KR100557877B1 - Apparatus and method for channel estimating and ofdm system for the same - Google Patents

Abstract

The present invention relates to a channel estimation apparatus and method for receiving digital terrestrial television broadcasting using an orthogonal frequency division multiplexing (OFDM) transmission method, and to an OFDM system using the same. In particular, a mobility variance is calculated from a pilot signal extracted from a received signal. The effect of noise is reduced by determining the number of symbols whose time-varying characteristics of the channel can be ignored compared to the noise variance after pre-channel compensation, and compensating the signal by linearly interpolating the sample average of the pilot signal within the determined symbol. The output SNR after channel compensation is maximized.

OFDM

Description

Apparatus and Method for Channel Estimation and Orthogonal Frequency Division Multiplexing Using Them

1 is a block diagram showing the overall configuration of a typical DVB-T receiver

2 is a diagram illustrating a transmission state of a pilot signal inserted into an active carrier in a 2k mode frame structure of the DVB-T standard.

3 is a block diagram of an apparatus for estimating a channel of an OFDM system according to the present invention;

4 is a graph showing the relationship between the mobility variance and the noise variance obtained from the Rayleigh channel model according to an embodiment of the present invention.

5 is a graph showing an output SNR according to the number of symbols in a Rayleigh channel model according to an embodiment of the present invention.

Explanation of symbols for main parts of the drawings

204: FFT unit 206: signal correction unit

301: pilot extractor 302: mobility estimation unit

302-1: noise variance calculator 302-2: mobility variance calculator

302-3: Comparator 303: Pilot signal sample average

304: linear interpolation unit

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an Orthogonal Frequency Division Multiplexing (OFDM) system using multiple carriers. In particular, a channel estimating apparatus for estimating the degree of mobility or channel change in an OFDM transmission digital TV receiver And a method and an OFDM system using the same.

In general, the OFDM transmission system is a multi-carrier transmission scheme in which data is transmitted using a plurality of carriers having orthogonality to each other, unlike a single carrier scheme in which data is transmitted using one carrier. That is, unlike the conventional serial data transmission system in which the frequency spectrum of each data symbol occupies the entire effective bandwidth, the OFDM transmission scheme transmits data in parallel by dividing the entire signal frequency band into N non-overlapping subchannels. In this case, each carrier has a very small bandwidth and is affected by the channel change. However, in the case of the entire frequency band, in the case of the multi-interference channel, each carrier transmitted is only affected by the channel and only the amplitude is reduced, so that it can be sufficiently recovered. Do. Due to its characteristics, there is a strong advantage in the multipath channel, which is a major problem among the characteristics of the terrestrial channel.

In addition, when the rear end of the effective symbol period exceeding the delay spread of the channel in the time domain of data transmission is used as a guard interval by cyclic prefixing to the front end of the valid symbol, inter-symbol interference experienced on the channel In addition to preventing inter symbol interference (ISI), channel equalization can be performed simply by a single tap equalizer at the receiving end.

Recently, the OFDM transmission scheme has been studied with high interest as a high speed wireless data transmission scheme.

In addition, OFDM systems that require high data rates through mobile radio channels, such as multimedia terrestrial broadcasting and multimedia wireless services, use multiple amplitude modulation schemes such as quadrature amplitude modulation (QAM).

At this time, the multi-amplitude modulation structure in the mobile radio channel requires compensation of channel distortion through dynamic channel estimation.

Therefore, as well as a study on selecting a pilot signal for efficiently performing OFDM channel estimation while maintaining a high data rate, a method of obtaining a channel transfer function by taking a sample average of a pilot signal in a time domain, and a mean square error of a pilot signal Various researches on OFDM channel estimation techniques have been proposed, such as a method of estimating channel characteristics in the frequency domain using (MMSE / LMMSE) and applying them to signal compensation.

However, these findings are based on time-varying or very slow fading channels in consideration of the channel environment.

That is, averaging the pilot signals that exist in multiple consecutive symbols in the time domain in time-invariant or very slow fading channels can significantly reduce the effects of noise, whereas in fast fading channels with many temporal changes, Using the symbol of, serious channel estimation error occurs due to the time-varying nature of the channel.

Therefore, in order to adaptively perform optimal channel estimation according to a channel environment having various practical conditions, a structure for accurately determining time-varying characteristics of a corresponding channel is required.

For example, in the case of digital terrestrial broadcasting systems using the frequency range of VHF (54-216 MHz) or UHF (470-890 MHz), the actual broadcast channel is time-varying when considering mobile reception (eg, a TV installed in a car). to be. In addition, such a time-varying channel gives an OFDM symbol distortion due to intercarrier interference (ICI).

In other words, in channel estimation using a pilot signal, the ICI and additive white gaussian noise (AWGN) act as serious factors that hinder accurate channel estimation.

Therefore, more accurate channel estimation can be achieved by finding out how much interference and noise are caused by the time-varying characteristics of the channel.

FIG. 1 is a block diagram illustrating a general OFDM transmission / reception system, based on a 2k mode of 1705 subcarriers in the digital video broadcasting-terrestrial (DVB-T) standard, which is a standard for European digital terrestrial television. The same applies to the 8k mode in which the number of carriers is 6817.

In other words, when a binary source occurs, the signal mapping unit 101 maps data to be transmitted (that is, binary source) according to a modulation method. In general, the signal mapping unit 101 maps data in a manner called quadrature amplitude modulation (QAM). Modulation methods include quadrature phase shift keying (QPSK), 16-QAM, and 64-QAM.

The data mapped by the signal mapping unit 101 is converted in parallel by a serial / parallel (S / P) conversion unit 102 and a pilot is inserted by the pilot insertion unit 103.

That is, the transmitting side of the OFDM transmission system inserts and transmits a pilot signal having a value known to the receiving side between data to be transmitted to assist the demodulation of the receiving side. In this case, according to the European DVB-T standard, in the 2K mode, 45 continuous pilots per OFDM symbol exist in every symbol, that is, in the same subcarrier position on the time axis. However, the spacing on the frequency axis is random without regularity. The scattered pilot positions are inserted every 12 carriers on the frequency axis and every four symbols on the time axis. That is, the regularity of insertion is repeated every four symbols.

At this time, there are various research results on the pilot insertion position selection, Figure 2 is a DVB-T OFDM frame structure showing an example of which, the black portion represents the position of the continuous pilot, the hatched portion of the distributed pilot The white portion represents the active carrier, that is, the portion of data to be transmitted. Here, the black portion may overlap the continuous pilot and the distributed pilot.

As described above, the pilot-inserted data is output to the inverse fast Fourier transform (IFFT) unit 104 and IFFT, wherein the time-domain sample x (n) is expressed by Equation 1 below.

, n = 0, 1, ..., N-1

, n = 0, 1, ..., N-1

Where N is the number of subcarriers.

In order to prevent intersymbol interference, the IFFT data is inputted to the guard interval inserting section 105, and then converted into a series in the P / S section 106 after the guard interval is inserted longer than the delay spread on the channel. The analog / digital (D / A) converter 107 is converted to analog and then transmitted through a multipath fading channel.

The data passing through the transmission channel as described above is converted to digital in the A / D conversion unit 201 of the receiving side, and again converted in parallel in the S / P unit 202 and then the guard interval as shown in Equation 2 below Output is rejected (203).

  yg (n) = xg (n) * h (n) + w (n)

Where h (n) is the impulse response of the transmission channel, w (n) is the AWGN included in the signal received through the transmission channel, * denotes linear convolution, and subscript g denotes the guard interval. Indicates that it contains.

Therefore, when the guard section is removed by the guard section remover 203, a form as shown in Equation 3 below is obtained.

  y (n) = x (n) * h (n) + w (n)

At this time, it is necessary to examine the impulse response of the transmission channel in detail for the channel estimation technique to determine the influence of the channel.

That is, the impulse response of the channel may be expressed as Equation 4 below.

Where M is the number of propagation paths, h _m is amplitude attenuation, tau _m is the time delay of the m th propagation path, and θ _m is the phase shift.

Therefore, the guard interval at the transmitting side should be inserted longer than τ _max with the maximum time delay in τ _m .

In addition, the data from which the guard interval is removed by the guard interval remover 203 is output to the channel estimator 205 and the signal corrector 206 through an FFT transform action by the FFT unit 204.

In this case, since the signal sequence of the l-th OFDM symbol that has undergone the FFT transform in the received signal y (n) from which the guard interval is removed may be represented by the frequency domain characteristic of Equation 3, this may be expressed as Equation 5 below. have.

Y (l, k) = X (l, k) H (l, k) + W (l, k)

Here, (l, k) represents the k-th subcarrier of the l-th OFDM symbol.

Therefore, in the OFDM frame structure as in the European DVB-T standard, simple channel compensation in the frequency domain may be performed using Equation 5 above.

That is, in the transmission of the frame structure as described above, the channel estimator 205 extracts a pilot signal existing at a known subcarrier position from the FFT signal and then transfers the channel with respect to the pilot signal through comparison with a known reference pilot signal. You can get a function.

)를 추정한다.Then, in order to determine the influence of the channel on the transmitted information data, a linear interpolation method on the frequency axis is performed on the channel transfer function of the pilot signal.

Estimate).

Accordingly, the signal correction unit 206 may recover the transmitted data sample by dividing the received signal sequence, that is, the FFT data Y (k) by the estimated channel transfer function. That is, the FFT signal may be divided by a channel impulse response to compensate for carriers distorted by the channel.

The relationship is as shown in Equation 6 below.

(l, k)는 H(l,k)에 대해 추정된 채널 전달 함수이다.here

(l, k) is the estimated channel transfer function for H (l, k).

The signal compensated by the signal correcting unit 206 is demapped by the signal demapping unit 208 via the P / S unit 207 and then decoded in the form of binary data.

(l, k)가 분모에 있기 때문에 보다 정확한 채널 전달 함수의 추정이 실제 전송된 신호열에 근접한 값을 구하는데 중요한 역할을 한다는 것을 알 수 있다.At this time, in Equation 6

Since (l, k) is in the denominator, it can be seen that a more accurate estimate of the channel transfer function plays an important role in obtaining a value close to the actual transmitted signal sequence.

That is, in the European DVB-T standard, as shown in the OFDM frame structure of FIG. 2, since the regularity is repeated every four symbols while the insertion position of the pilot signal is different for each symbol, a slow fading channel having small time-varying characteristics of the channel is used. In other words, when there is no change in the channel environment over time, as shown in FIG. 1, a more accurate channel effect can be estimated, which is significantly reduced in noise effect, by obtaining a sample average of pilot signals in several symbols before and after the corresponding symbol. have.

However, in case of mobile reception, channel compensation using a plurality of symbols deteriorates the SNR performance of the demodulator due to the time-varying characteristics of the channel environment according to the movement of the receiver. In other words, in a fast fading channel such as moving object reception, the use of multiple symbols reduces the influence of noise, but the time-varying channel effect increases rather than the channel estimation.

As described above, if the multipath on the transmission channel has a time varying characteristic, the fixed channel estimation scheme of FIG. 1 should not be applied.                         

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to provide a dynamic channel estimation apparatus and method for tracking time-varying channel characteristics and an OFDM system using the same.

Another object of the present invention is to determine the time-varying characteristics of the channel to determine the number of symbols that can ignore the time-varying characteristics of the channel, and to estimate the channel transfer function from the pilot signal within the determined symbols, thereby reducing the effects of noise and reducing A channel estimation apparatus and method for maximizing an output signal to noise ratio (SNR) and an OFDM system using the same are provided.

The channel estimation apparatus of the OFDM system according to the present invention for achieving the above object, the number of symbols that can ignore the influence of the time-varying characteristics of the channel by determining the degree of dispersion of the extracted pilot signal using the mobility estimation unit The signal compensation is performed by linearly interpolating a sample average of the pilot signals within the determined symbol.

The mobility estimator includes a mobility variance calculator that calculates a variance of the continuous pilot from the pilot signal, and selects only a distributed pilot from the pilot signal to perform pre-channel compensation through linear interpolation, and then performs a continuous pilot from the compensated signal. The noise variance calculator which calculates the variance of the noise mixed with the variance and the variance variance obtained by the mobility variance calculator and the noise variance obtained by the noise variance calculator determine the number of symbols that can ignore the time-varying characteristics of the channel. It is characterized by consisting of a comparison unit.

The mobility variance calculator calculates the variance in the time domain for each continuous pilot included in one OFDM symbol and takes an average.
In the channel estimation method according to the present invention,
(a) receiving the FFT signal and the pilot signal, calculating noise variance and mobility variance at the pilot position, respectively, and determining the number of symbols that can ignore time-varying characteristics of the channel by comparing the two variances;
(b) obtaining a sample average of pilot signals present in the number of symbols determined in step (a);
(c) linearly interpolating a sample average of the pilot signal on a time axis and a frequency axis, respectively; And
(d) dividing the FFT signal by the linear interpolated signal in step (c) to compensate for carriers distorted by the channel.
In the step (a), the variance in the time domain is obtained for each continuous pilot included in one OFDM symbol, and then the average is calculated to calculate the mobility variance of the pilot signal, and only the distributed pilot is selected from the pilot signals. Calculating the variance of the noise mixed in the continuous pilot from the compensated signal after performing the pre-channel compensation through linear interpolation, and ignoring the time-varying characteristics of the channel by comparing the mobility variance with the noise variance obtained in the step. Determining a number of possible symbols.
An OFDM system according to the present invention includes: a received signal processor for digitizing an Orthogonal Frequency Division Multiplexed (OFDM) signal received through a transmission channel and performing FFT conversion after removing a guard interval inserted at a transmitting side; A pilot extractor for extracting a pilot signal present at a known subcarrier position from the FFT signal; A mobility estimator which receives the FFT signal and the pilot signal, calculates noise variance and mobility variance at a pilot position, and determines the number of symbols that can ignore time-varying characteristics of the channel by comparing the two variances; A signal compensator for linearly interpolating a sample average of pilot signals existing in the number of symbols determined by the mobility estimator to compensate for the signal; And a signal correction unit for dividing the FFT signal by the output of the signal compensation unit to compensate carriers distorted by the channel.
Other objects, features and advantages of the present invention will become apparent from the following detailed description of embodiments taken in conjunction with the accompanying drawings.

delete

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

The present invention is to determine the number of symbols used for channel estimation to adaptively determine the time-varying characteristics of the channel to maintain the maximum output SNR.

The channel estimation apparatus of the OFDM system according to the present invention for realizing this is shown in FIG.

Referring to FIG. 3, a pilot extractor 301 extracting a pilot signal from the FFT data, calculates a mobility variance from the extracted pilot signal, and then calculates a mobility variance from the extracted pilot signal. A mobility estimator 302 for determining the number of symbols to be applied to the final channel compensation through comparison, a pilot signal sample averaging unit 303 for averaging a pilot signal in symbols determined by the mobility estimator 302, And a linear interpolation unit 304 for linearly interpolating the sample averaged signal and outputting the linearly interpolated signal to the signal correction unit 206.

The mobility estimator 302 selects only a distributed pilot from the extracted pilot signals to perform pre-channel compensation through linear interpolation, and then calculates a variance of noise mixed in the continuous pilot from the compensated signal. 302-1), a mobility variance calculator 302-2 for calculating the variance of the continuous pilot inserted in the same subcarrier position in every symbol among the extracted pilot signals, and the noise variance calculator 302- The comparator 302-3 determines the number of symbols that can ignore the time-varying characteristics of the channel by comparing the noise variance obtained in 1) with the mobility variance calculated by the mobility variance calculator 302-2.

In the present invention configured as described above, the pilot extractor 301 extracts a pilot signal included in an FFT transform, that is, a received signal sequence, that is, the FFT signal, to obtain a noise variance calculator 302-1 and a mobility variance calculator 302-2. )

The mobility variance calculator 302-2 calculates the variance of the continuous pilot inserted in the same subcarrier position for every symbol among the extracted pilot signals.

 In the case of the fixed receiver, since the channel is invariant in time, the mobility variance is almost constant since the influence of the channel is considered only by the noise component due to AWGN. If the moving line variance has a constant value, i.e., in the case of a fixed antenna, the larger the number of symbols to be used for channel estimation, the larger the use of the symbol, the greater the effect of noise.

However, when the receiver moves, the larger the number of symbols used for the mobility variance calculation due to the time-varying characteristics of the channel, the smaller the effect of noise but the greater the mobility variance. High variance means that the degree of distortion of the channel is large.

In the present invention, a more accurate calculation of mobility variance uses a structure in which each of the 45 continuous pilots is averaged over 45 again after obtaining a variance in the time domain.

으로 정의하고 다음의 수학식 7과 같이 표현할 수 있다.Mobility distribution

It can be defined as Equation 7 below.

, Nl = 2, 3, ..., lmax

, Nl = 2, 3, ..., lmax

Here, Y (l, k _c ) is a continuous pilot signal received at the k _c subcarrier position of the l th symbol, and μ _k is an average value of Y (l, k _c ).

In addition, since only the continuous pilot is considered, k _c represents only 45 fixed positions, and N ₁ is the number of symbols used. Since Equation 7 precedes an operation for obtaining a variance in the time domain, variance is calculated for one symbol. Has no meaning.

Therefore, N _l is considered from the operation on two or more symbols, and l _max should be selected to satisfy the tradeoff between the complexity of the actual implementation and the accuracy of time-varying characteristics of the channel.

For example, the mobility variance calculator 302-2 may calculate the mobility variance from the continuous pilot and the distributed pilot.

In this case, the noise variance calculator 302-1 calculates a noise variance corresponding to the criterion to determine the degree of the noise component due to the time-varying multipath included in the mobility variance.

To this end, the noise variance calculator 302-1 first selects only a distributed pilot from the extracted pilot signals to obtain a prechannel transfer function through linear interpolation to perform prechannel compensation.

으로 표현하면 다음의 수학식 8과 같다. From the compensated signal obtained here, we can find the variance of the noise mixed in the continuous pilot.

When expressed as shown in Equation 8 below.

, Nl = 1, 2,..., lmax

, Nl = 1, 2, ..., lmax

Where N (l, k _c ) is the noise mixed in the pre-channel compensated continuous pilot, μ _N is the mean value of N (l, k _c ), where k _c also represents only a fixed position of 45 continuous pilots .

In this case, the noise component N (l, k _c ) is represented by the following equation (9).

는 분산형 파일롯만을 선택하여 선형보간을 통해 추정한 채널 전달 함수이고, P_c는 송신단에서 미리 삽입한 알려진 연속형 파일롯 값이다.here

Is a channel transfer function estimated through linear interpolation by selecting only a distributed pilot, and P _c is a known continuous pilot value previously inserted by a transmitter.

에서 추정 오류가 없다면 N(l, k_c)는 순수한 잡음만을 나타낸다.In Equation 9

If there is no estimation error in, N (l, k _c ) represents pure noise only.

In this case, after the pre-channel compensation is performed in the noise variance calculator 302-1, calculating the variance of the noise mixed in the continuous pilot from the compensated signal is to increase the reliability of the system. The noise variance may be calculated from a continuous pilot or from a continuous pilot and a distributed pilot.

)과 잡음 분산(

)을 비교하여 해당 채널 환경에 적합한 최적의 심볼 수를 판단한다.The noise variances obtained from Equations 8 and 9 serve as a criterion for ignoring the time-varying effects of the channel and finding the number of symbols that can take a sample average, and the comparator 302-3 calculates the mobility variance (

) And noise variance (

) To determine the optimal number of symbols for the channel environment.

Accordingly, when the mobility estimator 302 takes a sample average of the pilot signals present in the determined number of symbols and performs linear interpolation to compensate the signal, the output SNR is the maximum output SNR in the channel environment.

That is, the pilot signal sample averaging unit 303 takes a sample average of pilot signals existing in the number of symbols determined by the mobility estimation unit 302, extracts frequency characteristics, and outputs the frequency characteristic to the linear interpolation unit 304. The linear interpolation unit 304 linearly interpolates this on the time axis and the frequency axis to extract frequency characteristics in the total subcarriers. That is, since the output of the pilot signal sample averaging unit 303 is a frequency characteristic at the pilot position, the frequency characteristic at the actual data position should also be obtained. To this end, the linear interpolator 304 infers the channel characteristics of the required active carrier that we actually want to know by interpolation from a known pilot signal.

In this case, the linear interpolation unit 304 may first interpolate on a time axis and then interpolate on a frequency axis. Here, the reason for interpolating on the time axis first is that interpolation on the time axis first results in the pilot not being present in every 12 carriers but in every three carriers on the frequency axis. You can expect the effect to decrease to 1.

The signal corrector 206 compensates for the carriers distorted by the channel by applying the output of the linear interpolator 304 to the FFT signal. For example, the FFT signal may be divided by the output of the linear interpolator 304 to compensate for a carrier distorted by the channel.

In this case, the process of the mobility estimator 302 comparing the mobility variance and the noise variance to determine the optimal number of symbols suitable for the channel environment is as follows.

That is, the process of determining the optimal number of symbols suitable for the channel environment through the computer simulation results of applying the channel estimation apparatus according to the present invention to time-varying multipath will be described.

To this end, the present invention uses the 2k mode frame structure of the DVB-T standard, the modulation scheme uses 64-QAM, and the insertion position of the pilot signal is shown in FIG.

Then, it is assumed that the data flow is allocated to 10 bits, and the carrier and symbol timing at the transmitter and the receiver are synchronized.

In this case, the channel model used in the simulation is the Rayleigh channel of GSM Recommendation 05.05, and the main characteristics corresponding to the channel model are shown in Table 1 below.

  Route

위상size

Phase    Time delay   Amount of time change (degrees / sec)     One 0.50

30 °      0.0㎲      None     2 1.00

0 °      0.1㎲      None     3 0.63

140 °      0.2㎲   -198 ° / sec     4 0.25

60 °      0.3 ㎲    238 ° / sec     5 0.16

180 °      0.4 ㎲    278 ° / sec     6 0.10

220 °      0.5 ㎲   -218 ° / sec

 In the present invention, as shown in Table 1, a time-varying channel environment is realized by changing a phase by adding a time-varying amount item.

In the rayleigh channel model of Table 1, the path 5 having the maximum amount of time variation causes a phase shift of 278 ° per second, which is a rate of bringing a phase shift of 70 ° per 100 OFDM symbols. In addition, the white Gaussian noise added on the transmission channel was such that the C / N was 20.1 dB.

In this case, the mobility variance has a substantially constant value in a multipath channel having no change over time, and the number of symbols to be used for channel estimation is approximately one frame, considering the noise reduction effect and the memory size. Can be determined.

Accordingly, approximately 50 symbols are used in the time-varying multipath channel, and up to 68 symbols corresponding to one frame are used to obtain mobility and noise variance. That is, it means that l _max in Equations 7 and 8 is considered as 68. This is a sufficient number of symbols to understand the time varying characteristics of the channel.

4 shows the mobility variance and the noise variance obtained as a result of the simulation under the above conditions. At this time, the output SNR of the signal compensated for the influence of the channel is applied as a criterion for evaluating the performance of the channel estimation scheme proposed by the present invention.

5 shows the output SNR for the Rayleigh channel model, where the horizontal axis represents the number of symbols that take a sample average over the pilot signal.

That is, as shown in FIG. 4, since the channel change characteristic with time distorts the continuous pilot signal transmitted with the same value, the value of mobility variance continues to increase as the number of symbols used to obtain variance increases. Since the noise variance is the variance of the small mixed in the continuous pilot in the pre-channel compensated signal, the influence of the time-varying factor is reduced, indicating only the value due to the noise effect by the AWGN.

Therefore, the number of corresponding symbols at the point where the mobility variance and the noise variance meet is a time interval in which the influence of the time-varying characteristics of the channel can be ignored.

Therefore, the comparator 302-3 of the mobility estimator 302 determines that 20 symbols for the Rayleigh channel are the maximum number of symbols that can ignore time-varying characteristics of the channel, and uses the 20 symbols to average the sample of the pilot signal. Decide to take.

That is, in FIG. 4 and FIG. 5, after applying the 20 symbols corresponding to the point where the mobility variance and the noise variance are applied, the channel transfer function is estimated and the compensated signal is the best as the output SNR of 14.8 dB. This means that the output SNR of the compensation signal can be increased by taking the sample average of the pilot signal present in the number of symbols up to the corresponding symbol in the time domain where the influence of the time-varying characteristics of the channel can be ignored.

However, the use of a larger number of symbols results in greater distortion of the received signal due to the time-varying nature of the channel, resulting in poor performance of the output SNR. That is, in the channel estimation algorithm using the sample average of the pilot signal, it is most important to select the number of symbols used to estimate the channel transfer function according to the channel environment.

Therefore, the proposed channel estimator for the various time-varying multipath fading channels experienced by a high-speed mobile object estimates an optimal channel transfer function suitable for the channel environment to maintain demodulator SNR performance after channel compensation to the maximum. This can be confirmed by computer simulation.

As described above, according to the channel estimation apparatus and method of the digital TV receiver for receiving digital terrestrial television broadcasting using the OFDM transmission method according to the present invention, and the OFDM system using the same, the mobility variance is found from the pilot signal extracted from the received signal. By comparing the channel variance with the noise compensation over the channel compensation, the number of symbols that can be ignored by the time-varying nature of the channel can be neglected, and the channel transfer function can be estimated from the pilot signal within the determined symbol, thereby reducing the effect of noise and then after channel compensation. Since the output SNR of is maximized, there is an effect that an optimal channel estimation is made.

Claims (30)

A channel estimating apparatus of an orthogonal frequency division multiplexing (OFDM) system that takes a sample average of a received pilot signal, obtains a channel transfer function, and then applies a fast Fourier transform (FFT) signal to compensate for a signal distorted by a channel. A mobility estimator that receives the FFT signal and the pilot signal, calculates noise variance and mobility variance at the pilot position, respectively, and determines the number of symbols that can ignore time-varying characteristics of the channel by comparing the two variances; And a signal compensator for linearly interpolating a sample average of pilot signals existing in the number of symbols determined by the mobility estimator to compensate for the signal. The method of claim 1, wherein the mobility estimating unit A mobility variance calculator for calculating a mobility variance of the pilot signal; A noise variance calculator configured to select only a distributed pilot from the pilot signals to perform pre-channel compensation through interpolation, and then calculate a variance of noise mixed in the continuous pilot from the compensated signal; Orthogonal frequency division multiplexing system comprising a comparison unit for determining the number of symbols that can ignore time-varying characteristics of a channel by comparing the mobility variance obtained by the mobility variance calculator with the noise variance obtained by the noise variance calculator Channel estimation device. The method of claim 2, wherein the mobility variance calculator An apparatus for channel estimation in an orthogonal frequency division multiple system, characterized in that the average is obtained after obtaining variance in the time domain for each continuous pilot included in one OFDM symbol. The method of claim 2, wherein the mobility variance calculator A channel estimation apparatus of an orthogonal frequency division multiple system, characterized in that the average is obtained after obtaining variances in the time domain for each of the included continuous pilots and the distributed pilots in one OFDM symbol. The method of claim 2, Variance of noise mixed in a pilot obtained by the noise variance calculator (

) Is a channel estimating apparatus of an orthogonal frequency division multiplexing system characterized by the following equation.

, N _l = 1, 2, ..., l _max Where N (l, k _c ) is the noise mixed in the pre-channel compensated continuous pilot, μ _N is the mean value of N (l, k _c ), and k _c represents the predetermined position of the continuous pilot. The method of claim 5, The noise component N (l, k _c ) of the noise variance calculation unit can be expressed by the following equation.

Here, Y (l, k _c ) is a continuous pilot signal received at the k _c th subcarrier position of the l th symbol,

Is the channel transfer function estimated by linear interpolation by selecting only the distributed pilot, and P _c is a known continuous pilot value previously inserted by the transmitter. The method of claim 6, The channel transfer function

If there is no estimation error in the noise component N (l, k _c ) channel estimation apparatus of orthogonal frequency division multiple system, characterized in that it represents only pure noise. The noise dispersion calculator of claim 2, wherein the noise variance calculator And performing pre-channel compensation by performing linear interpolation on the distributed pilot selected from the pilot signals to obtain a pre-channel transfer function. The noise dispersion calculator of claim 2, wherein the noise variance calculator And selecting only a continuous pilot from the pilot signals and calculating a variance of noise without prior channel compensation. The noise dispersion calculator of claim 2, wherein the noise variance calculator And a continuous pilot and a distributed pilot from among the pilot signals, and then calculate a variance of noise without prior channel compensation. The method of claim 2, wherein the mobility variance calculator A channel estimation apparatus of an orthogonal frequency division multiple system, characterized in that the average is obtained after obtaining variance in the time domain for each of 45 continuous pilots included in one OFDM symbol. The method of claim 11, wherein the mobility variance calculation unit Mobility distribution by applying the following equation

And a channel estimating apparatus for an orthogonal frequency division multiplexing system.

, N _l = 2, 3, ..., l _max Where Y (l, k _c ) is a continuous pilot signal received at the k _c subcarrier position of the l th symbol, μ _k is the mean value of Y (l, k _c ), k _c is 45 predetermined positions, N _l is the number of symbols used. The method of claim 1, wherein the signal compensator A pilot signal sample averaging unit for taking out and outputting a sample average of pilot signals existing in the number of symbols determined by the mobility estimator; And a linear interpolation unit configured to extract frequency characteristics of a total subcarrier by linearly interpolating the output of the pilot signal sample average unit on a time axis and a frequency axis, respectively. A received signal processor for digitizing the orthogonal frequency division multiplexed (OFDM) signal received through the transmission channel, removing the guard interval inserted by the transmitter, and performing FFT conversion; A pilot extractor for extracting a pilot signal present at a known subcarrier position from the FFT signal; A mobility estimator which receives the FFT signal and the pilot signal, calculates noise variance and mobility variance at a pilot position, and determines the number of symbols that can ignore time-varying characteristics of the channel by comparing the two variances; A signal compensator for linearly interpolating a sample average of pilot signals existing in the number of symbols determined by the mobility estimator to compensate for the signal; And And a signal correction unit for dividing the FFT signal by the output of the signal compensation unit to compensate carriers distorted by the channel. 15. The method of claim 14, wherein the mobility estimating unit A mobility variance calculator for calculating a mobility variance of the pilot signal; A noise variance calculator for selecting only a distributed pilot from the pilot signals to perform pre-channel compensation through linear interpolation and calculating a variance of noise mixed in the continuous pilot from the compensated signal; Orthogonal frequency division multiplexing system comprising a comparison unit for determining the number of symbols that can ignore time-varying characteristics of a channel by comparing the mobility variance obtained by the mobility variance calculator with the noise variance obtained by the noise variance calculator . The method of claim 15, wherein the mobility variance calculator An orthogonal frequency division multiplexing system, characterized in that the average is obtained after finding variance in the time domain for each continuous pilot included in one OFDM symbol. The method of claim 15, wherein the mobility variance calculator Orthogonal frequency division multiplexing system characterized in that the average is obtained after variance is obtained in the time domain for each of the 45 continuous pilots included in one OFDM symbol. The method of claim 17, wherein the mobility variance calculator Mobility distribution by applying the following equation

Orthogonal frequency division multiplexing system comprising:

, N _l = 2, 3, ..., l _max Where Y (l, k _c ) is a continuous pilot signal received at the k _c subcarrier position of the l th symbol, μ _k is the mean value of Y (l, k _c ), k _c is 45 predetermined positions, N _l is the number of symbols used. The method of claim 15, wherein the mobility variance calculator An orthogonal frequency division multiplexing system, characterized by taking the variance in the time domain for each of the included continuous pilots and distributed pilots in one OFDM symbol. The method of claim 15, Variance of noise mixed in a pilot obtained by the noise variance calculator (

) Is orthogonal frequency division multiplexing system characterized by the following equation.

, N _l = 1, 2, ..., l _max Where N (l, k _c ) is the noise mixed in the pre-channel compensated continuous pilot, μ _N is the mean value of N (l, k _c ), and k _c represents the predetermined position of the continuous pilot. The method of claim 20, The noise component N (l, k _c ) of the noise variance calculator can be expressed by the following equation.

Here, Y (l, k _c ) is a continuous pilot signal received at the k _c th subcarrier position of the l th symbol,

Is the channel transfer function estimated by linear interpolation by selecting only the distributed pilot, and P _c is a known continuous pilot value previously inserted by the transmitter. The method of claim 21, The channel transfer function

In the orthogonal frequency division multiplexing system, if there is no estimation error in, the noise component N (l, k _c ) represents pure noise only. 16. The apparatus of claim 15, wherein the noise variance calculator And selecting only a continuous pilot from the pilot signals and calculating a variance of noise without prior channel compensation. 16. The apparatus of claim 15, wherein the noise variance calculator Orthogonal frequency division multiplexing system characterized in that for selecting the continuous pilot and distributed pilot from the pilot signal and calculate the variance of noise without prior channel compensation. 15. The apparatus of claim 14, wherein the signal compensator Orthogonal frequency division multiplexing, wherein the sample average of the pilot signals existing in the number of symbols determined by the mobility estimator is linearly interpolated on the time axis and then linearly interpolated on the frequency axis. A channel estimation method of an orthogonal frequency division multiplexing (OFDM) system that takes a sample average of a received pilot signal, obtains a channel transfer function, and then applies a fast Fourier transform (FFT) signal to compensate for a signal distorted by a channel. (a) receiving the FFT signal and the pilot signal, calculating noise variance and mobility variance at a pilot position, respectively, and determining the number of symbols that can ignore time-varying characteristics of the channel by comparing the two variances; (b) obtaining a sample average of pilot signals present in the number of symbols determined in step (a); (c) linearly interpolating a sample average of the pilot signal on a time axis and a frequency axis, respectively; And (d) dividing the FFT signal by the linear interpolated signal in step (c) to compensate carriers distorted by the channel. The method of claim 26, wherein step (a) Calculating the variance variance of the pilot signal by taking an average after variance is obtained in the time domain for each continuous pilot included in one OFDM symbol, Calculating only the variance of the noise mixed in the continuous pilot from the compensated signal after performing pre-channel compensation through linear interpolation by selecting only a distributed pilot from the pilot signal; And determining the number of symbols that can ignore time-varying characteristics of the channel by comparing the mobility variance and the noise variance obtained in the above step. 28. The method of claim 27, wherein the mobility dispersing step is Mobility distribution by applying the following equation

And channel estimation method of an orthogonal frequency division multiplexing system.

, N _l = 2, 3, ..., l _max Where Y (l, k _c ) is a continuous pilot signal received at the k _c subcarrier position of the l th symbol, μ _k is the mean value of Y (l, k _c ), k _c is 45 predetermined positions, N _l is the number of symbols used. 28. The method of claim 27, wherein the noise variance step is Dispersion of Noise Mixed in the Pilot

) Is a channel estimation method of an orthogonal frequency division multiplexing system, characterized by the following equation.

, N _l = 1, 2, ..., l _max Where N (l, k _c ) is the noise mixed in the pre-channel compensated continuous pilot, μ _N is the mean value of N (l, k _c ), and k _c represents the predetermined position of the continuous pilot. The method of claim 29, The noise component N (l, k _c ) can be represented by the following equation.

Here, Y (l, k _c ) is a continuous pilot signal received at the k _c th subcarrier position of the l th symbol,

Is the channel transfer function estimated by linear interpolation by selecting only the distributed pilot, and P _c is a known continuous pilot value previously inserted by the transmitter.

Images