KR100747852B1 - Maximum-likelihood symbol timing and carrier frequency offset estimation method and apparatus for OFDM systems based on a preamble with two identical part - Google Patents

Abstract

Method and apparatus for estimation of maximum likelihood symbol timing and carrier frequency offset using one preamble having a repetitive characteristic and a cyclic prefix (CP) for a preamble in an Orthogonal Frequency Division Multiplexing (OFDM) system Is disclosed. According to the present invention, when a sequence is loaded on even-numbered subcarriers and an Inverse Fast Fourier Transform (IFFT) is obtained, an OFDM symbol having a structure that is repeated once in the time axis is obtained. A terminal station samples a downlink signal including an OFDM symbol having a structure that is repeated once as a preamble and defines a timing metric based on an autocorrelation calculation between samples corresponding to the preamble. Here, the autocorrelation calculation between samples includes not only autocorrelation between samples in the effective OFDM symbol but also autocorrelation calculation between samples corresponding to the cyclic prefix and the effective OFDM symbol. The position of the specimen that maximizes this timing metric is considered the start position of the preamble. The carrier frequency error can also be estimated by calculating autocorrelation from a certain number of samples starting from the position of the sample maximizing the timing metric and measuring the phase of the result.

Description

Maximum likelihood symbol timing and carrier frequency offset estimation method and apparatus for OFDM systems based on a single preamble having a repetitive characteristic and a cyclic prefix for a preamble in an OFM system preamble with two identical part}

1 is a conceptual diagram for conceptually explaining a method and apparatus for estimating maximum likelihood symbol timing and carrier frequency offset according to the present invention;

2 is a block diagram of a preamble applied to a method and apparatus for estimating a maximum likelihood symbol timing and a carrier frequency offset according to the present invention.

3 is a block diagram of a synchronization acquisition device applied to the present invention.

4 is a detailed configuration diagram of the timing metric calculator of FIG.

5 is a detailed configuration diagram of the first autocorrelator of FIG. 4.

FIG. 6 is a detailed configuration diagram of the second autocorrelator of FIG. 4. FIG.

7 is a detailed block diagram of the energy calculator of FIG.

8 is a correlator that calculates the correlation of b samples with a distance between two matched filter output samples a .

9 is a simplified detailed schematic diagram of the first autocorrelator of FIG.

10 is a simplified detailed schematic diagram of the energy calculator of FIG.

11 is a simplified detailed schematic diagram of the energy calculator of FIG.

12 is a graph of the comparison between the variance of the symbol timing estimate estimated according to the present invention and the variance of the symbol timing estimate estimated according to the prior art.

13 is a graph of comparison between mean square error (MSE) of frequency offset estimate estimated according to the present invention and MSE of frequency offset estimate estimated according to the prior art;

<Description of the symbols for the main parts of the drawings>

1 ... Sample matched filter 10 ... Timing metric calculator

20 ... Sample position selector 30 ... Sample counter

50 Phase Calculator 112 First Autocorrelator

114 ... First absolute value calculator 120 ... Second autocorrelator

122 ... 2nd absolute calculator 124 ... Square

132 Energy Calculator 136 Reciprocal Calculator

The present invention relates to a method and apparatus for estimating a maximum likelihood symbol timing and a carrier frequency offset using a cyclic prefix for a preamble and a preamble having a repetitive characteristic in an orthogonal frequency division multiplexing (OFDM) system. Maximum likelihood symbol timing and carrier using cyclic prefix for one preamble and preamble having repetitive characteristics in an OFDM system in which a UE acquires symbol timing synchronization and carrier frequency error synchronization for a downlink signal including a preamble having a structure A method and apparatus for frequency offset estimation.

As is well known to those skilled in the art, orthogonal frequency division multiplexing (OFDM) based communication systems are in the spotlight as the modulation and demodulation technology of the next generation broadband mobile communication systems because they enable robust and efficient frequency use for multipath channels. In particular, the use of a Cyclic Prefix (CP) eliminates Inter-Symbol Interference (ISI) and at the same time, the allowable error is relatively high when acquiring symbol timing synchronization.

Methods for acquiring symbol timing synchronization in an OFDM based wireless communication system are classified into two categories. The first method is to use autocorrelation between CP and effective OFDM symbol. This method requires a long time for acquiring synchronization because the relatively short CP length requires the accumulation of autocorrelation values for several symbols. The second method is to utilize a preamble having a specific structure in a system for transmitting and receiving data in packet units. Compared to the method using CP, faster symbol timing synchronization can be obtained, but bandwidth efficiency is inferior because a preamble for acquiring synchronization is included every frame.

In the conventional method of acquiring symbol timing synchronization using a preamble, a timing metric defined based on a structure of a preamble, which is repeated once, is calculated, and the position of a sample that maximizes the result is set as the start position of the preamble. To consider. Here, the preamble is obtained by taking an IFFT with a specific sequence on even-numbered subcarriers, resulting in a structure that is repeated once in the time axis, and there is a correlation between samples having a distance of N / 2 between two samples within the preamble. exist. Where N is defined as the Fast Fourier Transform (FFT) magnitude. The existing timing metric utilizes only the correlation between repetition patterns existing in the effective OFDM symbol of the preamble, and does not consider the correlation between the CP and the effective OFDM symbol. The result is a flat area in the timing metric as a function of the sample's position and ambiguity in the symbol timing estimates. As long as the estimated symbol timing is present in the CP, it is possible to recover the data without ISI. However, since the larger the variance of the symbol timing estimate may cause the performance degradation of other algorithm blocks, an estimation algorithm having a smaller variance of the symbol timing estimate is required.

In order to solve the above problem, there is a small symbol timing estimation method proposed in the related art, in which a result of performing an FFT on a real number sequence is conjugated to a Hermitian symmetry around the N / 2th sample. This method is based on calculating the sum of the products of samples separated by a certain distance from the N / 2th sample. This method has a disadvantage in that the computational complexity is very large because it cannot be calculated by the recursive method.

Another way to reduce the variance of symbol timing estimates is to increase the number of repeating patterns and assign different symbols to each repeating pattern to ensure that the timing metric is maximized at the sample position corresponding to the start of the preamble. . In this method, when the FFT size N is fixed, in order to make the number of repeating patterns L in the time axis, each alphabetic sequence of L subcarriers must be loaded in the frequency domain so that the length of the sequence becomes N / L. Therefore, as L increases, N / L decreases. Since the length of the sequence is related to the number of base stations that can be supported in the communication system using the corresponding preamble, the length of the sequence constituting the preamble is preferably longer. As a result, the above method requires an algorithm for reducing the variance of the symbol timing estimate without increasing the number of repeating patterns.

Accordingly, the technical problem to be achieved by the present invention is a symbol timing estimation method and apparatus and carrier frequency error estimation method in which the variance of the symbol timing estimate and the repetitive calculation of the timing metric are not possible and the calculation complexity of the timing metric itself is not high. The present invention provides a method and apparatus for estimating maximum likelihood symbol timing and carrier frequency offset using one preamble having a repetition characteristic and a cyclic prefix for a preamble in an orthogonal frequency division multiplexing system capable of providing an apparatus.

Another technical problem to be achieved by the present invention is to maintain a length of a sequence for configuring a preamble as N / 2 , one preamble having a repetition characteristic in an orthogonal frequency division multiplexing system that does not reduce the number of sequences for supporting base station identification. And a maximum likelihood symbol timing and carrier frequency offset estimation method and apparatus using a cyclic prefix for a preamble.

In order to achieve the above object, the present invention provides a method for estimating maximum likelihood symbol timing and carrier frequency offset using a cyclic prefix for one preamble and a preamble having a repetitive characteristic in an orthogonal frequency division multiplexing (OFDM) system, a The timing metric using the correlation between the repetitive patterns present in the effective orthogonal frequency division multiplexing (OFDM) symbol interval of the preamble and the correlation between the cyclic prefix (CP) and the effective OFDM symbol is obtained from the sampled matched filter output. Calculating for sample location; b) selecting a sample position that maximizes the calculated timing metric; c) calculating autocorrelation from a sampled matched filter output to estimate a carrier frequency offset at the selected sample position; And d) calculating a phase of the calculated autocorrelation output through an autocorrelator. A symbol timing and carrier frequency offset estimation method is provided.

In a preferred embodiment of the method, the method further comprises: calculating weighted autocorrelation between samples in which the distance between two samples is N / 2 (where N is a fast Fourier transform size) in the timing metric calculation; Calculating weighted autocorrelation between samples having a distance N between the two samples; Calculating the sum by taking the absolute value or the square of the absolute value of each calculated autocorrelation value and performing squares; Calculating an energy of a sampled matched filter output and calculating the inverse of the squared value; And multiplying the inverse of the autocorrelation result and the energy calculation result.

On the other hand, the present invention, in order to achieve the above technical problem, in the apparatus for estimating the symbol timing and the carrier frequency offset, the correlation between the repetition pattern present in the effective orthogonal frequency division multiplexing (OFDM) symbol interval of the preamble and Means for calculating a timing metric using a correlation between a Cyclic Prefix (CP) and a valid OFDM symbol for all sample positions from the sampled matched filter output; Means for selecting a sample position that maximizes the calculated timing metric; Means for calculating autocorrelation from a sampled matched filter output to estimate a carrier frequency offset at the selected sample position; And means for calculating the phase of the calculated autocorrelation output through an autocorrelator.

Hereinafter, preferred embodiments of a symbol timing and carrier frequency offset estimation method according to the present invention will be described in detail with reference to the accompanying drawings. In the following description of the present invention, when it is determined that detailed descriptions of related well-known technologies or configurations may unnecessarily obscure the subject matter of the present invention, the detailed description will be omitted. In addition, terms to be described below are terms defined in consideration of functions in the present invention, which may vary according to the intention or custom of a user or an operator. Therefore, the definition should be made based on the contents throughout the specification.

First, the symbol timing and carrier frequency offset estimation method and apparatus according to the present invention are based on the maximum likelihood estimation algorithm. The system model considered for derivation of the maximum likelihood estimation algorithm according to the present invention is shown in FIG. 1.

이다. 그리고

,

,

는 각각 CP의 길이, 상향 링크에서 하향 링크로의 천이 시간에 해당하는 표본의 개수, 표본화된 정합 여파기(1; 도 3)의 출력이 저장될 수신 벡터의 길이이다. 도 1과 같은 시스템 모델 하에서 심벌 타이밍

를 추정하기 위해 최대 우도 추정 알고리즘을 유도하였을 때 도달하게 되는 결론은 도 2를 참조하여 설명될 수 있다. Referring to FIG. 1, a variable corresponding to a symbol timing error to be estimated is

to be. And

,

,

Are the length of the CP, the number of samples corresponding to the transition time from the uplink to the downlink, and the length of the reception vector in which the output of the sampled matched filter 1 (FIG. 3) is to be stored. Symbol timing under the system model as shown in FIG.

The conclusion reached when the maximum likelihood estimation algorithm is derived in order to estimate can be described with reference to FIG. 2.

Referring to FIG. 2, symbol timing with small variance may be estimated from a timing metric defined by including autocorrelation between CP and effective OFDM symbols in addition to autocorrelation between repetitive patterns in the effective OFDM symbol of the preamble. The timing metric obtained by approximating the maximum likelihood algorithm derivation to a high signal-to-noise ratio is shown in Equation 1 below.

[Equation 1]

와

는 [수학식 2]와 같이 정의된다. In Equation 1

Wow

Is defined as shown in [Equation 2].

[Equation 2]

The symbol timing and carrier frequency error estimation using the timing metric of Equation 1 can be obtained based on Equation 3.

[Equation 3]

과

에 곱해져 있는 서로 다른 가중치들 간의 차이가 없도록 만들면, [수학식 4]와 같이 변형된다. In the timing metric of Equation 1

and

If there is no difference between the different weights multiplied by, Equation 4 is modified.

[Equation 4]

Using the timing metric of Equation 4, symbol timing and carrier frequency error estimation can be obtained based on Equation 5.

[Equation 5]

에 대한 식이 남아있는데 현실적으로 정확한

의 길이를 알 수는 없으므로 [수학식 4]에서

에 의존하는 식을 제외했을 때 결과 식은 [수학식 6]과 같다.The timing metrics of Equations 1 and 4 are

The expression for remains realistic

Since we don't know the length of Equation 4,

Excluding equations that depend on, the resulting expression is shown in [Equation 6].

[Equation 6]

Similarly, symbol timing and carrier frequency error estimation can be obtained based on Equation 7 using the timing metric of Equation 6.

[Equation 7]

To calculate Equations 1, 4, and 6 from the sampled matched filter (1; FIG. 3) output, Equation 2 must be calculated. It can be calculated by the recursive method as shown in Equation 8]. Therefore, the complexity of the timing metric calculation can be greatly reduced.

[Equation 8]

Hereinafter, the operation of the symbol timing and carrier frequency offset estimation method and apparatus according to the preferred embodiment of the present invention, and more specifically, the synchronization acquisition process according to the preferred embodiment of the present invention will be described.

As described above, the representative method for synchronous acquisition of the prior art has a disadvantage in that the variance is large due to the ambiguity of the timing estimate. One of the proposed methods is then to increase the number of repetitions in order to reduce variance and to assign different signs of each repetition pattern to ensure that the newly defined timing metric is maximized at the start of a valid OFDM symbol. Another way is to use Hermitian symmetry, which is the result of taking an IFFT on a series of real numbers. These methods have the disadvantage of reducing the number of sequences for identifying the required base station when generating the preamble or increasing the complexity of the timing metric calculation process. However, the method according to the present invention can reduce the complexity because the dispersion of the timing estimate is small and the number of sequences for base station identification is not reduced and the recursive calculation of the timing metric is possible.

3 is a block diagram of an estimation apparatus for implementing a symbol timing and carrier frequency offset estimation method according to an embodiment of the present invention.

3, a symbol timing estimator according to the present invention comprises a timing metric calculator 10 for calculating a timing metric from an output of a sampled matched filter 1; A matched filter 1, a sample counter 30 for identifying the position of the output sample; And a sample position selector 20 that selects the sample position at which the timing metric is maximum.

를 곱하는 곱셈기(55)를 포함하여 이루어진다. 여기서 제1 자기상관기(112)는 타이밍 메트릭을 최대로 하는 표본의 위치를 시작으로 선택한

개의 표본들로부터 계산한다.The carrier frequency offset estimator includes: a first autocorrelator 112 constituting a timing metric in FIG. 3; A phase calculator 50 for calculating a phase of the output of the first autocorrelator 112;

It comprises a multiplier 55 to multiply. Here, the first autocorrelator 112 selects the starting position of the sample maximizing the timing metric.

Calculate from two samples.

4 shows the structure of the timing metric calculator 10 used in the symbol timing estimating apparatus according to the present invention. The timing metric calculator 10 includes an autocorrelation calculator 110 and an energy calculator 130. The autocorrelation calculation unit 110 includes a first autocorrelator 112 for calculating autocorrelation of samples having a distance of N / 2 between two samples in time; A second autocorrelator 120 for calculating autocorrelation of samples having a distance N between two samples; First and second absolute value calculators 114 and 122 for calculating the absolute values of the outputs of the first autocorrelator 112 and the second autocorrelator 120, respectively; And a squarer 124 that calculates the square of the sum of the absolute values of the outputs of the first autocorrelator 112 and the second autocorrelator 120. The energy calculator 130 includes an energy calculator 132 that calculates a sum by calculating and weighting a square of an absolute value of each sample; A squarer 134 that squares the output of the energy calculator 132; And a reciprocal calculator 136 that calculates the reciprocal of the output of the squarer 134.

또는

개의 켤레 곱들의 합을 계산하는 상관기들(1110a, 1110b, 1110c)과, 이렇게 계산한 각

,

,

에 가중치를 부여하고 합을 계산하는 부분으로 구성된다. 5 shows the configuration of the first autocorrelator 112 of the timing metric calculator 10. The first autocorrelator 112 calculates the conjugate product between the samples having a distance of N / 2 between two samples, or

or

Correlators 1110a, 1110b, and 1110c that calculate the sum of the two conjugate products, and

,

,

It consists of a part that weights and computes the sum.

개의 켤레 곱들의 합을 계산하는 상관기(1110d)와, 계산결과인

에 가중치를 곱하는 부분으로 구성된다. 6 shows the configuration of the second autocorrelator 120 of the timing metric calculator 10. The second autocorrelator 120 calculates the conjugate product between samples with a time difference of N,

A correlator 1110d for calculating the sum of the conjugates of

It consists of a part of multiplying the weight by.

,

,

,

,

,

에 가중치를 곱하여 모두 합하는 부분 및 다수의 지연기(1108e, ..., 1108j)를 포함하여 구성된다. FIG. 7 shows the configuration of an energy calculator 132 that calculates the sum of the squares of the absolute values of several sampled matched filter outputs in FIG. 4 showing the timing metric calculator 10. The energy calculator 132 of FIG. 7 includes correlators 1110e, ..., 1110j for calculating the sum of squares of absolute values of samples corresponding to a certain number, and respective correlators 1110e, ..., 1110j. Obtained from the output of

,

,

,

,

,

Multiplied by a weight, and the sum of the sum parts and the multiple delayers 1108e, ..., 1108j.

FIG. 8 is a block diagram that calculates the correlation between the matched filter 1 outputs sampled as corresponding to correlator 1110 in FIGS. 5 and 6 and 7. The basic function is to calculate the conjugate of two samples with a distance between two samples and the sum of the b conjugates. As can be seen in Figure 8 consists of the group delay, 2 (b-1) of delay, b multiplication of one group of parallel adders for delaying one sample of delay of a.

를 제외했을 경우에는 도 11과 같이 에너지 계산기(132")의 구조를 단순화 시킬 수 있다.When the weights are the same in FIG. 5, the structure of the first autocorrelator 112 ′ may be simply expressed as shown in FIG. 9. If the weights are the same in FIG. 7, the structure of the energy calculator 132 ′ may be simplified as shown in FIG. 10.

Except for FIG. 11, the structure of the energy calculator 132 ″ may be simplified as shown in FIG. 11.

를 곱해줌으로써 반송파 주파수 오차를 추정한다. Hereinafter, the operation of the symbol timing and the carrier frequency offset configured as described above will be described. When the samples of the matched filter 1 output enter the input of the timing metric calculator 10, the timing metric at that sample position is calculated. The size of the timing metric calculated with respect to the previous sample position is compared in the selector 20, and the timing metric and the sample position corresponding to the value of the timing metric are large are stored. If this process is performed until the value of the sample counter 30 becomes K , the sample position corresponding to the maximum timing metric for the received vector of length K can be known, which is the position of the sample corresponding to the start of the preamble. Is the result of estimating. Then, calculate the phase of the output result of the autocorrelator 1 112 at the position where the timing metric is maximum.

Multiply by to estimate the carrier frequency error.

12 and 13 compare variances of estimates and mean square error (MSE) when symbol timing and carrier frequency offset are estimated according to three methods and the related art according to an embodiment of the present invention. The parameters applied for this performance experiment are as follows.

FFT size: 1024

CP length: 128 samples

Channel model: Vehicular A channel, each multipath component undergoes Rayleigh fading independent of each other.

12 and 13 it can be seen that the performance of the present invention is superior to the performance according to the prior art. In addition, it can be seen that the dispersion of the carrier frequency offset estimation error according to the present invention is closer to the Cramer-Rao boundary than in the prior art.

In the OFDM system according to the present invention as described above, a method and apparatus for estimating a symbol timing and a carrier frequency offset using a cyclic prefix for one preamble and a preamble having a repetitive characteristic may be performed between the CP and the effective OFDM without changing the structure of the preamble. By adding autocorrelation of, it shows better performance than the prior art. In addition, in the present invention, since the structure of the preamble is not changed, there is no reduction in the number of sequences for identifying the supportable base station generated by increasing the number of repetitions. In addition, the present invention provides an advantage that a low complexity system can be implemented due to the recursive calculation of timing metrics.

Although a preferred embodiment of the present invention has been described in detail above, those skilled in the art to which the present invention pertains may make various changes without departing from the spirit and scope of the invention as defined in the appended claims. It will be appreciated that modifications or variations may be made. Therefore, changes in the future embodiments of the present invention will not be able to escape the technology of the present invention.

Claims (10)

A method of estimating symbol timing and carrier frequency offset, a) a timing metric using the correlation between repetitive patterns present in the effective orthogonal frequency division multiplexing (OFDM) symbol interval of the preamble and the correlation between the Cyclic Prefix (CP) and the effective OFDM symbol from the sampled matched filter output Calculating for all sample locations; b) selecting a sample position that maximizes the calculated timing metric; c) calculating autocorrelation from a sampled matched filter output to estimate a carrier frequency offset at the selected sample position; And d) calculating the phase of the calculated autocorrelation output via an autocorrelator. The method of claim 1, Calculating weighted autocorrelation between samples in which the distance between two samples is N / 2 (where N is a fast Fourier transform size) in calculating the timing metric; Calculating weighted autocorrelation between samples having a distance N between the two samples; Calculating the sum by taking the absolute value or the square of the absolute value of each calculated autocorrelation value and performing squares; Calculating an energy of a sampled matched filter output and calculating the inverse of the squared value; And And multiplying the autocorrelation result by the inverse of the energy calculation result. The method of claim 1, Calculating weighted autocorrelation between samples in which the distance between two samples is N / 2 (where N is a fast Fourier transform size) in calculating the timing metric; Calculating weighted autocorrelation between samples having a distance N between the two samples; Calculating a sum by taking an absolute value or a square of an absolute value of each calculated autocorrelation value; Calculating the energy of the sampled matched filter output and calculating the inverse of the calculated value; And multiplying the autocorrelation result by the inverse of the energy calculation result. The method of claim 1, Calculating weighted autocorrelation between samples in which the distance between two samples is N / 2 (where N is a fast Fourier transform size) in calculating the timing metric; Calculating weighted autocorrelation between samples having a distance N between the two samples; And calculating the sum by taking the absolute value or the square of the absolute value of each of the calculated autocorrelation values. The method of claim 1, Calculating weighted autocorrelation between samples in which the distance between two samples is N / 2 (where N is a fast Fourier transform size) in calculating the timing metric; And calculating the absolute value of the calculated autocorrelation value or the square of the absolute value. An apparatus for estimating symbol timing and carrier frequency offset, All samples are sampled from the sampled matched filter output using a timing metric that uses the correlation between repetitive patterns present in the effective orthogonal frequency division multiplexing (OFDM) symbol interval of the preamble and the correlation between the cyclic prefix (CP) and the effective OFDM symbol. Means for calculating with respect to the location; Means for selecting a sample position that maximizes the calculated timing metric; Means for calculating autocorrelation from a sampled matched filter output to estimate a carrier frequency offset at the selected sample position; And Means for calculating a phase of the calculated autocorrelation output via an autocorrelator. The method of claim 6, Means for calculating a weighted autocorrelation between samples in which the distance between two samples is N / 2 (where N is a fast Fourier transform size) in calculating the timing metric; Means for calculating a weighted autocorrelation between samples having a distance N between the two samples; Means for taking the absolute value or the square of the absolute value of each calculated autocorrelation value, calculating a sum, and then performing squares; Means for calculating an energy of a sampled matched filter output and calculating the inverse of the squared value; And And means for multiplying the inverse of the autocorrelation result and the energy calculation result. The method of claim 6, Means for calculating a weighted autocorrelation between samples in which the distance between two samples is N / 2 (where N is a fast Fourier transform size) in calculating the timing metric; Means for calculating a weighted autocorrelation between samples having a distance N between the two samples; Means for calculating the sum by taking the absolute value or the square of the absolute value of each calculated autocorrelation value; Means for calculating the energy of the sampled matched filter output and calculating the inverse of the calculated value; And means for multiplying the inverse of the autocorrelation result and the energy calculation result. The method of claim 6, Means for calculating a weighted autocorrelation between samples in which the distance between two samples is N / 2 (where N is a fast Fourier transform size) in calculating the timing metric; Means for calculating a weighted autocorrelation between samples having a distance N between the two samples; And means for calculating the sum by taking the absolute value or the square of the absolute value of each calculated autocorrelation value. The method of claim 6, Means for calculating a weighted autocorrelation between samples in which the distance between two samples is N / 2 (where N is a fast Fourier transform size) in calculating the timing metric; And means for calculating the absolute value of the calculated autocorrelation value or the square of the absolute value.

Images