KR20090013957A - Apparatus and method for compensation of channel impulse response estimation error in orthogonal frequency division multiplexing systems - Google Patents

Abstract

An apparatus and a method for compensation of estimation error are provided to improve spectrum leakage and noises by compensating for estimation error of a channel impulse response before selecting a valid channel impulse response. A initial channel estimation machine(403) estimates a primary channel. By performing IFFT calculation in the presumed primary channel value, An IFFT(Inverse Fast Fourier Transform) operator(404) estimates CIR(Channel Impulse Response). A maximum power route detection part(420) detects route has the maximum power reserve about presumed CIR. A CIR error compensation part(430) uses the CIR value and the CIR estimation error function of the maximum power route in order to compensate for the CIR estimation error about all routes.

Description

Apparatus and method for compensating the estimation error of channel impulse response in orthogonal frequency division multiplexing system

The present invention relates to an Orthogonal Frequency Division Multiplexing (OFDM) system, and more particularly to an apparatus and method for compensating for an error of estimation of a channel impulse response (CIR) in a Discrete Fourier transform (DFT) based channel estimation algorithm. It is about.

Orthogonal Frequency Division Multiplexing (hereinafter, referred to as 'OFDM') is a method of slowing the transmission of input data in parallel on multiple carriers instead of high-speed transmission on a single carrier, thereby affecting frequency selective fading or narrowband interference. It is a way to receive less. The OFDM scheme has good spectral efficiency because the spectra of subchannels overlap each other while maintaining mutual orthogonality.

In an OFDM system, the transmitted signal is modulated by an Inverse Fast Fourier Transform (hereinafter referred to as IFFT) and the received signal is demodulated by a Fast Fourier Transform (hereinafter referred to as 'FFT'). In this way, an efficient configuration of the digital modulation and demodulation section is possible. The biggest advantage of this configuration is that the receiver can be simply configured with an equalizer that requires only one complex multiplication for each carrier.

The interference caused by multipath fading in the OFDM scheme can be reduced by increasing the symbol period in proportion to the number of subchannels. For this purpose, a guard interval is inserted for every predetermined number of transmitted symbols. In the guard period, the last symbols of a frame composed of a number of symbols corresponding to the number of carriers are typically copied, and these symbols are called a cyclic prefix (hereinafter, referred to as 'CP'). The length of the CP should be longer than the length of the channel impulse response (hereinafter, referred to as 'CIR') indicating the channel characteristics of the radio channel. If the length of the CIR is longer than the length of the CP, the performance of the system is degraded due to the influence of inter-channel interference (ICI) and inter-symbol interference (ISI). In an actual system, the length of CP is given in advance. Due to the nature of the air interface having random characteristics, CIRs having a period longer than a given CP length may appear during transmission. In this case, the transmission characteristics of the radio channel also appear according to the frequency selective characteristics.

In the OFDM system, the received signal is significantly influenced by the frequency selective fading effect due to the multipath delay of the channel and the Doppler frequency due to the increase of the moving speed. In particular, the coherence time is increased when the moving speed increases and the Doppler shift becomes deeper. Due to the decrease of, the channel change increases even within the subframe.

In the channel estimation algorithm for compensating the influence of the channel, LS (Least Square) using a pilot of an OFDM symbol, a channel frequency response (Channel CFR) is estimated in the frequency domain. Minimum Mean Squares Error (MMSE) using correlation. In addition, there are channel estimation methods based on Discrete Fourier transform (DFT), Discrete Cosine Transform (DCT), and Time Domain Processing (TDP) as a method of estimating a channel impulse response (CIR) in the time domain. In the above three channel estimation methods, only the components that are valid in the CIR estimated through the IFFT are selected, and the components other than the effective CIR values are regarded as noise and forcibly replaced by zero, thereby reducing the influence of noise. There is a common feature to improve estimation performance.

However, in the three channel estimation methods, all of the components of the CIR selected according to the environment of the channel used may not be valid CIRs. In this case, there is a disadvantage that does not sufficiently eliminate the influence of noise. Another problem of the above scheme is that a guard band (hereinafter, referred to as 'GB') in which information is not transmitted in order to reduce interference with an adjacent channel in an OFDM communication system is a high frequency band and a low frequency band. The subcarriers are assigned to subcarriers, respectively, and CIR estimation errors occur due to the guard band setting. That is, a zero value is assigned to a GB in which information is not transmitted, which causes a spectral leakage by causing a sudden change in a signal at an available subcarrier section and a boundary section of a GB section. Due to the spectral leakage, leakage components appear in the CIR in the region beyond the maximum delay time of the channel, and even within the maximum delay time of the channel, the CIR spreads around the ideal position compared to the impulse model (Equation 2). Will appear. Therefore, if a CIR estimation error occurs and only the influence of noise is eliminated without supplementing the CIR estimation error, the channel estimation signal CFR after the FFT is used to correct the distortion of the channel estimation at both ends of the frequency domain. It becomes visible and has a large performance degradation.

1 is a diagram illustrating an example of an error of estimation of CIR due to the setting of GB in an OFDM system. Referring to FIG. 1, it can be seen that a phenomenon in which CIR having an impulse delay profile is dispersed by setting of GB. In addition, FIG. 2 is an exemplary diagram illustrating MSE performance for each subcarrier of the conventional scheme in the frequency domain. Referring to FIG. 2, it can be seen that the estimated CFR due to the CIR estimation error shows distortion of channel estimation in both bands of the frequency domain. Lastly, FIG. 3 is a graph illustrating performance degradation due to the setting of GB in the conventional scheme with an uncoded bit error rate (BER). 3 is a graph illustrating an example of an OFDM system to which a DFT-based channel estimation scheme is applied, as compared with the conventional scheme. As shown in FIG. It can be seen that it shows deterioration.

Therefore, there is a need for a method for compensating CIR estimation error in an OFDM system in a multipath fading environment.

It is an object of the present invention to provide an apparatus and method for compensating for an estimation error of a CIR in an OFDM system.

Another object of the present invention is to compensate for the estimation error of the CIR before selecting the effective CIR in the DFT-based channel estimation scheme in an OFDM system, thereby effectively eliminating the effects of noise present in the initial channel estimate and the spectral leakage effect due to the GB setting. In addition, to provide an apparatus and method for overcoming performance degradation.

Another object of the present invention is to calculate a CIR estimation error function in advance in consideration of the size of the GB and the placement of pilot symbols in an OFDM system, and to compensate for the estimation error of the CIR using the calculated CIR estimation error function and In providing a method.

According to an embodiment of the present invention, a channel estimating apparatus of a receiver includes an initial channel estimator for estimating an initial channel, and an inverse fast fourier transform (IFFT) operation on the estimated initial channel value to perform a CIR. Using an IFFT operator for estimating a channel impulse response, a maximum power path detector for detecting a path having the maximum power with respect to the estimated CIR, a CIR value of the detected maximum power path, and a CIR estimation error function, And a CIR error compensator for compensating CIR estimation errors for all paths.

In accordance with an embodiment of the present invention, a channel estimation method of a communication system includes estimating an initial channel and performing an inverse fast fourier transform (IFFT) operation on the estimated initial channel value. Channel impulse response), detecting the path having the maximum power with respect to the estimated CIR, and using the CIR value and the CIR estimation error function of the detected maximum power path, the CIR for all paths. Comprising a process for compensating the estimation error.

The present invention compensates for the estimation error of the CIR before selecting the effective CIR in the DFT-based channel estimation scheme in an OFDM system, thereby effectively removing the influence of noise present in the initial channel estimate and the spectral leakage effect due to the GB setting. Compensating for the degradation of the performance due to the GB setting provides excellent channel estimation performance, which has the advantage of improving the performance of the receiver. In addition, since there is no need for signal-to-noise ratio (SNR) or advance information about a channel required in a minimum mean-square error (MMSE) method, there is an advantage of convenience.

Hereinafter, exemplary embodiments of 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 a detailed description of a related known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted.

Hereinafter, an apparatus and method for compensating for an estimation error of a CIR in an OFDM system will be described. In particular, the present invention relates to an apparatus and method for compensating for the estimation error of a CIR before selecting an effective CIR in a DFT based channel estimation scheme, which may be adaptively considered depending on the state of the channel. In the present invention, a CIR estimation error function is used to compensate for the estimation error of the CIR, which is calculated in advance in consideration of the size of GB and the placement of pilot symbols.

4 illustrates a receiver structure using a channel estimation method based on DFT in an OFDM system according to an embodiment of the present invention.

As shown, the receiver structure includes a guard interval remover 401, a first FFT operator 402, an initial channel estimator 403, an IFFT operator 404, a CIR estimation error compensator 405. ), An effective CIR selector 406, a second FFT operator 407, an equalizer 408, a guard band canceler 409, a pilot canceler 410, data symbol demapping And a data symbol demapper 411.

Referring to FIG. 4, the guard interval remover 401 removes the guard interval from the baseband signal received through the antenna through the multipath fading channel and outputs the guard interval.

The first FFT operator 402 performs an FFT operation on the signal from which the guard interval is removed, outputs a signal in a frequency domain to the equalizer 408, and outputs a pilot symbol among the signals in the frequency domain to the initial channel estimator. Output as (403). Here, the signal in the frequency domain is represented by Equation 1 below. Here, the signal in the frequency domain is set with GB at both ends of the band.

인 AWGN(Additive White Gaussian Noise)를 나타낸다. 상기

는 GB의 크기를 나타내고, 상기 N은 전 체 부반송파의 개수를 나타낸다. Here, Y (m, k) represents a signal for the kth subcarrier of the mth symbol interval received through the multipath fading channel, and X (m, k) represents the kth mth symbol interval generated by the transmitter. Represents an OFDM symbol for a subcarrier. H (m, k) represents the frequency response of the multipath fading channel for the kth subcarrier in the mth symbol interval, and W (m, k) has an average of zero and a variance.

AWGN (Additive White Gaussian Noise). remind

Represents the size of GB, and N represents the total number of subcarriers.

Here, the frequency response H (m, k) and the time response h (m, n) of the multipath fading channel are represented by Equation 2 below.

는 i번째 채널 경로의 지연 시간을 나타내며, 상기

는 복소 채널 이득을 나타낸다. 여기서, 상기

는 크로네커 델타 함수로서, [ ]안의 값이 0일 때만 1의 값을 갖는 함수이다. Where H (m, k) represents the frequency response of the multipath fading channel for the kth subcarrier in the mth symbol interval, and h (m, n) represents the multipath fading channel for the nth sample in the mth symbol interval. Represents the time response. L represents the number of delay paths of the channel,

Denotes the delay time of the i-th channel path, and

Denotes the complex channel gain. Where

Is a Kronecker delta function that has a value of 1 only when the value in [] is 0.

The initial channel estimator 403 estimates an initial channel using pilot symbols in the signal of the frequency domain and outputs an initial channel value. Here, as a channel estimation algorithm using pilot symbols, there is a LS (Least Square) using pilot symbols arranged at equal intervals, and when the channel is estimated by applying the LS, the initial channel value is represented by Equation 3 below. Appears as

는 주파수 영역에서 파일럿 심볼이 할당된 부반송파 집합을 나타낸다. Where

Denotes a subcarrier set to which pilot symbols are allocated in the frequency domain.

)하고, 추정된 CIR을 출력한다. The IFFT operator 404 estimates CIR by performing an IFFT operation on the initial channel value.

) And output the estimated CIR.

The CIR estimation error compensator 405 is configured to include a maximum path path detector 420 and a CIR error compensation unit 430. Compensate for the error for the estimated CIR that occurs due to the setting of. Here, the CIR estimation error function is a function calculated in advance in consideration of the size of GB and the arrangement of pilot symbols, and is represented by Equation 4 below.

That is, the CIR estimation error function may be calculated by IFFT operation on an OFDM symbol in which only pilot symbols are allocated to available subcarriers. 6 shows an example of the CIR estimation error function in the time domain. Here, R {} represents a function that takes only real values in {}.

를 검출하고, 상기 CIR 추정 오류 보상기(405)의 CIR 오류 보상부(430)는 상기 검출된 최대 파워 위치의 CIR 값과 CIR 추정 오류 함수를 이용하여, 하기 <수학식 5>와 같이, 모든 경로에 대하여 CIR 추정 오류를 보상한다.Here, the maximum power path detector 420 of the CIR estimation error compensator 405 is the position of the path having the maximum power with respect to the estimated CIR.

The CIR error compensator 430 of the CIR estimation error compensator 405 uses the CIR value and the CIR estimation error function of the detected maximum power position, as shown in Equation 5 below. Compensate for the CIR estimation error.

은 모듈로(modulo) N의 연산을 의미한다. 상기

은 0에서 제일 큰 값을 가지고, 상기 0을 기준으로 상기 0에서 멀어질수록(예를 들어, -1, 1) 작은 값을 가지는 sync 함수로 나타난다. 따라서, 상기 검출된 최대 파워 경로의 CIR, 즉

을 상기 sync 함수의 가장 큰 값을 가지는

으로 노멀라이즈(normalize)함으로써, GB 설정으로 인해 감소된 상기 검출된 최대 파워 경로의 파워를 보상한다. 또한, 상기 검출된 최대 파워 경로를 제외한 나머지 경로에 대해서도 상기 <수학식 5>를 이용하여

으로 노멀라이즈(normalize)함으로써, 해당 경로의 파워를 보상한다. Where

Denotes a modulo N operation. remind

Is represented by a sync function having the largest value at 0 and having a smaller value as the distance from the zero with respect to 0 (eg, -1, 1). Thus, the CIR of the detected maximum power path, i.e.

Has the largest value of the sync function

By normalizing to, compensates for the power of the detected maximum power path reduced due to GB setting. Also, for the remaining paths except the detected maximum power path, Equation 5 may be used.

By normalizing to, we compensate for the power of that path.

를 검출한다. 이와 같이 CIR 추정 오류를 보상함으로써, 상기 유효 CIR 선택기(406)가 채널 환경에 따라 선택하는 CIR이 모두 유효한 CIR이 될 수 있다. 즉, OFDM 심볼에 보호 대역을 설정하지 않았을 경우와 같이, 상기 유효 CIR 선택기(406)가 채널 환경에 따라 선택하는 CIR이 모두 유효한 CIR이 될 수 있다.Here, the CIR estimation error compensation process is repeatedly performed, and the number of repetitions may be determined as the number of delay paths L of the channel. At this time, in the next iteration process, the position of the path having the maximum power for the other paths except for the path having the maximum power detected in the previous iteration process

Detect. By compensating for the CIR estimation error as described above, all the CIRs selected by the valid CIR selector 406 according to the channel environment may be valid CIRs. That is, as in the case where the guard band is not set in the OFDM symbol, all the CIRs selected by the valid CIR selector 406 according to the channel environment may be valid CIRs.

The valid CIR selector 406 selects and outputs a valid CIR from the error compensated CIR.

에 IFFT 연산을 수행하여 추정한 CIR은 하기 <수학식 6>과 같이 나타난다.In general, when the guard band is not set in the OFDM symbol,

The CIR estimated by performing the IFFT operation on the Equation 6 is expressed by Equation 6 below.

은

를 나타내고, 상기 h(m,n)은 하기 <수학식 7>과 같이 나타낼 수 있다.Where

silver

H (m, n) may be represented by Equation 7 below.

Here, the CIR estimated through Equations 6 and 7 can be divided into 0≤n≤L-1, which is an effective channel estimation interval, and L≤n≤N-1, which is a pure noise interval. It is expressed as Equation 8 below.

)으로 설정하고 임계값(

)을 초과하는 성분만을 선택하는 방법이 있으며, 각각의 방법에 대한 DFT 기반 채널 추정 방식의 최종 채널 추정값은 하기 <수학식 9>와 <수학식 10>과 같이 나타난다. Here, the best performance among the methods for selecting the effective CIR is to select only a valid component of the estimated CIR, that is, a multipath component of a channel, and forcibly replace the remaining components by considering them as noise. The above method is for removing the influence of noise, and in the frequency domain, the method smoothes the estimated channel value. However, this method assumes the ideal situation of accurately knowing the positional components of the multipath of the channel in the estimated CIR, so it is not practically applicable. Alternatively, the method considers the maximum delay time of the channel to force zero substitution for the components after the maximum delay time, and calculates the variance of the noise in the interval from the estimated CIR.

) And the threshold (

There is a method of selecting only the components exceeding), and the final channel estimation value of the DFT-based channel estimation method for each method is represented by Equations 9 and 10 below.

)을 설정하여 유효 CIR을 선택하는 방법의 경우, 상기 임계값의 설정에 따라서 채널 추정 성능이 상이하게 나타나는데, 유효 채널 추정 구간에서 MSE(Mean-Square Error)를 최소화하기 위해, 상기 임계값은 하기 <수학식 11>과 같이 잡음 전력(

)의 두 배로 설정한다. Where the threshold (

In the case of selecting the effective CIR by setting the C), the channel estimation performance is different according to the setting of the threshold. In order to minimize the mean-square error (MSE) in the effective channel estimation interval, the threshold is As shown in Equation 11, the noise power (

Set to twice.

는 순수 잡음 구간인 L≤n≤N-1에 대한 평균 잡음 전력을 나타낸다. Where

Denotes the average noise power for the pure noise interval L≤n≤N-1.

The second FFT operator 407 performs an FFT operation on the selected effective CIR to estimate a final channel.

The equalizer 408 compensates for the influence of the channel on the signal in the frequency domain using the estimated final channel and outputs the compensated signal. That is, the noise generated in the wireless channel is compensated for and output. In other words, the influence of the channel is compensated for by using the channel frequency response input from the second FFT operator 407 to the signal in the frequency domain input from the first FFT operator 402.

The guard band remover 409 removes and outputs a guard band from a signal whose influence of the channel is compensated.

The pilot remover 410 removes pilot symbols from the signal from which the guard band is removed and outputs only data symbols.

The data symbol demapper 411 restores information through demapping and demodulation of the data symbol, and outputs binary data.

5 is a flowchart illustrating a procedure of a method for compensating CIR estimation error in DFT-based channel estimation in a receiver of an OFDM system according to an embodiment of the present invention.

Referring to FIG. 5, in step 501, a receiver removes a guard interval from a baseband signal received through an antenna through an multipath fading channel and performs an FFT operation on a signal from which the guard interval is removed. .

In step 503, the receiver estimates an initial channel using pilot symbols among signals in the frequency domain generated by the FFT operation. In step 505, the receiver performs an IFFT operation on the estimated initial channel value to perform CIR. Estimate. Here, as a channel estimation algorithm using pilot symbols, there is a LS (Least Square) using pilot symbols arranged at equal intervals.

In step 507, the receiver detects the position of the path having the maximum power with respect to the estimated CIR, and in step 509, using the CIR value and the CIR estimation error function of the detected maximum power position, 5>, we compensate for the CIR estimation error for all paths. Here, the CIR estimation error function is a function calculated in advance in consideration of the size of GB and the arrangement of pilot symbols. As shown in Equation 4, IFFT operation is performed on an OFDM symbol in which only pilot symbols are assigned to available subcarriers. Can be calculated

Here, the CIR estimation error compensation processes of steps 507 and 509 may be repeated, and the number of repetitions may be determined by the number of delay paths of the channel. At this time, in the next iteration process, the position of the path having the maximum power is detected for the remaining paths except for the path having the maximum power detected in the previous iteration process.

In step 511, the receiver selects a valid CIR from the error-compensated CIR. Here, the best performance among the methods for selecting the effective CIR is to select only a valid component of the estimated CIR, that is, a multipath component of a channel, and forcibly replace the remaining components by considering them as noise. The above method is to remove the influence of noise, and when viewed in the frequency domain, the method smoothes the estimated channel value. However, this method assumes the ideal situation of accurately knowing the positional components of the multipath of the channel in the estimated CIR, so it is not practically applicable. Alternatively, the method considers the maximum delay time of the channel to force zero substitution for the components after the maximum delay time, calculates the variance of noise in the interval from the estimated CIR, and sets it as the threshold. And selecting only components that exceed the threshold.

In step 513, the receiver performs an FFT operation on the selected effective CIR to estimate a final channel, and then, in step 515, the receiver uses the estimated last channel to generate a signal for the frequency domain signal generated in step 501. Compensate for the effects of the channel. That is, it compensates for various noises generated in the wireless channel.

The receiver then terminates the algorithm according to the invention.

7 is a diagram illustrating a CIR estimation error compensation result of the present invention.

Referring to FIG. 7, it can be seen that the CIR estimation error compensation proposed by the present invention is excellent in compensating the spectrum leakage phenomenon due to the presence of GB in the OFDM communication system.

Simulation for the performance verification of the proposed method was performed by recording statistical performance values through a sufficient number of iterations in a randomly varying multipath fading channel environment, and the results are shown in FIGS. 10 to 12. same. Here, the channel estimation method set as a comparison object for performance evaluation is as follows.

① Linear Interpolation (first-order linear interpolation technique among conventional Lagrange interpolation techniques)

② Conventional Method (conventional DFT-based channel estimation technique)

③ Proposed Method (DFT-based channel estimation method including CIR estimation error compensation method of the present invention)

FIG. 8 is a graph illustrating a comparison between MSE performance of each subcarrier in the frequency domain and the scheme proposed by the present invention.

이하의 매우 우수한 성능을 나타냄을 확인할 수 있다. Referring to FIG. 8, the same parameters as in the experimental environment of FIG. 2 were used, and in the case of the linear interpolation method, the MSE component at the pilot position was not indicated. The conventional DFT-based channel estimation algorithm shows better performance than linear interpolation except for the degradation of MSE performance at both ends of the frequency band. The proposed method improves the MSE performance degradation of both ends of the frequency band by effectively eliminating the channel estimation error of the CIR, which was the cause of the performance deterioration of the conventional DFT-based channel estimation algorithm. Uniformly across the band

It can be seen that the following very excellent performance is shown.

FIG. 9 is a graph illustrating comparison of Uncoded BER performance of the conventional technique and the scheme proposed by the present invention according to change of Eb / No.

에서

사이의 성능을 목표한다. 기존의 DFT기반의 채널 추정 알고리즘의 경우 GB설정으로 인한 성능의 열화로 비교 대상의 선형 보간법보다 복잡도가 높지만 거의 유사한 성능을 나타내며, 약 Eb/No=14dB에서 오히려 선형 보간법보다 성능이 열화되는 특징이 있다. 또한, 60km/h의 이동체 속도와 Eb/No=20dB 이하의 실험 환경에서 Uncoded BER 성능이

이하로 내려가지 않는 것을 확인할 수 있다. 본 발명의 DFT기반의 채널 추정 알고리즘을 적용하였을 경우 약 Eb/No=14.5dB에서

이하의 성능을 나타내며 Eb/No=20dB에서 약

을 나타낸다. 이러한 성능은 선형 보간법과 기존의 DFT기반의 채널 추정 알고리즘보다 전체 Eb/No구간에서 월등히 우수하며 IDEAL BER 성능과 매우 근접한 성능을 나나냄을 확인할 수 있다. Referring to FIG. 9, the same parameters as those of the experimental environment of FIG. 3 are used, and the IDEAL BER performance shown in FIG. 9 shows Uncoded BER performance when ideally compensating for channel influence. This performance represents the best performance in the multipath fading channel used. In addition, with Uncoded BER performance

in

Aim for performance between. In the conventional DFT-based channel estimation algorithm, due to the performance deterioration due to the GB setting, the complexity is higher than that of the linear interpolation method compared to the comparison target, but shows almost the same performance, and the performance is worse than the linear interpolation method at about Eb / No = 14dB. have. In addition, Uncoded BER performance is improved at 60 km / h of mobile speed and under experimental conditions of Eb / No = 20dB.

You can see that it does not go down. When applying the DFT-based channel estimation algorithm of the present invention at about Eb / No = 14.5dB

It exhibits the following performances and approx. At Eb / No = 20dB

Indicates. This performance is superior to the linear interpolation method and the existing DFT-based channel estimation algorithm in the entire Eb / No interval and shows that the performance is very close to the IDEAL BER performance.

10 is a graph illustrating a comparison of MSE performance of the conventional technique and the scheme proposed by the present invention according to the change of Eb / No.

의 MSE성능을 나타낸다.Referring to FIG. 10, the conventional DFT-based channel estimation algorithm exhibits similar performance to the linear interpolation method similar to the Uncoded BER performance, and exhibits a performance degradation compared to the linear interpolation method at Eb / No = 12dB. When the DFT-based channel estimation algorithm of the present invention is applied, it is much better in the entire Eb / No section and is weak at Eb / No = 20dB.

MSE performance is shown.

8 to 10, the DFT-based channel estimation algorithm, which adds the CIR estimation error compensation scheme of the present invention, has a Doppler according to the frequency selective fading effect of the multipath fading channel and the increase of the moving velocity. It can be seen that it can show better performance than the existing method even in a severe environment. Especially, it is considered as a technique that can effectively compensate for the effects of spectrum leakage in OFDM communication system that sets GB.

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 determined not only by the scope of the following claims, but also by the equivalents of the claims.

1 is a diagram illustrating an example of an error of estimation of a CIR due to the setting of GB in an OFDM system;

2 is an exemplary diagram illustrating MSE performance for each subcarrier of the conventional scheme in the frequency domain;

FIG. 3 is a graph illustrating performance degradation due to the setting of GB in the conventional scheme with an Uncoded Bit Error Rate (BER). FIG.

4 is a diagram illustrating a receiver structure using a channel estimation method based on DFT in an OFDM system according to an embodiment of the present invention;

5 is a flowchart illustrating a procedure of a method for compensating CIR estimation error in DFT based channel estimation in a receiver of an OFDM system according to an embodiment of the present invention;

6 illustrates an example of a CIR estimation error function of the present invention in a time domain;

7 is a diagram illustrating a CIR estimation error compensation result of the present invention;

8 is a graph illustrating a comparison of MSE performance for each subcarrier in the scheme proposed in the prior art and the present invention in the frequency domain;

9 is a graph illustrating a comparison of Uncoded BER performance of the conventional technique and the scheme proposed by the present invention according to the change of Eb / No, and

10 is a graph illustrating a comparison of MSE performance of the conventional technique and the scheme proposed by the present invention according to the change of Eb / No.

Claims (20)

In the channel estimation apparatus of the receiver, An initial channel estimator for estimating an initial channel, An IFFT calculator for estimating a channel impulse response (CIR) by performing an inverse fast fourier transform (IFFT) operation on the estimated initial channel value; A maximum power path detector for detecting a path having the maximum power with respect to the estimated CIR; And a CIR error compensator for compensating CIR estimation errors for all paths by using the CIR value and the CIR estimation error function of the detected maximum power path. The method of claim 1, Further comprising a first FFT operator for performing a fast fourier transform (FFT) operation on the received signal to generate a signal in the frequency domain, And the initial channel estimator estimates an initial channel using pilot symbols among the signals in the frequency domain. The method of claim 1, And the initial channel is estimated using Equation 12 below.

Here, Y (m, k) represents a signal for the kth subcarrier of the mth symbol interval received through the multipath fading channel, and X (m, k) represents the kth mth symbol interval generated by the transmitter. Represents an OFDM symbol for a subcarrier. H (m, k) represents the frequency response of the multipath fading channel for the kth subcarrier in the mth symbol interval, and W (m, k) has an average of zero and a variance.

AWGN (Additive White Gaussian Noise). remind

Denotes a set of subcarriers to which a pilot symbol is allocated in the frequency domain.

Denotes the initial channel for the kth subcarrier in the mth symbol interval. The method of claim 1, The CIR estimation error function is a function calculated in consideration of the guard band size and the placement of pilot symbols. The method of claim 4, wherein The CIR estimation error function is a function calculated using Equation 13 below.

Where R {} represents a function that takes only real values in {},

Denotes a subcarrier set to which pilot symbols are allocated in the frequency domain. K denotes a subcarrier number and n denotes a sample number. The method of claim 1, wherein the maximum power path detector, The path having the maximum power is detected as many times as the number of repetitions, and in this case, the position of the path having the maximum power is detected for the remaining paths except for the path having the maximum power detected in the previous repetition process. Device. The method of claim 6, Wherein the number of repetitions is equal to the number of delay paths of a channel. The method of claim 1, And the CIR error compensator compensates for CIR estimation errors for all paths using Equation (14).

Where

Denotes the CIR for the n th sample of the m th symbol interval,

Represents the CIR estimation error function,

Represents the path with the detected maximum power. The method of claim 1, A valid CIR selector for selecting a valid CIR from the error compensated CIR; A second FFT operator for performing an FFT operation on the selected effective CIR to estimate a final channel; And an equalizer for compensating for the influence of the channel in the received signal using the estimated final channel. The method of claim 9, The effective CIR selector selects only the multipath components of the channel in the estimated CIR and replaces them with zero by considering them as noise for the remaining components, considering the maximum delay time of the channel, and selecting the component after the maximum delay. For example, selecting an effective CIR using one of a method of forcing it to zero, one of calculating a variance of noise in an estimated CIR, setting it as a threshold, and selecting only components above the threshold Characterized in that the device. In the channel estimation method of the communication system, Estimating the initial channel, Estimating a channel impulse response (CIR) by performing an inverse fast fourier transform (IFFT) operation on the estimated initial channel value; Detecting a path having the maximum power with respect to the estimated CIR; And compensating for CIR estimation errors for all paths by using the CIR value and the CIR estimation error function of the detected maximum power path. The method of claim 11, wherein estimating the initial channel comprises: Generating a signal in a frequency domain by performing a fast fourier transform (FFT) operation on the received signal; Estimating an initial channel using a pilot symbol among the signals in the frequency domain. The method of claim 11, The initial channel is estimated using the following equation (15).

Here, Y (m, k) represents a signal for the kth subcarrier of the mth symbol interval received through the multipath fading channel, and X (m, k) represents the kth mth symbol interval generated by the transmitter. Represents an OFDM symbol for a subcarrier. H (m, k) represents the frequency response of the multipath fading channel for the kth subcarrier in the mth symbol interval, and W (m, k) has an average of zero and a variance.

AWGN (Additive White Gaussian Noise). remind

Denotes a set of subcarriers to which a pilot symbol is allocated in the frequency domain.

Denotes the initial channel for the kth subcarrier in the mth symbol interval. The method of claim 11, The CIR estimation error function is a function calculated in consideration of the guard band size and the placement of pilot symbols. The method of claim 14, The CIR estimation error function is a function calculated using Equation 16 below.

Where R {} represents a function that takes only real values in {},

Denotes a subcarrier set to which pilot symbols are allocated in the frequency domain. K denotes a subcarrier number and n denotes a sample number. The method of claim 11, wherein the maximum power path detection process, And performing a repetition number of times, wherein in the next repetition process, the position of the path having the maximum power is detected for the remaining paths except for the path having the maximum power detected in the previous repetition process. The method of claim 16, Wherein the number of repetitions is equal to the number of delay paths of a channel. The method of claim 11, CIR estimation error for all the paths, characterized in that to compensate for using the equation (17).

Where

Denotes the CIR for the n th sample of the m th symbol interval,

Represents the CIR estimation error function,

Represents the path with the detected maximum power. The method of claim 11, Selecting a valid CIR from the error compensated CIR; Estimating a final channel by performing an FFT operation on the selected effective CIR; And compensating for the influence of the channel in the received signal using the estimated final channel. The method of claim 19, The effective CIR selects only the multipath components of the channel in the estimated CIR and considers noise as the remaining components, and replaces them with zeros, and considers the components after the maximum delay in consideration of the maximum delay time of the channel. A method of forcibly substituting to zero, the method comprising selecting one of the methods of calculating a variance of noise in a predetermined period from an estimated CIR, setting it as a threshold, and selecting only a component that exceeds the threshold .

Images