KR20030017057A - Frequency error detection mehod for wireless lan receiver - Google Patents

Abstract

PURPOSE: A method for detecting a frequency error of a wireless LAN(Local Area Network) receiver is provided to use a repetition characteristic and the relation between a real number part and an imaginary number part in a short preamble of 802.11a, that is a wireless LAN standard, so as to detect a frequency error and estimate a phase error. CONSTITUTION: A correlation value among identical data having the shortest distance from received signals is calculated, and a frequency error is estimated using the calculated correlation value. The frequency error is corrected using a frequency error estimation value. A real number part and an imaginary number part of sample data, that are arranged in the latter part corresponding to a half of samples in a short preamble for which the frequency error is corrected, are mutually changed. Correlation values between samples of each sample group are calculated, and a phase error is estimated using a calculated result. And the estimated phase error is corrected by the sample.

Description

Frequency error detection method of wireless LAN receiver {FREQUENCY ERROR DETECTION MEHOD FOR WIRELESS LAN RECEIVER}

The present invention relates to a reception technology of a wireless LAN system using Orthogonal Frequency Division Multiplexing (OFDM), and more particularly, to a wireless LAN suitable for detecting a frequency error of a received signal by using a repetition characteristic of a preamble. The present invention relates to a frequency error detection method of a receiver.

As demand for multimedia services increases, research on high-speed data transmission is more active than ever. Orthogonal frequency division multiplexing is a typical transmission scheme. This orthogonal frequency division multiplexing not only enables high-speed transmission but also uses a plurality of subcarriers so that the frequency band can be efficiently used. In addition, by inserting a guard interval in front of the symbol it can reduce the symbol interference (Inter-Symbol Interference). For this reason, it has been applied to DAB (Digital Audio Broadcasting) and DVB (DVB) in Europe and has been adopted in 802.11a, which is a short-range wireless LAN standard.

In order to demodulate a received signal in an orthogonal frequency division multiplexing system, two synchronization elements are largely detected and a process of correcting the received signal is required. The first is the symbol synchronization process to find the starting point of the symbol, and the second is the frequency synchronization process to find the frequency error caused by the oscillator error of the receiver.

Conventional techniques include a method of detecting sync using a guard interval inserted before a symbol and a method of detecting sync using a preamble.

In general, symbol synchronization and frequency synchronization are performed using the result of the correlator calculation, and the FFT is performed by correcting the received signal with the estimated result. This demodulates the quadrature frequency division multiplexing receiver.

1 illustrates a preamble structure of 802.11a, a wireless LAN standard. The same data string is repeated from the first to the tenth section (Short Preamble), which are denoted as S1 to S10. It is recommended that the first seven of these 10 intervals be used for signal detection, automatic gain control (AGC), and the other three intervals for time synchronization and integer frequency error detection. The section denoted by GI becomes a guard interval of L1 to L2 symbols which follow. The L1 and L2 intervals should be used for channel estimation and prime frequency error estimation.

Here, the integer frequency error means a frequency error corresponding to the distance between subcarriers in the frequency spectrum of OFDM, and the decimal frequency error means less than an integer multiple error.

In the prior art, integer multiple error was estimated using S8 to S10, and prime frequency error was estimated using L1 to L2. In both cases, the frequency error is estimated using the correlation between the same data. When using S8 to S10, data at 16 sample distances are calculated. When using L1 to L2, data at 64 sample distances are calculated. The calculation method for estimating the frequency error basically uses the Moose method. In estimating this frequency error, it is necessary to assume that the time synchronization of the symbol is already known.

Equation 1 below is a calculation of correlation values between the same data when estimating a frequency error.

Here, R () represents a received sample and assumes that the symbol time synchronization is already known. Therefore, the correlation value is obtained by using the repeating section that is separated by exactly 16 sample distances.

Equation 2 below shows the process of obtaining the frequency error from the calculated correlation value.

here, Denotes an estimated integer error. The range of integer frequency error that can be finally estimated in Equation 2 is -2.0 < <2.0.

Integer frequency error estimates are also expressed in decimal frequency error ranges. However, since the accuracy is a problem, it is necessary to estimate the exact minority frequency error later using the L1 to L2 preambles. The equation for estimating prime frequency error is similar to the case of estimating integer frequency error. In other words, the distance between the samples for calculating the correlation values is the same except for the difference.

Equation 3 below is a calculation of correlation values between the same data when estimating the prime frequency frequency error.

In Equation 3, a correlation value is obtained by using a repetition interval in which a 64 sample distance is apart.

Equation 4 below shows a process of obtaining a frequency error from the calculated correlation value.

here, Represents the estimated frequency error, and the range of integer frequency error that can be finally estimated in Equation 4 is -0.5 < <0.5.

Since the result obtained through [Equation 2] represents the integer part, the integer frequency error is corrected based on this. Subsequently, the minority frequency error is corrected as a result obtained through Equation 4.

However, in the conventional frequency error detection method, the same data between 16 samples is used for the short preamble, which is used to calculate the integer frequency error as shown in [Equation 2], and the estimation range is Is less than absolute value 2. Therefore, when the frequency range of the local oscillator of the receiver generates a frequency error larger than this range, there is a defect that the frequency error cannot be estimated using only a correlation value of 16 sample distances.

Accordingly, an object of the present invention is to provide a frequency error detection method for improving the performance of a receiver in an orthogonal frequency division multiplexing system.

1 is a format diagram of a preamble of 802.11a, which is a wireless LAN standard.

2 is an explanatory diagram of a reception process of a receiver of an orthogonal frequency division multiplexing system according to the present invention;

3 is a sample format diagram showing characteristics of a preamble used when estimating an integer multiple frequency error.

4 is a graph showing the characteristics of the real part and the imaginary part of the short preamble used when estimating the phase error.

5 is a sample format diagram showing an arrangement of preambles used for phase error estimation.

6 is a signal flowchart of a phase error estimation process according to the present invention;

** Description of symbols for the main parts of the drawing **

S1-S6: Steps 1-6

According to the first aspect of the present invention, the performance of the frequency error detection apparatus is improved by using the repetition feature of the short preamble of the 802.11a, which is a wireless LAN standard.

According to the second aspect of the present invention, the phase error is estimated using the relationship between the real part and the imaginary part of the short preamble of the 802.11a wireless LAN standard.

In the method of detecting a frequency error of a WLAN receiver according to the present invention, a correlation value is calculated between the same data in a received signal by multiplying the received data first by the public complex number of the received data and estimating the frequency error using the same. A first process of doing; A second step of correcting a frequency error using the frequency error estimate; The sample of the short preamble whose frequency error is corrected is classified into the first half of the sample group and the remaining half of the sample group, and then the real part and the imaginary part of the rear part sample data are interchanged. Calculating a value and then estimating a phase error using the calculated result; The fourth step of correcting the estimated phase error in each sample unit will be described in detail with reference to FIGS. 2 to 6 attached to the operation of the present invention.

First, the reception process of the receiver will be described with reference to FIG. 2.

Correlation values are calculated between the same data in the received signal, and the frequency error is estimated using the same. The correlation value is obtained by multiplying the first received data among the same data by the conjugate complex number of the later received data. Embodiment of the present invention relates to a wireless LAN standard 802.11a. When estimating the frequency error, the process of estimating the frequency error using the same samples having a short distance was added to the conventional method. The feature of this method is that it can widen the range of frequency errors that can be estimated. In addition, in the present invention, the phase error may be estimated using the real part and the imaginary part of the short preamble of the 802.11a standard. These estimates are used to correct the received signal. The corrected signal is restored to a transmission signal through a Fast Fourier Transform (FFT).

The present invention deals with the estimation of the frequency error and the estimation of the phase error in the time domain. In addition, in the present invention, it is assumed that the symbol start point estimation is performed first. Frequency error estimation is performed using a short preamble and a long preamble, and this estimation method is added to a conventional method to improve performance. After the estimated frequency error is corrected, the phase error is estimated and corrected using the real and imaginary parts of the short preamble in the time domain.

For reference, in the conventional method, the same data between 16 samples was used for the short preamble. This is used to find the integer frequency error as in [Equation 2], the estimated range is less than 2. If the local oscillation frequency of the receiver produces an error larger than this range, it cannot be estimated using only the correlation value of 16 sample distances.

Therefore, in the present invention, it is possible to estimate a wider range of frequency error by using a correlation value between samples with close distances. The short preamble is repeated 10 times using 16 samples as a unit. In the present invention, samples having a short distance among the 16 samples are used for frequency error estimation.

3 shows the same samples used in the present invention among the 16 samples of the short preamble. Basically, it can be seen that the same data is repeated in units of 16 samples. In addition, there is also the same data that is 8 samples like A and 4 samples such as B and C.

Equation 5 below calculates a correlation between A samples.

Here, since it is assumed that the time synchronization of the symbol is already known, n indicates the position of A (1) or A (2).

Equation 6 below is a formula for estimating a frequency error using a correlation value between A samples.

The estimated range according to [Equation 6] is -4.0 < <4.0.

Equation 7 below calculates the correlation between B and C samples.

Here, since it is assumed that the time synchronization of the image symbol is known, n indicates the position of B (1) or C (1).

Equation 8 below is a formula for estimating a frequency error using a correlation value between B or C samples.

The estimated range according to Equation 8 is -8.0 < <8.0.

The frequency error estimation method according to the present invention estimates a wider range of frequency errors using samples spaced apart by 4,8 sample distances in addition to data spaced apart by the conventional 16 sample distances. Since the number of identical data is small and the accuracy of the estimate is relatively low, it should be reflected in the calculation along with the data of 16 sample distances.

After the estimation of the frequency error, the frequency error is corrected by this. After that, the phase error is estimated using the data corrected for the frequency error.

Equation 9 below shows a phase error included in the received signal.

Here, T (n) represents a transmission signal when the influence of the channel is not taken into account. R (n) represents a received signal that includes a frequency error and a phase error. And, Denotes a phase error. N represents the number of subcarriers in accordance with the 802.11a standard. Phase error is equally included in all data. Therefore, in the equation for estimating the frequency error, the phase error item is eliminated, which is not a problem. In the present invention, the feature of the short preamble is used to estimate the phase error. The short preamble has 16 samples repeated 10 times in succession.

4 shows data of a short preamble drawn in the time domain. If these data are drawn separately from the real part and the imaginary part, it can be seen that they appear in the same form after a certain delay time. Delaying exactly eight samples will match the real and imaginary parts. In the present invention, the phase error psi is obtained based on these characteristics of the real part and the imaginary part.

5 shows the characteristics of the real part and the imaginary part as described above. 16 complex samples can be represented by C1-C16. Each complex sample is multiplied by a phase error.

FIG. 6 illustrates a process of estimating a phase error using characteristics of data as shown in FIG. 4. here, Represents an arbitrary phase error, Denotes a frequency error.

First, in order to obtain a phase error in the short preamble, 16 samples having a frequency error corrected are prepared (S1).

The 16 samples are divided into eight samples at the front and eight samples at the rear, and then the real part and the imaginary part of the last eight sample data are replaced with each other because the phase error item disappears when the conventional method is processed. (S2)

Equation 10 below shows an example of replacing the real part and the imaginary part.

Here, it can be seen that when the real part and the imaginary part are interchanged with each other, the sign of the phase error also changes.

Then, the correlation value between the first eight samples and the last eight samples is calculated. Equation 11 below shows an equation for calculating the correlation value of the first sample.

Then, the phase error using the correlation value Where is estimated phase error The range of does not represent the full range from Intrinsic phase error because it represents only the range up to Or It may be (S4).

Therefore, through the following fifth step (S5) Will be determined.

The estimated phase error to determine the phase error The real part and the imaginary part are substituted into the sample that is replaced. In this case, the phase is subtracted from the phase of the front sample and added to the phase of the rear sample, and then calculated as in the case of obtaining the correlation values for these samples. If this result is positive Is a phase error, negative Becomes the phase error (S5).

Finally, in the phase error correction step S6, the estimated phase error is corrected for each sample unit (S6).

In the first embodiment of the present invention, when a frequency error is estimated by a receiver of an orthogonal frequency division multiplexing system conforming to the 802.11a standard, data having a distance of 4 or 8 of preambles is used.

On the other hand, in the second embodiment of the present invention, it is used to estimate the phase error using the identity of the real part and the imaginary part of the preamble. In addition, the method of changing the position of the real part and the imaginary part includes using the phase error estimation process.

As described in detail above, the present invention detects a frequency error up to an absolute value of 8 using a repetition characteristic of a short preamble of 802.11a, which is a wireless LAN standard, in a receiver of an orthogonal frequency division multiplexing (OFDM) system, and performs a short preamble. By estimating the phase error using the relationship between the real part and the imaginary part of, the performance of the receiver is further improved.

In addition, there is an effect that can stably detect the frequency error, phase error in the time domain at the time of data reception.

Claims (7)

Calculating a correlation value between the same data having a short distance from the received signal and estimating a frequency error using the calculated correlation value; A second step of correcting a frequency error using the frequency error estimate; Replace the real part and the imaginary part of the latter sample data corresponding to half of the samples of the short preamble whose frequency error is corrected, calculate the correlation value between the samples of each sample group, and then use the calculated result to phase Estimating an error; And a fourth process of correcting the estimated phase error on a per-sample basis. The method of claim 1, wherein the first process includes estimating a phase error using the real part and the imaginary part of the short preamble of the 802.11a standard. The method of claim 1, wherein the correlation value of the first process is obtained by multiplying the received complex data by the empty complex number of the received data later. The method of claim 1, wherein the first process uses the following equation when obtaining a correlation value between 8 samples (A) repeated among 16 samples of the short preamble. . Here, assume that you already know the time synchronization of symbols, where n is the position of A (1) or A (2). The method of claim 1, wherein the third process changes the real part and the imaginary part as shown in the following equation. The method of claim 1, wherein the third process is performed by using the following equation in the case of the first sample when calculating a correlation value between samples of each sample group. The method of claim 1, wherein the third process comprises: estimated phase error A first step of subtracting the value from the phase of the front sample and adding the phase of the rear sample when assigning a value to a sample in which the real part and the imaginary part are replaced; Calculate the correlation value for the samples, Is determined by the phase error, And detecting a phase error as a phase error.