KR20060038070A - Channel estimation method using linear prediction in an ofdm communication system with virtual subcarriers, and device thereof - Google Patents

Abstract

The present invention relates to a channel estimation method of an OFDM communication system having a virtual subcarrier, comprising the steps of: (a) estimating a frequency response of a pilot channel from a received OFDM signal by LS; and (b) from the frequency response of the pilot channel. Linearly predicting the frequency response of the virtual subcarrier channel, (c) constructing an overall channel response including a pilot channel and a virtual carrier channel from the frequency response calculated in steps (a) and (b); (d) performing an IFFT operation on the entire channel response, (e) performing amplitude adjustment to reduce noise effects on the entire channel response on which the IFFT operation is performed, and (f) performing the amplitude Converting the adjusted total channel response into a channel estimate in the frequency domain by an FFT operation. According to the present invention, the channel estimation performance can be improved by effectively improving the time-domain spreading phenomenon generated by the virtual subcarriers.

OFDM, virtual subcarrier, channel estimation, linear prediction, DFT, LS, MMSE,

Description

CHANNEL ESTIMATION METHOD USING LINEAR PREDICTION IN AN OFDM COMMUNICATION SYSTEM WITH VIRTUAL SUBCARRIERS, AND DEVICE THEREOF}

1 is a block diagram of an apparatus for conventional DFT based channel estimation in an OFDM system.

FIG. 2 is a graph showing least square error (MSE) performance of the DFT-based channel estimation scheme of FIG.

FIG. 3 is a diagram illustrating a spread phenomenon of a channel impulse response due to a virtual subcarrier in the DFT-based channel estimation technique of FIG. 1. FIG.

4 is a diagram illustrating a subcarrier configuration of an OFDM symbol.

5 is a block diagram of an apparatus for performing channel estimation in accordance with a preferred embodiment of the present invention.

6 is a conceptual diagram of forward and reverse channel prediction techniques of virtual subcarrier positions using LS channel estimates of pilot subcarrier positions, according to an embodiment of the present invention.

7 is a comparison diagram of MSE performance of an optimal MMSE channel estimation technique, a conventional DFT based channel estimation technique, and a channel estimation technique according to a preferred embodiment of the present invention in a given system model.

8 is a comparison of BER performance of an optimal MMSE channel estimation technique, a conventional DFT based channel estimation technique, and a channel estimation technique in accordance with a preferred embodiment of the present invention in a given system model.

The present invention relates to an OFDM communication system, and more particularly, to a DFT-based channel estimation technique using linear prediction in an OFDM system having a virtual carrier.

Orthogonal Frequency Division Multiplexing (OFDM) is a wideband modulation scheme that divides a frequency bandwidth allocated for a communication session into a plurality of narrowband frequency subbands, each subband being wireless. Frequency subcarriers, each subcarrier being mathematically orthogonal to the RF subcarriers included in each of the other subchannels. The orthogonality of the subcarriers allows their individual spectra to overlap without interference with other carriers. By dividing the frequency bandwidth into a plurality of orthogonal subbands, the OFDM scheme enables high data rates and very efficient bandwidth usage.

The OFDM method first converts the data to be transmitted into a complex symbol in the form of M-ary quadrature amplitude modulation (M-ary QAM), and performs a serial-to-parallel conversion of the complex symbol sequence, which is a sequence of complex symbols. The multi-carrier modulation method converts a plurality of parallel complex symbols through a rectangular pulseshaping and sub-carrier modulation. In the multicarrier modulation scheme, frequency intervals between subcarriers are set such that all sub-carrier modulated parallel complex symbols are orthogonal to each other.

When the M-ary QAM modulated signal is transmitted through a wireless fading channel without using the OFDM scheme, the delay spread of the channel caused by the multipath delay is greater than the symbol period of the modulated signal. If large, inter-symbol interference occurs, and correct signal restoration is impossible at the receiving end. Therefore, an equalizer must be used to compensate for random delay spread, but the implementation of the equalizer is not only very complicated but also has a disadvantage in that the transmission performance is degraded due to input noise at the receiving end.

On the other hand, using the OFDM scheme, since the symbol period of each parallel complex symbol can be much longer than the delay spread of the channel, the interference between symbols can be relatively small. In particular, by setting the guard interval longer than delay spread, there is an advantage that the interference between symbols can be completely eliminated. Of course, there is no need to implement an equalizer that compensates for random delay spread due to multipath delay. Accordingly, the OFDM scheme is very effective for data transmission over a wireless fading channel and is currently adopted as a standard transmission scheme for terrestrial digital television and audio broadcasting systems in Europe. In addition, data transmission system through wired channel such as digital subscriber loop (DSL) and powerline communication eliminates degradation of transmission performance due to multipath reflection occurring in line network environment. It is used a lot.

The transmitting end of the data transmission system using the OFDM scheme first converts the data to be transmitted into coded data through channel encoding means and M-ary QAM and PSK through a mapper. After converting into complex symbols in the form of shift keying) and DPSK (differential PSK), and converting them into multiple parallel complex symbols through serial-to-parallel conversion, each parallel complex symbol is square-shaped and subcarrier modulated and then all subcarrier modulated And modulation means for carrier-modulating the sum of the signals, and transmission-end channel matching means, including an amplifier and an antenna, for transmitting the carrier-modulated signal through wireless and wired channels. In contrast to the transmitting end, the receiving end comprises a receiving end channel matching means, a demodulation means, a channel decoding means and the like.

As the channel encoding means, a number of methods including a convolutional encoding, a block encoding, a turbo encoding, or the like, or a suitable combination thereof are used. The square wave shaping and subcarrier modulation means of a plurality of parallel complex symbols among the transmitter modulation means are implemented as an inverse fast Fourier transform (IFFT) signal processing means based on a sampling theorem. Fast Fourier transform (FFT) signal processing means is used.

At the transmitting end of the OFDM system, the encoded data is converted into complex symbols through a mapper, and the complex symbols are complex symbols adjacent by a frequency interleaver and a frequency deinterleaver of the receiving end. They are subject to fading independent of each other. Therefore, the coded data reconstructed at the receiving end is prevented from severe performance deterioration due to the loss of the burst form. However, the probability of loss of information due to fading is still very high, which results in a severe degradation of transmission performance compared to data transmission through an unfaded channel.

Meanwhile, in an OFDM scheme using a plurality of orthogonal subcarriers, each subcarrier demodulated by a receiver is represented as a product of data symbols and frequency non-selective fading (frequency response in each subcarrier). In the OFDM method using a coherent modulation scheme, the channel fading distortion of each subcarrier is estimated for data detection, and the result is used as a coefficient of a single tap equalizer to remove fading distortion from each demodulated subcarrier. . In this data detection process, a significant factor affecting the detection performance is channel estimation, and many studies have been conducted for this purpose.

The channel estimation method for OFDM is based on the channel estimation method based on Linear Minimum Mean Square Error (LMMSE) and the least square (LS) based on which criterion was used in deriving the channel estimator. It can be divided into channel estimation method. Among them, the channel estimation method based on LS is very simple to calculate, but has a disadvantage of greatly affected by noise. The channel estimation method based on the MMSE shows an excellent estimation performance compared to the LS reference method because the channel estimation method considers the influence of noise.

In addition, the channel estimation method for OFDM is basically a channel estimation method using a training symbol, a channel estimation method using a pilot tone, and a decision directed according to the type of data used for the channel estimation. ) Can be classified into channel estimation method.

Among them, the channel estimation method using the pilot tone and the MMSE criterion has the advantage of tracking the change of time-varying channel with excellent estimation performance, but it is difficult to implement when the number of subcarriers is large due to excessive computation. have. This problem can be solved using the Discrete-Time Fourier Transform (DFT) based channel estimation method which processes the channel estimation process using both the frequency domain and the time domain. However, this scheme is basically designed on the assumption that pilot tones are allocated over the entire frequency band. Therefore, the scheme is applied to an OFDM system using a virtual subcarrier that does not allocate a subcarrier to a specific frequency band. In this case, interference occurs between the paths of the multipath channel in the time domain, which greatly degrades the estimation performance. Insertion of virtual subcarriers is the simplest way to prevent inter-channel interference by limiting the out-of-band power spectrum density of a particular channel.

1 is a block diagram illustrating an apparatus for a conventional DFT-based channel estimation in an OFDM system, in which N represents the total number of subcarriers, N _v represents the number of subcarriers allocated as virtual subcarriers, and N _p is The number of pilot tones is shown. In the conventional scheme, it is assumed that N _v is zero.

의 각 파일럿 톤으로부터 부채널 응답

을 구한다. 이는 각 파일럿 톤 위 치에서 복조된 수신 데이터 심볼

로부터 파일럿 톤에 할당된 데이터 심볼을 제거해줌으로써 이루어진다. 이를 주파수영역 신호에서의 LS 채널추정방식이라고 하며, LS 채널 추정기(110)에 의하여 수행된다. 모든 파일럿 톤 위치에서 LS 추정 기법을 사용하여 파일럿 부채널을 추정한 후 나머지 데이터 부반송파 위치에 0을 삽입하고, 그 결과를 크기 N의 IFFT 연산기(또는 IDFT 연산기)(120)에 의하여 시간영역으로 변환한다. 이 과정에서 얻게 되는 신호는 파일럿 톤의 간격

이 샘플링 이론을 만족시켰다면 다중경로 채널의 순시 임펄스 응답(CIR; Channel Impulse Response)이 된다. 이 임펄스 응답은 M 번 반복되는 구조를 가지며, 반복 구간은 N/M이 된다. 다음으로, MMSE 가중기(130)에 의하여, 첫 번째 N/M구간의 신호에 적절한 계수를 곱하여 원래 채널 임펄스 응답과 유사한 값을 갖도록 조정한다. 마지막으로, 크기 N의 FFT 연산기(140)에 의하여 전술한 임펄스 응답을 주파수영역으로 변환시키면 채널의 주파수 영역 응답

을 얻게 된다. The channel estimator shown in FIG. 1 is first demodulated by an OFDM symbol of size N

Subchannel response from each pilot tone

Obtain This is the demodulated received data symbol at each pilot tone position.

By removing the data symbols assigned to the pilot tones from This is called an LS channel estimation method in a frequency domain signal and is performed by the LS channel estimator 110. After estimating the pilot subchannels using the LS estimation technique in all pilot tone positions, insert zeros into the remaining data subcarrier positions, and convert the result into a time domain by an IFFT operator (or IDFT operator) 120 having size N. do. The signal obtained in this process is the pilot tone spacing.

If this sampling theory is satisfied, it becomes instantaneous impulse response (CIR) of a multipath channel. This impulse response has a structure that is repeated M times, and the repetition interval is N / M. Next, the MMSE weighting unit 130 adjusts to have a value similar to the original channel impulse response by multiplying the signal of the first N / M interval by an appropriate coefficient. Finally, if the above-mentioned impulse response is converted into frequency domain by FFT operator 140 of size N , the frequency domain response of the channel

You get

However, in the frequency domain, the channel frequency response should be obtained at subcarrier positions between the pilot tones as well as the pilot tones. In other words, an interpolation process in the frequency domain is required. This is done by inserting NN / M zeros at the back of the impulse response with the N / M sample obtained earlier, then performing a DFT or FFT of size N on the result.

FIG. 2 illustrates the performance of Mean Square Error (MSE) of the DFT-based channel estimation scheme of FIG. 1 in an OFDM system using a virtual carrier. As shown, the basic DFT-based channel estimation method has a pilot tone interval smaller than the sampling interval considering the coherence bandwidth of the multipath channel, and N / M is equal to the maximum delay spread of the multipath channel. If it is small, it shows basically similar estimation performance as the frequency domain LMMSE channel estimation method. However, when this method is applied when the virtual carrier is inserted, it can be seen that the estimation performance is drastically degraded according to the number of virtual subcarriers in the sample interval channel. This is because the time domain signal output through the IFFT of the DFT-based channel estimation technique is dispersive due to the virtual subcarriers as shown in FIG. It will be degraded.

The root cause of this phenomenon occurs in the process of performing the DFT or FFT. In other words, the DFT of the frequency domain channel response is equivalent to the DFT of the signal obtained by multiplying the frequency domain channel response by a rectangular window. Therefore, the DFT result is a convolution of the DFT result of the signal and the DFT result of the rectangular window, that is, the sinc function. If the length of the rectangular window is equal to the bandwidth of the channel (without using virtual subcarriers), then in the discrete time domain, the sinc function becomes a delta function and the DFT result is the channel's impulse response itself. However, when using a virtual subcarrier, the length of the rectangular window is shorter than the bandwidth of the channel, which causes the sinc function to zero-cross at a time different from the discrete time sampling position. Therefore, since the time-domain signal of the rectangular window has a non-zero value at the sampling position, the DFT result is different from the impulse response of the original channel.

As described above, since the virtual subcarrier technology is basically used in most OFDM systems, it must be considered to obtain excellent estimation performance in the DFT-based channel estimation method using pilot tones.

Accordingly, the present invention proposes a new channel estimation method suitable for an OFDM system using a virtual subcarrier, and particularly aims at reducing distortion caused by a virtual subcarrier in a DFT-based channel estimation technique using a pilot tone.

In order to achieve the above object, according to the first aspect of the present invention, there is provided a channel estimation method of an OFDM communication system having a virtual subcarrier, (a) estimating the frequency response of the pilot channel from the received OFDM signal by the LS scheme (B) linearly predicting the frequency response of the virtual subcarrier channel from the frequency response of the pilot channel, and (c) the pilot channel and virtual carrier from the frequency response calculated in steps (a) and (b) Constructing an overall channel response comprising a channel, (d) performing an IFFT operation on the overall channel response, and (e) reducing a noise effect on the overall channel response on which the IFFT operation is performed; Performing amplitude adjustment, and (f) converting the amplitude-adjusted total channel response into a channel estimate in the frequency domain by FFT operation. The.

In this case, the virtual subcarriers are located to the left and right of the pass band of the OFDM signal, and step (b) includes (g) forward linear prediction of the frequency response of the virtual subcarrier channel located to the right of the pass band. And (h) linearly predicting a frequency response of the virtual subcarrier channel located to the left of the passband.

According to a second aspect of the present invention, there is provided a channel estimating apparatus of an OFDM communication system having a virtual subcarrier, comprising LS channel estimating means for estimating a frequency response of a pilot channel from a received OFDM signal by an LS scheme; Channel prediction means for linearly predicting the frequency response of the virtual subcarrier channel from the frequency response, and an IFFT operation on the entire channel response including the pilot channel and the virtual carrier channel from the frequency response calculated from the LS channel estimation means and the channel prediction means. IFFT calculation means for performing the following, amplitude adjustment means for performing amplitude adjustment for reducing the noise effect on the entire channel response of the time domain output from the IFFT calculation means, and the entire time domain output from the amplitude adjustment means. Channel response to the frequency domain Means for calculating FFTs to convert to channel estimates.

In this case, the virtual subcarriers are located to the left and right of the pass band of the OFDM signal, respectively, the channel predicting means is a linear forward prediction means for linearly predicting the frequency response of the virtual subcarrier channel located to the right of the pass band; Reverse linear prediction means for reverse linear prediction of the frequency response of the virtual subcarrier channel located to the left of the pass band.

Hereinafter, exemplary embodiments of the present invention will be described with reference to the accompanying drawings.

개의 전체 부반송파로 이루어진 OFDM 시스템에서

개의 부반송파 중

개는 실제 데이터와 파일럿 데이터 전송을 위해 사용되는 유효 부반송파이며, 나머지

개의 부반송파는 보호대역(guard band)으로 사용되는 가상 부반송파가 된다. 따라서, 전체 부반송파는

로 나타낼 수 있다. 한편,

는 짝수,

는 홀수라고 가정한다. 4 shows a subcarrier configuration of an OFDM symbol,

In an OFDM system consisting of four total subcarriers

Of subcarriers

Is the effective subcarrier used for the actual data and pilot data transmission, and

Subcarriers are virtual subcarriers used as guard bands. Therefore, the entire subcarrier is

It can be represented by. Meanwhile,

Is even,

Is assumed to be odd.

개가, 오른쪽에

개가 놓인다. 여기서,

은 실수

보다 작거나 같은 가장 가까운 정수를 나타낸다. 유효 부반송파는 M-ary 심볼

에 의해 변조된다. 여기서,

는 부반송파 번호를 나타낸다. 유효 부반송파에 데이터 할당이 끝난 후 IFFT에 의해 기저대역 변조가 이루어지며 그 결과는 다음과 같다.As shown in Fig. 4, the virtual subcarriers are on the left side in the frequency domain of the passband.

Dog on the right

The dog is placed. here,

Is a mistake

Represents the nearest integer that is less than or equal to. Effective subcarriers are M -ary symbols

Is modulated by here,

Indicates a subcarrier number. After allocating data to the effective subcarriers, baseband modulation is performed by IFFT. The result is as follows.

은 이산시간 샘플시간을 나타낸다. 수학식 1의 신호에 보호구간으로 CP(cyclic prefix)가 더해지고 아날로그 신호로 변환된 후, RF 전처리부(RF front-end)와 안테나를 거쳐 무선채널로 전송된다. 수신단에서 수신된 시간영역 이산시간 신호는 다음과 같다.here,

Represents the discrete time sample time. A CP (cyclic prefix) is added to the signal of Equation 1 and converted into an analog signal, and then transmitted to a wireless channel through an RF front-end and an antenna. The time domain discrete time signal received at the receiver is as follows.

은 환형 컨벌루션,

은 평균이 0이고 분산이

인 가우시안 잡음,

는 보호구간의 샘플 수를 각각 나타낸다. 또한,

은

개의 경로를 갖는 다중경로 채널의 탭 지연 라인(TDL; tapped delay line) 모델을 나타낸다.

은

개의 탭을 가지며, 이 중

개의 탭이 0이 아닌 복소수 계수를 갖는다. 0이 아닌

개의 탭 계수는 각각 크기가 레일레이 분포를 가지며, 위상은 균일 분포를 갖는다. 수신단에서의 심볼 동기와 주파수 동기가 완벽하게 이루어졌다고 가정하면, 보호구간을 제거한 나머지 수신 샘플은 FFT에 의해 주파수 영역으로 변환되며 그 결과는 다음과 같이 표현된다.here,

Silver annular convolution,

Has a mean of 0 and a variance

Gaussian noise,

Represents the number of samples of the protection interval. Also,

silver

A tapped delay line (TDL) model of a multipath channel having two paths is shown.

silver

Tabs, of which

Taps have non-zero complex coefficients. Nonzero

The tapped coefficients each have a Rayleigh distribution of magnitude and a phase distribution of uniformity. Assuming that symbol synchronization and frequency synchronization at the receiving end are completed, the remaining received samples with the guard interval removed are transformed into the frequency domain by the FFT, and the result is expressed as follows.

은 잡음의 주파수 영역 표현이고,

은

번째 부반송파에서의 부채널 응답으로 다음과 같이 표현된다.here,

Is the frequency domain representation of the noise,

silver

The subchannel response in the first subcarrier is expressed as follows.

(단,

는 짝수)개의 파일 럿 부반송파가

개의 유효 부반송파에 아래 수식과 같이 삽입된다.To estimate the multipath fading channel,

(only,

Is even) pilot subcarriers

The two effective subcarriers are inserted as in the following formula.

는 두 파일럿 부반송파 사이의 간격을 나타내며 정수이다.

은 삽입되는 파일럿의 수를 나타낸다. 그러나 앞서 언급하였듯이 실제로 사용되는 개수는

이다. here,

Is an integer between two pilot subcarriers.

Represents the number of pilots to be inserted. However, as mentioned earlier, the actual number used

to be.

5 shows a channel estimation technique according to a preferred embodiment of the present invention.

First, by using the LS channel estimator 500, the frequency response of the channel at the pilot position may be estimated by the LS scheme, and may be expressed by Equation 6.

A vector consisting of LS estimates obtained through Equation 6 is defined as follows.

와 동 일하다. This vector is used to linearly predict the frequency response of the channel at the location of the virtual subcarrier. The interval of the predicted channel frequency response is the pilot subcarrier interval.

Do the same with

Regarding the functions of the forward linear predictor 510 and the reverse linear predictor 520, FIG. 6 shows a conceptual diagram of the forward and reverse channel prediction techniques of the virtual subcarrier positions using LS channel estimates of the pilot subcarrier positions.

First, the forward linear prediction by the forward linear predictor 510 of FIG. 5 is the frequency response of the virtual carrier channel located on the right side (forward) of the pass band according to the algorithm expressed step by step in Equations 8 to 10 below. Can be obtained.

는 래그(lag)로 선형예측에 사용되는 과거의 데이터 수를 나타낸다.

은

로 정의되며, 짝수라고 가정한다. 순방향 선형예측 계수

은 다음의 식을 통해 구할 수 있다.here,

Denotes the number of past data used for linear prediction in lag.

silver

It is defined as, and is assumed to be even. Forward Linear Prediction Coefficient

Can be obtained from the following equation.

는 순방향 선형예측계수 벡터,

는 채널의 자기상관 행렬이다. 또한,

,

로 정의된다.

은 예측 에러 분산이며, 예측 에러는

이다.

는 행렬의 트랜스포즈(transpose) 연산자를 나타내고,

는 복소 켤레 연산자를 나타낸다. 수학식 11을 이용하여 선형예측계수 벡터를 구하는 방법 중에서, 본 발명의 바람직한 실시예에서는 가장 간단한 역행렬을 이용한 방법 즉,

을 사용하였다. 이렇게 구한 선형예측계수 벡터는 채널추정과정에서 항상 동일하게 사용된다. 순방향 선형예측에 의해 구한 채널 응답을 다음의 벡터로 나타낸다.here,

Is the forward linear predictive coefficient vector,

Is the autocorrelation matrix of the channel. Also,

,

Is defined as

Is the prediction error variance, and the prediction error is

to be.

Represents the transpose operator of the matrix,

Denotes a complex conjugate operator. Among the methods for obtaining the linear predictive coefficient vector using Equation 11, in the preferred embodiment of the present invention, the method using the simplest inverse matrix,

Was used. The linear predictive coefficient vectors thus obtained are always used the same in the channel estimation process. The channel response obtained by forward linear prediction is represented by the following vector.

Next, the backward linear prediction by the backward linear predictor 520 of FIG. 5 obtains the frequency response of the virtual carrier channel located on the left side (reverse direction) of the pass band through the stepwise algorithms of Equations 13 to 15 below. Can be.

이다.

은 역방향 선형예측계수이며 다음의 수학식 16으로부터 구할 수 있다.here,

to be.

Is a reverse linear predictive coefficient and can be obtained from Equation 16 below.

는 역방향 선형예측 계수벡터이다. 역방향 선형예측에 의해 구한 채널 응답은 다음의 벡터와 같다.here,

Is the reverse linear predictive coefficient vector. The channel response obtained by the backward linear prediction is as follows.

Referring back to FIG. 5, since the subchannel response at the pilot tone location is estimated in the entire frequency band including the virtual carrier interval, the time-domain impulse response is obtained using this result. To do this, first reconstruct the entire channel response as follows:

의 IFFT 연산기(또는 IDFT 연산기)(530)에서, 수학식 18에 IFFT를 취하면 다음의 결과를 얻을 수 있다.size

In the IFFT operator (or IDFT operator) 530, the IFFT in Equation 18 can be obtained.

At this time, in order to reduce the influence of noise from the impulse response, the MMSE weighting unit 540 multiplies the MMSE coefficients as shown in Equation 20.

은 MMSE 계수이다.

는 이산시간 임펄스 응답의

번째 샘플 시간에서의 평균 전력을 나타낸다.

는

이고,

은 M-ary QAM 심볼이다. here,

Is the MMSE coefficient.

Of the discrete time impulse response

Average power at the first sample time.

Is

ego,

Is an M-ary QAM symbol.

의 FFT를 통해 위 결과를 주파수 영역으로 되돌리며, 그 결과는 수학식 21과 같이 표현된다.Subsequently, the size of the FFT operator 550

The above result is returned to the frequency domain through the FFT, and the result is expressed by Equation 21.

Simulation was performed to prove the superiority of the channel estimation technique according to the preferred embodiment of the present invention described above. The center frequency of the system is 2.3 GHz and the frequency band is 10 MHz. The sampling time interval is "1 / frequency band". Data subcarriers are QPSK modulated and pilot subcarriers are BPSK modulated. The remaining parameters are shown in Table 1, and the multipath fading channel (“Vehicular A” channel model) is shown in Table 2 and is assumed to be a sample interval channel.

System parameters value

: 부반송파의 개수

: Number of subcarriers 1024

: Number of effective subcarriers 865

: Number of subcarriers for data 648 : Number of pilot subcarriers 217

: Number of virtual subcarriers 80 (left), 79 (right)

= Number of guard interval samples 128

: Spacing between two adjacent pilots 4

Tab Average power (dB) Time delay (ns) Time delay (sample cycle) One 0 0 0 2 -One 310 3 3 -9 710 7 4 -10 1090 10 5 -15 1730 17 6 -20 2510 25

7 and 8 show the MSE and BER performance for each scheme as the simulation results described above. In Figures 7 and 8, "MMSE" represents the MSE performance of the optimal LMMSE channel estimation technique, "DFT" in Figure 5 MSE performance of the conventional DFT-based channel estimation technique that does not use the forward and backward linear predictor Indicates. The remaining three graphs show the MSE performance of the channel estimation method according to the preferred embodiment of the present invention, according to the number of lags (P = 108, 40, 30). The proposed method, “Perfect,” is a bit error rate (BER) performance under the assumption that the channel is fully known. Analysis of the results shows that the optimal LMMSE channel estimation technique and the proposed method show the same BER performance as when “Perfect”.

로 나타낼 수 있으며, 제안 한 방식의 복소 곱셈 수는 다음의 수학식 22와 같다.Table 3 compares the complexity of the optimal LMMSE channel estimation technique with the proposed scheme as a preferred embodiment of the present invention. For comparison, we consider the number of complex multiplications. The complex multiplication number of the optimal LMMSE channel estimation technique is

The complex multiplication number of the proposed scheme is given by Equation 22 below.

Channel estimation method Number of complex multiplications Optimal LMMSE 187,705 Conventional DFT base 7,172 Proposed scheme according to a preferred embodiment of the present invention (P = 108) 11,384 Proposed scheme according to a preferred embodiment of the present invention (P = 40) 8,732

Through Table 3, it can be seen that the complexity of the channel estimation scheme according to the preferred embodiment of the present invention is lower than the optimal MMSE channel estimation technique.

As described above, according to the present invention, the proposed channel estimation method first estimates a pilot tone in a virtual subcarrier interval, performs a DFT-based channel estimation technique using the result, and performs forward and reverse linear prediction to estimate the pilot tone. By using the algorithm, the channel estimation performance can be improved by effectively improving the distortion caused by the virtual subcarrier, that is, the time domain spreading phenomenon. Moreover, according to the present invention, since the complexity of the system is low compared to the conventional LMMSE technique, there is an advantage that it is easy to implement.

Claims (14)

A channel estimation method of an OFDM communication system having a virtual subcarrier, (a) estimating a frequency response of the pilot channel from the received OFDM signal by the LS scheme; (b) linearly predicting the frequency response of the virtual subcarrier channel from the frequency response of the pilot channel; (c) constructing an entire channel response including a pilot channel and a virtual carrier channel from the frequency response calculated in steps (a) and (b); (d) performing an IFFT operation on the entire channel response; (e) performing amplitude adjustment on the entire channel response on which the IFFT operation is performed to reduce noise influence; (f) converting the amplitude-adjusted total channel response into a channel estimate in the frequency domain by FFT operation Channel estimation method of an OFDM communication system having a virtual subcarrier comprising a. The method of claim 1, The virtual subcarriers are located at the left and right of the pass band of the OFDM signal, respectively. In step (b), (g) forward linearly predicting the frequency response of the virtual subcarrier channel located to the right of the passband; (h) backward linearly predicting the frequency response of the virtual subcarrier channel located to the left of the passband. A channel estimation method of an OFDM communication system having a virtual subcarrier. 3. The frequency response of claim 2, wherein the frequency response of the i-th pilot channel estimated in step (a) is determined.

(here,

Is the spacing between pilot subcarriers,

Is the number of subcarriers actually used, and the frequency response of the virtual subcarrier channel that is linearly predicted in step (g)

(here,

Is the number of virtual subcarriers located on the right side of the passband,

Is the total number of subcarriers,

silver

Defined as Step (g) is an algorithm for each step

(here,

Is the number of historical data used for linear prediction,

Is the vector of forward linear predictive coefficients) The frequency response of the virtual subcarrier channel by

Channel estimation method in an OFDM communication system having a virtual subcarrier. The method of claim 3, wherein the vector of the forward linear predictive coefficient (

)

(here,

Is the channel's autocorrelation matrix,

Is a prediction error

Dispersion,

) The channel estimation method of the OFDM communication system having a virtual subcarrier determined by. The method according to claim 3 to 4, In step (h), the frequency response of the virtual subcarrier channel reversely predicted is

(here,

) Step (h) is an algorithm for each step

(here,

Is a vector of inverse linear prediction coefficients) The frequency response of the virtual subcarrier channel by

Channel estimation method in an OFDM communication system having a virtual subcarrier. The method of claim 5, wherein the vector of the linear linear predictive coefficients (

)

The channel estimation method of the OFDM communication system having a virtual subcarrier determined by. 7. The overall channel response as defined in claim 6, wherein

)silver

A channel estimation method of an OFDM communication system having a virtual subcarrier. A channel estimating apparatus of an OFDM communication system having a virtual subcarrier, LS channel estimation means for estimating the frequency response of the pilot channel by the LS method from the received OFDM signal; Channel prediction means for linearly predicting the frequency response of the virtual subcarrier channel from the frequency response of the pilot channel; IFFT calculation means for performing an IFFT operation on the entire channel response including the pilot channel and the virtual carrier channel from the frequency response calculated by the LS channel estimation means and the channel prediction means; Amplitude adjusting means for performing amplitude adjustment to reduce the noise effect on the entire channel response in the time domain output from the IFFT calculating means; FFT calculation means for converting the entire channel response in the time domain output from the amplitude adjusting means into the channel estimate value in the frequency domain by FFT calculation. Channel estimation apparatus of the OFDM communication system having a virtual subcarrier comprising a. The method of claim 8, The virtual subcarriers are located at the left and right of the pass band of the OFDM signal, respectively. The channel prediction means, Forward linear prediction means for forward linear prediction of the frequency response of the virtual subcarrier channel located on the right side of the pass band; Comprising reverse linear prediction means for linearly predicting the frequency response of the virtual subcarrier channel located to the left of the pass band. Channel estimation apparatus of an OFDM communication system having a virtual subcarrier. The frequency response of the i-th pilot channel estimated by the LS channel estimating means.

(here,

Is the spacing between pilot subcarriers,

Is the number of reverse virtual subcarriers located on the left side of the passband,

Is the number of subcarriers actually used, and the frequency response of the forward virtual subcarrier channel predicted by the forward linear prediction means.

(here,

Is the total number of subcarriers,

silver

Defined as The forward linear prediction means is a step-by-step algorithm

(here,

Is the number of historical data used for linear prediction,

Is the vector of forward linear predictive coefficients) The frequency response of the forward virtual subcarrier channel by

Channel estimation apparatus of an OFDM communication system having a virtual subcarrier. 11. The method of claim 10, wherein the vector of the forward linear predictive coefficients (

)

(here,

Is the channel's autocorrelation matrix,

Is a prediction error

Dispersion,

) Channel estimation apparatus of an OFDM communication system having a virtual subcarrier as determined by. The method according to claim 10, wherein The frequency response of the reverse virtual subcarrier channel predicted by the reverse linear prediction means

(here,

) The reverse linear prediction means is a following step-by-step algorithm

(here,

Is a vector of inverse linear prediction coefficients) The frequency response of the reverse virtual subcarrier channel by

Channel estimation apparatus of an OFDM communication system having a virtual subcarrier. 13. The vector of claim 12, wherein

)

Channel estimation apparatus of an OFDM communication system having a virtual subcarrier as determined by. 14. The overall channel response of claim 13, wherein an IFFT operation is performed by said IFFT operation means.

)silver

Channel estimation apparatus of the OFDM communication system having a virtual subcarrier.

Images