSCs with 26 symbol lengths that can be used in the GERAN system. In FIG. 4, two sequences A 'and B' may be generated by copying the last L symbols of the two sequences A and B, respectively. That is, TSC is expressed as in Equation (8).

The present invention describes two embodiments of generating sequence pairs (A'A) and (B'B) in the TSC defined in Equation (8).

<First Embodiment>

First step: Search all sequences for two sequences A and B that satisfy the equation 9 and have aperiodic autocorrelation.

Considering TSC having 26 symbol lengths in Equation 9, K = 5 and N = 8.

Second step: TSC candidates having the structure shown in FIG. 4 are configured. For reference, if the two sequences A and B are Golay complementary sequences, the sequences A and B are also complementary to cyclic autocorrelation. Therefore, there is a characteristic of Equation 10 for periodic autocorrelation.

Third step: The TSC candidates generated in the second step are optimized using the minimum-maximum optimization method described in Korean Patent Application No. 10-2007-0012938 and find the required number of TSCs.

Step 4: Save the TSC found in step 3.

5 shows an example of a TSC set found according to the first embodiment of the present invention. This TSC set is an example suitable for the normal burst of GSM / EDGE.

Second Embodiment

First step: Search all sequences for two sequences A'A and B'B having periodic autocorrelation characteristics while satisfying Equation (10).

Second step: TSC candidates having the structure shown in FIG. 4 are configured.

Third step: The TSC candidates generated in the second step are optimized for SNR degradation using the minimum-maximum optimization method described in Korean Patent Application No. 10-2007-0012938 and the required number of TSCs are found.

Step 4: Save the TSC found in step 3.

6 shows an example of a TSC set found according to the second embodiment of the present invention. This TSC set is an example suitable for the normal burst of GSM / EDGE.

7 illustrates a training sequence code generation procedure according to an embodiment of the present invention.

Referring to FIG. 7, a sequence is searched for in step 701. In the first embodiment of the present invention, A and B are searched, and in the second embodiment of the present invention, A'A and B'B are searched.

In step 702, a TSC candidate as shown in FIG. 4 is generated.

In step 703, the optimization is performed through SNR degradation evaluation of the generated TSC candidates. At this time, as the optimization method, the minimum-maximum optimization method described in Korean Patent Application No. 10-2007-0012938 is used.

Finally, in step 704, the optimized TSC set is stored.

In the following description, an embodiment of generating a TSC having 30 symbol lengths in addition to the above-described first and second embodiments will be described as a third embodiment, and an embodiment of generating a TSC having 31 symbol lengths will be described as a fourth embodiment. Let's explain.

Third Embodiment

The TSC structure generated from the complementary sequence according to the third embodiment of the present invention is the same as FIG. 4 described above.

The TSC length is determined to be N '= 2N + 10. If N = 10, the TSC symbol length is 30. If X is a TSC sequence having a length of 30 symbols, it is represented by Equation 11 below.

A binary TSC having a length of 30 symbols generated as described above is the same as a result of omitting the first bit from each sequence TSC # of FIGS. 9 and 10.

Fourth Embodiment

Suppose sequence A '' is a copy of the last six symbols into sequence A. A TSC having a 31 symbol length is shown in FIG. 8A. It can be seen that A '' is a one-symbol long sequence when compared to the sequence B '. It is also possible to generate another form of TSC as shown in FIG. 8B. 8A and 8B have the same correlation characteristics. Therefore, considering only FIG. 8A, a TSC having a length of 31 symbols may be defined as in Equation 12 below.

이라 하면, 동일 채널 간섭(CCI)이 없는 상황에서 수신기에서의 수신 신호 샘플은

과 같다. Channel impulse response

In this case, in the absence of co-channel interference (CCI), the received signal sample at the receiver

Is the same as

이기 때문에,

의 관계가 성립한다. 그러므로, i=15에서의 수신 신호 샘플

는

에 의해서 구할 수 있다. 시퀀스 A''내의 부가적인 1개 심볼이 설계한 TSC의 자기 상관 특성을 저하 시킬 수 있기 때문에 이 심볼은 행렬 X에 포함되지 말아야 한다. i=6,...,14 및 i=21,...,30에 대하여

라 정의하면, 수신 신호의 수정된 수식 표현은 벡터 형태로

로 쓸 수 있다. 여기에서 n는 잡음벡터로 (20x1) 차원을 갖는 행렬로 다음 수학식 13과 같이 표현할 수 있다.For reference

Because

The relationship is established. Therefore, the received signal sample at i = 15

Is

Can be obtained by This symbol should not be included in the matrix X because an additional symbol in sequence A '' can degrade the autocorrelation characteristics of the designed TSC. About i = 6, ..., 14 and i = 21, ..., 30

In this case, the modified mathematical expression of the received signal is in vector form.

Can be written as Here, n is a noise vector as a matrix having a (20 × 1) dimension, as shown in Equation 13 below.

Therefore, Least-squares error estimates (LSEE) for the channel is expressed by Equation 14 below.

는 X의 켤레 변환(conjugate Transpose)이다. 두 개의 신호를 갖는 경우에 대하여도 행렬

는 심볼 길이가 30인 TSC인 경우와 비교하여 변하지 않는다. 그러므로, 심볼 길이 30에 대한 TSC의 SNR 저하(degradation) 계산에서 얻은 결과를 상기에서 생성한 31 심볼 길이의 TSC에도 적용할 수 있다. In Equation 14

Is the conjugate transpose of X. Matrix even for two signals

Does not change compared to the case of TSC having a symbol length of 30. Therefore, the result obtained from the calculation of the SNR degradation of the TSC for the symbol length 30 can also be applied to the TSC of the 31 symbol length generated above.

Binary TSCs having the above description and the 31 symbol length generated by FIG. 8A are the same as those of FIGS. 9 and 10. In addition, a binary TSC having a length of 31 symbols based on FIG. 8B can be easily generated according to the above description.

Meanwhile, in the detailed description of the present invention, specific embodiments have been described, but various modifications are possible without departing from the scope of the present invention. Therefore, the scope of the present invention should not be limited to the described embodiments, but should be defined not only by the scope of the following claims, but also by those equivalent to the scope of the claims.

In the present invention operating as described in detail above, the effects obtained by the representative of the disclosed invention are briefly described as follows.

According to the present invention provides a TSC in consideration of the cross-correlation characteristics, the use of the TSC according to the present invention enables efficient data transmission and reception in the GERAN system without performance degradation. In addition, the TSC according to the present invention can be extended to 16-QAM and 32-QAM employed in the GERAN system.

Claims (2)

In the method for generating a training sequence code in a communication system, Generating sequence A and sequence B, which are sequence pairs having aperiodic autocorrelation, Copying the last L symbols of the sequence A to generate a protection sequence A ', copying the last L symbols of the sequence B to generate a protection sequence B', And optimizing the signal-to-noise ratio degradation of the training sequence code candidates to retrieve and store the required number of training sequence codes. In the method for generating a training sequence code in a communication system, Generating a sequence pair A'A and a sequence B'B having sequence autocorrelation characteristics from sequence A and sequence B having aperiodic autocorrelation characteristics, and Searching for all sequences for the sequence A'A and sequence B'B to construct training sequence code candidates; Performing the optimization through the signal-to-noise ratio deterioration evaluation of the training sequence code candidates to retrieve and store as many training sequence codes as necessary; Wherein the sequence A 'is generated by copying the last L symbols of the sequence A, and the sequence B' is generated by copying the last L symbols of the sequence B.

Images