KR101081936B1 - Method for estimating frequency offset using training symbol in ofdm system and apparatus using the same - Google Patents

Abstract

The present invention is an OFDM system frequency offset estimation method using training symbols, S2 step of measuring the SNR value, S2 based on the measured SNR value, one of the ML method, EPP method or phase difference method is selected as the offset estimation method The method may include: step S3 of offset estimation according to the selected offset estimation method and step S4 of offset compensation at the receiving end according to the estimated offset.
An efficient OFDM system frequency offset estimation method is selected by selecting an offset estimation method that is optimal for the current communication environment according to the SNR value.

Description

Method and apparatus for estimating off-grid system frequency offset using training symbol TECHNICAL FIELD OF ESTIMATING FREQUENCY OFFSET USING TRAINING SYMBOL IN OFDM SYSTEM AND APPARATUS USING THE SAME

The present invention relates to a method and apparatus for estimating frequency offset in an OFDM system using training symbols. In particular, the present invention relates to a method and apparatus for estimating a frequency offset using a training symbol for performing a frequency offset by selecting an effective one of three techniques for estimating a frequency offset in an OFDM system using a training symbol.

Orthogonal Frequency Division Multiplexing (OFDM) systems use frequencies efficiently and efficiently perform signal propagation even in impulse noise and multipath fading environments.

OFDM systems have been adopted and used in various wireless communication standard modulation schemes such as Wireless Local Area Networks (WLANs), Digital Audio Broadcasting (DAB), and Digital Video Broadcasting (DVB).

However, the performance of the OFDM system is very sensitive to the frequency offset (offset) caused by the Doppler effect due to the movement of the transceiver. The frequency offset can be normalized to the interval of subcarriers and then divided into integer parts and fractional parts according to their size.

Integer frequency offset results in interference to the shift of the subcarrier index of the demodulated OFDM symbol via the Fast Fourier Transform (FFT), and fractional frequency offset breaks the orthogonality between the subcarriers, thereby interfering subcarrier interference (Intercarrier Interference: ICI). Therefore, various frequency offset estimation techniques using multiple symbols have been proposed to prevent serious performance degradation of OFDM system due to frequency offset.

However, the conventional method of estimating the frequency offset using training symbols is not an optimal method according to the situational performance change of each offset technique. As a result, the conventional technology cannot provide a frequency offset estimation method using training symbols suitable for a situation such as a communication environment of each system, and thus does not provide accurate offset estimation or unnecessary waste of excessive resources for offset estimation. There was this.

An OFDM system frequency offset estimation method and apparatus using training symbols according to the present invention aims to solve the following problems.

First, it is intended to enable optimal data communication using the offset estimation technique according to the communication environment.

Second, we want to smooth the resource allocation for the entire communication system by selecting the optimal offset estimation method according to the communication environment.

Third, the communication system reacts sensitively to changes in the communication environment, thereby making effective offset estimation.

The solution to the problem of the present invention is not limited to those mentioned above, and other solutions not mentioned can be clearly understood by those skilled in the art from the following description.

The present invention relates to an OFDM system frequency offset estimation method using training symbols.

In the method according to the present invention, the S1 step in which the SNR value is measured, the S2 step in which one of the ML method, the EPP method, or the phase difference method is selected as the offset estimation method based on the measured SNR value, and the offset according to the selected offset estimation method The estimated S3 step and the S4 step of offset compensation at the receiving end according to the estimated offset.

Step S2 of the present invention includes selecting the ML technique or the EPP technique when the SNR value measured in the S1 stage is less than 24db.

Step S2 of the present invention includes selecting a phase difference technique when the SNR value measured in step S1 is 24 db or more.

The step S1 according to the present invention involves the repetition of the SNR value with a time interval.

The step S2 according to the present invention includes selecting the ML method or the EPP method when the phase difference method is selected as the current offset estimation method and the re-measured SNR value is less than 24db with the time interval in the step S1.

The step S2 according to the present invention includes selecting a phase difference method when the ML method or the EPP method is selected as the current offset estimation method, and the re-measured SNR value is 24 db or more in step S1.

In the step S3 according to the present invention, the offset is estimated according to the estimation technique selected in step S2 for the step of transmitting the n th OFDM sample X _n generated through the inverse fast Fourier transform and the frequency offset component v of the received signal r _n . It includes a step.

OFDM sample X _n according to the present invention is transmitted over a multipath, OFDM sample X _n And Transmitted signal r _n is represented by an expression specified in the following detailed description.

)이 추정되는 것을 포함한다.When the method selected in step S2 according to the present invention is the ML technique, the offset (

) Is estimated.

When the technique selected in the step S2 according to the present invention is an EPP technique, the final frequency offset is estimated by adding all integer frequency offsets, fractional frequency offsets, and remaining frequency offsets estimated by the equation specified in the following detailed description. do.

When the technique selected in the step S2 according to the present invention is a phase difference technique, it includes generating an OFDM symbol repeated m times in the time domain, and the offset is estimated in a manner specified in the following detailed description.

The present invention relates to an OFDM system frequency offset estimation apparatus using training symbols.

The apparatus according to the present invention is an SNR measurement unit for measuring an SNR value, an estimation technique selector for selecting any one of an ML technique, an EPP technique, or a technique using a phase difference based on the value measured by the SNR measurement unit. An offset estimator for which the offset is estimated by the technique selected by the technique selector and an offset estimator for performing offset compensation at the receiver according to the offset value estimated by the offset estimator.

The estimation method selection unit according to the present invention includes selecting the ML technique or the EPP technique when the SNR value measured by the SNR measurement unit is less than 24db.

The estimation method selection unit according to the present invention includes selecting a phase difference technique when the SNR value measured by the SNR measurement unit is 24db or more.

The SNR measurement unit according to the present invention includes the SNR value being repeatedly measured at regular time intervals.

The estimation method selection unit according to the present invention includes selecting the ML technique or the EPP technique when the current phase difference technique is selected and the SNR measurement unit has a constant time interval and the re-measured SNR value is less than 24 db.

The estimation method selection unit according to the present invention includes selecting a phase difference method when the current ML technique or the EPP technique is selected and the SNR measurement unit has a constant time interval and the measured SNR value is 24 db or more.

The offset estimator according to the present invention selects an estimation technique selected by the estimation method selector with respect to the frequency offset component v of the received signal r _n transmitted from the transmitter and the transmitter for transmitting the _nth OFDM sample X _n generated through the inverse fast Fourier transform. And an estimator for estimating the offset according to the technique.

The present invention provides an OFDM system frequency offset estimation method and apparatus using training symbols for selecting an optimal offset estimation technique on the basis of signal-to-noise ratio (SNR).

The present invention provides an OFDM system frequency offset estimation method and apparatus using training symbols that can be changed by an offset estimation technique that is optimal for a current communication environment when the SNR value is changed.

The present invention provides an OFDM system frequency offset estimation method and apparatus using training symbols that do not waste resources of the entire communication system because the optimal offset estimation scheme is selected according to the communication environment (SNR).

The effects of the present invention are not limited to those mentioned above, and other effects not mentioned can be clearly understood by those skilled in the art from the following description.

1 is a flowchart schematically illustrating a method for estimating frequency offset of an OFDM system using training symbols according to the present invention.
2 is a flowchart illustrating an example of changing an offset estimation scheme according to a communication environment.
3 is a block diagram illustrating a training symbol having various repetitive structures in the frequency offset estimation method using the phase difference.
4 is a flowchart illustrating a step-by-step procedure of the frequency offset estimation method using the phase difference.
5 is a block diagram schematically showing the configuration of an OFDM system frequency offset estimation apparatus using training symbols according to the present invention.
6 shows experimental data comparing the performance of three offset estimation techniques according to SNR values in a multipath fading environment.

Hereinafter, an OFDM system frequency offset estimation method and apparatus using training symbols according to the present invention will be described in detail with reference to the accompanying drawings.

The OFDM system frequency offset estimation method using a training symbol according to the present invention is any one of an S1 step in which an SNR value is measured, a maximum likelihood (ML) method, an envelope equalization processing (EPP) method, or a phase difference method based on the measured SNR value. The method includes a step S2 in which the technique is selected as an offset estimation technique, a step S3 in which an offset is estimated according to the selected offset estimation technique, and a step S4 in which the offset is compensated at the receiving end according to the estimated offset.

In the present invention, the signal to noise ratio (SNR) value measured in step S1 is an important criterion. Since the SNR value can be measured using a method or apparatus well known in the communication industry, a description thereof will be omitted.

The reference value of the offset selection according to the present invention is based on experimental results (experimental will be described later) using the ML method, the EPP method, and the phase difference method according to SNR in a multipath fading environment.

Multipath fading refers to a phenomenon in which radio waves encounter obstacles and are alternately received due to interference by reflected waves. After all, the present invention is configured based on the real-life environment in which the actual OFDM is utilized.

Once the offset estimation technique is selected, offset estimation by the selected method is then performed, and offset compensation is performed at the receiver with the estimated value, thereby enabling data transmission and reception with low error.

In an embodiment according to the present invention, the offset estimation selection in step S2 is characterized in that, when the SNR value measured in step S1 is less than 24db, the ML technique or the EPP technique is selected.

As another example, the ML technique and the EPP technique may also be selected based on specific criteria. Since the ML technique is complex but allows the most accurate estimation, it may be selected again based on system performance or factors affecting the communication environment in addition to the SNR value.

For example, if there are enough resources (resources) of the OFDM system to send the current signal, the ML technique is preferably selected. On the contrary, when the performance of the OFDM system is low or there are not enough resources, it is preferable to select the EPP scheme.

In another embodiment according to the present invention, the offset estimation selection in step S2 is characterized in that the phase difference technique is selected when the SNR value measured in step S1 is 24db or more.

Since the phase difference technique is less complicated than the ML technique and the EPP technique, it is preferable to select the phase difference technique if the offset estimation is properly performed even when using the phase difference technique.

As another embodiment according to the present invention, the SNR value measurement may be repeatedly performed at regular time intervals in the step S1. A regular time interval will typically be in seconds or decades, but in some cases it may be in microseconds or nanoseconds. This time interval may vary according to communication technology and communication environment changes such as large data being able to be transmitted in a very short time.

2 is a flowchart illustrating an example of changing an offset estimation scheme according to a communication environment (change in SNR value).

If the re-measured SNR value with a predetermined time interval is different from the selection criterion of the previous offset estimation technique, the offset estimation technique may be changed. For example, the SNR value measured immediately before the re-measurement is less than 24db, so that the ML method or the EPP method is selected. When the re-measured SNR value is more than 24db, the offset estimation technique is changed to the phase difference technique.

That is, there may be two embodiments. First, when the phase difference method is selected as the current offset estimation method, and the re-measured SNR value with a time interval is less than 24db, the offset estimation method may be changed by the ML method or the EPP method.

Second, when the ML technique or the EPP technique is selected as the current offset estimation technique, and the re-measured SNR value with a time interval is 24 db or more, it may be changed to a phase difference technique.

Hereinafter, a method according to the present invention will be described as a test example of transmitting the n-th OFDM sample X _n generated through an inverse FTT in multipath.

In the step S3 where the offset is estimated according to the present invention, the step of transmitting the _nth OFDM sample X _n generated through the inverse fast Fourier transform and the estimation scheme selected in step S2 for the frequency offset component v of the received received signal r _n are performed. Estimating the offset accordingly. nth OFDM sample X _n And the received received signal r _n is expressed by the following equation.

In Equation 1, k denotes a subcarrier index, N denotes an inverse Fourier transform size, C _k denotes a complex symbol in a frequency domain transmitted by a k th subcarrier, and j denotes an imaginary number. number).

인 덧셈꼴 백색 정규 잡음, v는 주파수 옵셋 성분을 의미한다.In Equation 2, Sn is a signal component represented by Equation 3 below, w (n) has an average of 0 and a variance.

A white normal noise, v, denotes a frequency offset component.

In Equation 3, h _l represents the number of channels, L represents the number of multipaths, and Xn represents the nth OFDM sample X _n represented by Equation 1.

S2  The method chosen in step ML  Technique

The ML technique has very high frequency offset estimation performance for carrier frequency and channel response, and can estimate the frequency offset of the entire region, but it is complicated because it investigates the possibility of frequency estimation offset for all values.

The likelihood function for v, h _l of the received signal r _n can be expressed as Equation 4 below.

는

인 대각행렬,

는

의 시험값,

,

는

의 시험값,

,

는

의 시험값, A는

과

을 만족하는 행렬

를 의미한다. 수학식 4를 최대화 하는 조건은 하기의 수학식 5를 최대화 하는 조건과 같다.In equation (4)

Is

Diagonal matrix,

Is

Test value of,

,

Is

Test value of,

,

Is

Test value,

and

Matrix satisfying

Means. The condition for maximizing Equation 4 is the same as the condition for maximizing Equation 5 below.

를 추정 주파수 옵셋

로 결정하는 식은 하기의 수학식 6이다.The variable used in Equation 5 is the same as Equation 4, and H means Hermitian transpose. Test Value Maximizing Equation 5

Estimate frequency offset

Is determined by the following equation (6).

를 조사하여 추정 주파수 옵셋

를 추정한다. ML 기법은 정밀한 추정이 가능하고 주파수 옵셋 추정 범위가

로 넓지만, 구현이 복잡하고 계산량이 많다.
As shown in Equation 6

To estimate the estimated frequency offset

Estimate The ML technique allows for accurate estimation and frequency offset estimation range

Although wide, the implementation is complex and computational.

S2  The method chosen in step EPP  Technique

In order to estimate the frequency offset independently of the training symbol, the EEP technique is applied to the received OFDM symbol. The EPP technique estimates frequency offset by dividing the frequency offset into three stages: integer, decimal, and the rest using a periodogram. EEP is defined by Equation 7 below.

Where * means conjugate. The training symbol generated through the EPP at the receiving end is represented by Equation 8 below.

이고, 수신 시혼의 주파수 f_k에서 주기도표는 하기의 수학식 9와 같이 정의된다.In equation (8)

The period diagram at the frequency f _k of reception time horn is defined as in Equation 9 below.

The variables in the equations described in the EPP technique are the same as described in the ML technique. The EPP method estimates integer, decimal, and remaining frequency offsets in three steps, and then adds the estimated three frequency offsets to finally estimate the frequency offset. Each frequency offset equation is defined by Equations 10 (integer frequency offset), Equation 11 (fractional frequency offset), and Equation 12 (remaining frequency offset).

S2  The method chosen in step Phase difference  Technique

In the phase difference scheme, as shown in Equation 13, data can be transmitted every m-th subcarrier, and 0 can be transmitted to subcarriers other than the m-th subcarrier to obtain an OFDM symbol repeated m times in the time domain.

Assuming that the transmitter transmits this OFDM symbol, it can be seen that the OFDM symbol transmitted as shown in FIG. 3 has several repetition periods other than m (M = 2, 4, 8 (= m)). B in FIG. 3 means a block repeated in one OFDM symbol.

The frequency offset generates a positive phase rotation proportional to each sample index of the OFDM training symbol in the time domain. The frequency offset is estimated using the phase difference in the repeated training symbol generated as described above. A detailed frequency offset estimation equation is shown in Equation 14 below.

는 하기의 수학식 15와 같이 정의된다.Where i = 1,2, ...., log ₂ m. When estimating the frequency offset using the phase difference of the repeated training symbol, as the value of m increases, the estimation range of the frequency offset becomes wider, but the accuracy of the estimated frequency offset is lowered. Therefore, the process of FIG. 4 is performed to widen the estimation range and increase the accuracy.

Is defined as in Equation 15 below.

Hereinafter, an OFDM system frequency offset estimation apparatus using training symbols according to the present invention will be described in detail. However, descriptions common to the OFDM system frequency offset estimation technique using training symbols will be omitted and the description will be mainly focused on the core configuration of the apparatus.

5 schematically illustrates a configuration of an OFDM system frequency offset estimation apparatus using training symbols according to the present invention.

The OFDM system frequency offset estimation apparatus using the training symbol according to the present invention is based on the SNR measurement unit 100, the SNR measurement unit, the ML technique, the EPP technique, or the phase difference technique. An estimation method selecting unit 200 in which one of the techniques is selected, an offset estimator 300 and an offset estimator 300 in which an offset is estimated by the technique selected by the estimation method selecting unit 200 according to the estimated offset value The offset estimator 400 performs offset compensation at.

As one embodiment according to the present invention, the estimation method selection unit 200 of the present invention selects the ML method or the EPP method when the SNR value measured by the SNR measurement unit 100 is less than 24db, and the measured SNR value is In case of more than 24db, it is characterized by selecting a phase difference technique.

As another embodiment according to the present invention, the SNR measuring unit 100 measures the SNR value repeatedly at regular time intervals, and selects an optimal estimation method based on this.

That is, the estimation method selector 200 according to the present invention may select the ML method or the EPP method when the current phase difference method is selected and the SNR measured by the SNR measuring unit 100 has a predetermined time interval less than 24 db. have.

In another embodiment according to the present invention, the estimation method selection unit 200 is a phase difference method when the current ML method or the EPP method is selected and the SNR measurement unit has a predetermined time interval and the measured SNR value is 24 db or more. Can be selected.

The method of claim 12,

The offset estimator 300 according to the present invention selects an estimation scheme for a frequency offset component v of a transmitter for transmitting the _nth OFDM sample X _n generated through an inverse fast Fourier transform and a received signal r _n transmitted from the transmitter. An estimation unit for estimating the offset according to the estimation technique selected by the unit 200 is included.

FIG. 6 is experimental data comparing the performance of three offset estimation schemes according to SNR values in a multipath fading environment, and shows mean square error (MSE) according to SNRs of three techniques.

The conditions for comparing the performance of the three offset estimation techniques are as follows. OFDM symbols for performance comparison were generated by quadrature phase shift keying (Quadreture PSK, QPSK), and the size of inverse fast Fourier transform (IFFT) was 64.

The guard interval was set at 8 samples and the maximum Doppler frequency was set at 125 Hz. The multipath fading environment has six paths, each of which is delayed by 0, 1, 2, 3, 4, and 5 samples, respectively. The magnitude A _i of the l-th path of the channel varies independently of the Rayleigh distribution and is exponentially reduced.

 The experimental results show that the MSE of the ML technique (Method 1) is the lowest for all SNRs. However, the ML technique has a disadvantage in that it requires more investigation than all the test values, which is more complicated than other techniques.

In the case of the EPP technique (Method 2), which estimates the frequency offset over three stages, it can be seen that it closely follows the frequency offset estimation performance of Technique 1.

Finally, the phase difference technique (Method 3) shows a significant difference in the frequency offset estimation performance according to the estimation accuracy of the integer frequency offset. At low SNRs, integer frequency offsets cannot be estimated accurately, which results in very large MSE. However, at high SNR, the estimation accuracy of integer frequency offset is increased and it is similar to MSE of technique 1 and technique 2.

As a result, ML technique or EPP technique has good performance at low SNR and low complexity phase difference technique is more efficient at higher SNR of 24 dB or more. The present invention is to analyze the experimental results and to select the optimal offset estimation technique according to the communication environment (SNR).

The embodiments and drawings attached to this specification are merely to clearly show some of the technical ideas included in the present invention, and those skilled in the art can easily infer within the scope of the technical ideas included in the specification and drawings of the present invention. Modifications that can be made and specific embodiments will be apparent that both are included in the scope of the invention.

100: SNR measurement unit 200: estimation method selection unit
300: offset estimator 400: offset estimator

Claims (22)

S1 step of measuring the SNR value;
An S2 step of selecting one of an ML technique, an EPP technique, or a phase difference technique as an offset estimation technique based on the measured SNR value;
Step S3, in which an offset is estimated according to the selected offset estimation method; And
In step S4, the offset is compensated at the receiving end according to the estimated offset,
In step S2, if the SNR value measured in the step S1 is less than 24db, the ML method or the EPP method is selected, and when the SNR value measured in the step S1 is 24db or more, a training symbol, characterized in that for selecting a phase difference method OFDM system frequency offset estimation method. delete delete The method of claim 1,
In step S1, an OFDM system frequency offset estimation method using training symbols, wherein the SNR value is repeatedly measured at a time interval. The method of claim 4, wherein
The step S2
If the phase difference scheme is selected as the current offset estimation scheme and the re-measured SNR value is less than 24db in step S1, the OFDM system frequency offset estimation using the training symbol is selected. Way. The method of claim 4, wherein
The step S2
If the ML technique or the EPP technique is selected as the current offset estimation technique and the re-measured SNR value is 24 db or more in the step S1, the OFDM system frequency offset estimation using a training symbol is selected. Way. The method of claim 1,
The step S3
Transmitting the n th OFDM sample X _n generated through an inverse fast Fourier transform; And
And estimating the offset according to the estimation technique selected in step S2 for the frequency offset component v of the received received signal r _n . The method of claim 7, wherein
The OFDM sample X _n is transmitted over a multipath, and the OFDM sample X _n And An OFDM system frequency offset estimation method using a training symbol, _wherein the received received signal r _n is represented by the following equation.

(k is the subcarrier index, N is the inverse Fourier transform size, C _k is the complex symbol in the frequency domain transmitted by the kth subcarrier, j is an imaginary number)

(Sn is

Where h _l is the number of channels, L is the number of multipaths, w (n) is 0, and the variance is

Additive white normal noise) The method of claim 8,
If the method selected in the step S2 is an ML technique,
Offset by the formula

OFDM system frequency offset estimation method using a training symbol, characterized in that is estimated.

(

Is

Maximization condition, where

Is

Diagonal matrix, denoted by

Is

Test value of,

, A

and

To satisfy

Matrix, H is Hermitian transpose) The method of claim 8,
If the technique selected in step S2 is an EPP technique,
The final frequency offset is estimated by adding all integer frequency offsets (Equation 1 below), fractional frequency offsets (Equation 2 below), and remaining frequency offsets (Equation 3 below) estimated by the following equations. OFDM system frequency offset estimation method using training symbols.

(Eq. 1)

(Eq. 2)

(Eq. 3)
(

Is

Is a period diagram of the frequency of the received signal, where

Silver EPP,

Is the n th OFDM sample generated through the inverse fast Fourier transform, c _k is a complex symbol in the frequency domain transmitted by the k th subcarrier, N is the magnitude of the inverse Fourier transform, k is the subcarrier index,

Is a training symbol that has undergone EPP at the receiving end,
At this time

Where h _l is the number of channels, L is the number of multipaths, N is the magnitude of the inverse Fourier transform, w (n) is 0, and the variance

White normal noise, where x _n is the nth OFDM sample produced by an inverse fast Fourier transform) The method of claim 8,
If the technique selected in the step S2 is a phase difference technique,
OFDM system frequency offset estimation method using a training symbol characterized by generating an OFDM symbol repeated m times in the time domain and the offset is estimated by the following equation.

(i = 1,2, .... ,, log ₂ m, r _n is the received signal for x _n , where N is the magnitude of the inverse Fourier transform) An SNR measuring unit measuring an SNR value;
An estimation technique selection unit for selecting one of an ML technique, a technique using EPP, or a technique using a phase difference based on the value measured by the SNR measurement unit;
An offset estimator configured to estimate an offset by a technique selected by the estimation technique selector; And
An offset estimator configured to perform offset compensation at a receiving end according to the offset value estimated by the offset estimator,
The estimation method selection unit selects a ML method or an EPP method when the SNR measured by the SNR measuring unit is less than 24 db, and selects a phase difference technique when the SNR measured by the SNR measuring unit is 24 db or more. OFDM system frequency offset estimation apparatus using symbols. delete delete The method of claim 12,
And an SNR measurement unit having a predetermined time interval and repeatedly measuring an SNR value. 16. The method of claim 15,
If the current phase difference technique is selected and the SNR measurement unit has a predetermined time interval and the re-measured SNR value is less than 24db, the OFDM method using a training symbol, characterized in that for selecting the ML method or EPP method Frequency offset estimation device. 16. The method of claim 15,
The OFDM method using the training symbol selects the phase difference method when the current ML method or the EPP method is selected and the SNR measurement unit has a predetermined time interval and the re-measured SNR value is 24db or more. Frequency offset estimation device. The method of claim 12,
The offset estimator
A transmitter for transmitting the n th OFDM sample X _n generated through an inverse fast Fourier transform; And
And an estimation unit for estimating an offset according to the estimation technique selected by the estimation technique selection unit with respect to the frequency offset component v of the received signal r _n transmitted from the transmission unit. . The method of claim 18,
The transmitter transmits the OFDM samples X _n through a multipath, and the OFDM samples X _n and OFDM system frequency offset estimation apparatus using a training symbol characterized in that the received received signal r _n is represented by the following equation.

(k is the subcarrier index, N is the inverse Fourier transform size, C _k is the complex symbol in the frequency domain transmitted by the kth subcarrier, j is an imaginary number)

(Sn is

Where h _l is the number of channels, L is the number of multipaths, w (n) is 0, and the variance is

Additive white normal noise) 20. The method of claim 19,
If the method selected by the estimation method selection unit is an ML technique,
Offset by the formula

OFDM system frequency offset estimation apparatus using a training symbol, characterized in that is estimated.

(

Is

Maximization condition, where

Is

Diagonal matrix, denoted by

Is

Test value of,

, A

and

To satisfy

Matrix) 20. The method of claim 19,
If the technique selected by the estimation technique selection unit is an EPP technique,
The final frequency offset is estimated by adding all integer frequency offsets (Equation 1 below), fractional frequency offsets (Equation 2 below), and remaining frequency offsets (Equation 3 below) estimated by the following equations. An OFDM system frequency offset estimation apparatus using training symbols.

(Eq. 1)

(Eq. 2)

(Eq. 3)
(

Is

Is a period diagram of the frequency of the received signal, where

Is the EPP, x _n is the n th OFDM sample generated through the inverse fast Fourier transform, c _k is a complex symbol in the frequency domain transmitted by the k th subcarrier, N is the magnitude of the inverse Fourier transform, k is the subcarrier index, and j is Imagination,

Is a training symbol that has undergone EPP at the receiving end,
At this time

Where h _l is the number of channels, L is the number of multipaths, N is the magnitude of the inverse Fourier transform, w (n) is 0, and the variance

White normal noise, where x _n is the nth OFDM sample produced by an inverse fast Fourier transform) 20. The method of claim 19,
If the technique selected by the estimation technique selection unit is a phase difference technique,
OFDM system frequency offset estimation apparatus using a training symbol characterized by generating an OFDM symbol repeated m times in the time domain and the offset is estimated by the following equation.

(i = 1,2, .... ,, log ₂ m, r _n is the received signal for x _n , where N is the magnitude of the inverse Fourier transform)

Images