HK1138690A

HK1138690A - Ofdm channel estimation

Info

Publication number: HK1138690A
Application number: HK10104009.9A
Authority: HK
Inventors: 马晓强; A‧图滋尼
Original assignee: 高通股份有限公司
Priority date: 2006-08-28
Filing date: 2007-08-27
Publication date: 2010-08-27

Description

OFDM channel estimation

Background

Today, there are many forms of information that are transmitted from information sources, such as television content providers, to receivers, such as televisions in the homes of people. Thus, an example of such information is Digital Television (DTV) information. Communicating digital information typically involves converting the digital information to an analog signal and modulating the amplitude and/or phase of an RF (radio frequency) carrier frequency using the analog signal, and transmitting this modulated signal to a receiver over a propagation medium such as air.

Referring to fig. 1, a communication system 1 includes a transmitter 2 and a receiver 4. The transmitter 2 and receiver 4 have respective antennas 6 and 8, and although the antennas 6 and 8 are shown here as being external to the transmitter 2 and receiver 4, they may be considered as components of the transmitter 2 and receiver 4. The transmitter 2 is configured to transmit information (e.g., as with signals 14, 16, and 18) to the receiver 4 over a propagation medium, here a terrestrial broadcast system. Transmitting information over a propagation medium introduces signal distortion caused by noise (e.g., static electricity), intensity variations (fading), phase shift variations, doppler broadening, doppler fading, multipath delays, and the like. Multipath delays arise from the transmitted signal taking a different path between the transmitter and receiver through the propagation medium, e.g., because it is reflected from building 10 and/or relayed through relay station 12, etc. Different paths of the transmitted signal p (t) (e.g., signals 14, 16, and 18) result in different gains and different delay times, which results in time-delayed copies of the signal p (t) arriving at the receiver 4 at different times (like echoes) than the directly transmitted signal 16. The received signal r (t) is the directly transmitted signal and/or the combination of copies (if any). Multipath distortion results in inter-symbol interference (ISI) that causes the weighted effects of other symbols to be added to the current symbol, and/or inter-channel interference that causes separate subcarriers to interfere with each other. Noise and/or interference in the signal r (t) may also come from other sources such as a transmitter. These effects may lead to errors in the transfer of information from the transmitter 2 to the receiver 4 and/or in the interpretation of this information. The system fails when the Bit Error Rate (BER) of the system exceeds a threshold and the error tolerance of the system is exceeded.

Orthogonal Frequency Division Multiplexing (OFDM) can be used to transmit multiple DTV signals over a propagation medium. OFDM systems transmit signals, such as television signals, in parallel on one or more subcarriers using one or more time slots in each subcarrier. Each subcarrier is located at a different portion of the spectrum used to transmit the DTV signal. The spacing of the subcarriers is such that the frequency of each subcarrier is orthogonal to the frequency of each of the other subcarriers (e.g., the frequency spacing of the subcarriers is substantially equal to the inverse of the OFDM symbol duration). The orthogonality of the subcarrier frequencies provides greater resistance to RF interference and multipath distortion than when non-orthogonal frequencies are used as subcarriers. Each subcarrier includes, for example, data symbols, pilot symbols, and/or Transmission Parameter Signal (TPS) symbols, which are special type data symbols. The pilot symbols are predetermined known signals that are used to help the receiver estimate the transmission channel.

Once the receiver 4 receives an OFDM signal (e.g., r (t)), the channel estimate is used to cancel the effects of distortion on the transmitted signal (e.g., p (t)). For example, since r (t) is a linear combination of signals 14, 16, and 18, a particular mathematical function (i.e., a transfer function) may describe the relationship of p (t) to r (t). Once the transfer function of the propagation medium is known, a filter that is the inverse of the transfer function of the propagation medium may be used to reduce the effects of distortion introduced by the propagation medium. The transfer function of any given propagation medium is also changing due to the fact that the propagation medium is changing (e.g., objects causing multipath errors may move, weather may change, noise levels may change, etc.). When the propagation medium is changing rapidly, r (t) becomes a "fast fading" transmission channel, which increases the complexity of the estimation process. "long channels" occur when the propagation medium has a long delay spread. For example, if the distance that signal 18 travels to reach receiver 4 is long enough compared to signal 16, the symbols in signal 18 may not arrive until after the same symbols reach receiver 4 via signal 16, thereby causing ISI.

Summary of the invention

A receiver for receiving an Orthogonal Frequency Division Multiplexing (OFDM) Digital Video Broadcasting (DVB) signal comprising a set of OFDM symbols including data symbols and pilot symbols transmitted using a number of sub-carriers, the OFDM DVB signal being transmitted to the receiver via a transmission channel, the receiver comprising: an input module configured to receive the OFDM DVB signal via a transmission channel; and a channel estimation module coupled to the input module configured to compute a channel estimate for the transmission channel by performing a fourier transform on the set of OFDM symbols to produce a set of transformed symbols in the frequency domain and by performing Minimum Mean Square Error (MMSE) equalization on the set of transformed symbols using a subset of pilot symbols in this OFDM DVB signal.

Implementations of the invention may include one or more of the following features. The channel estimation module performs a single fourier transform on each received set of OFDM symbols. The subset of pilot symbols is substantially less than all of the pilot symbols in the set of OFDM symbols. The channel estimation module is configured to select, as the subset of pilot symbols, the N closest pilot symbols transmitted using subcarriers having a lower frequency than a frequency used to transmit the data symbol being viewed and the N closest pilot symbols transmitted using subcarriers having a higher frequency than the frequency used to transmit the data symbol being viewed. The channel estimation module is configured to select the N closest pilot symbols transmitted using subcarriers having a frequency higher than the frequency used to transmit the data symbol under examination as the subset of pilot symbols. The channel estimation module is configured to select the N closest pilot symbols transmitted using subcarriers having a frequency lower than the frequency used to transmit the data symbol under view as the subset of pilot symbols. The channel estimation module calculates channel estimates corresponding to each respective subcarrier used to transmit each of the sets of OFDM symbols.

Furthermore, implementations of the invention may also include one or more of the following features. The channel estimation module is configured to compute channel estimates for more than one data symbol in parallel. The constant N is equal to 2 and the channel estimation module is configured to perform channel estimation on each of the execution data symbols according to:

wherein c is_pAre the coefficients of the respective filters and are,is the channel estimation for the respective pilot symbol in the subset of pilot symbols, and p is the index value for the respective pilot symbol in the subset of pilot symbols. The channel estimation module is configured to be calculated according toCalculating the filter coefficient:

c(m)＝[c₁(m) c₂(m) c₃(m) c₄(m)]＝

[R((m-m₁)_M) R((m-m₂)_M) R((m-m₃)_M) R((m-m₄)_M)]R^-1

wherein

And is

Where M is the total number of subcarriers present in the OFDM DVB signal, L is an index representing each respective path of the multipath OFDM DVB signal, L is a number representing the delay spread of channel 115, M is an integer value representing the number of intervening channels between two channels selected for correlation, and

wherein h is_l(k) Is the channel impulse response of the l-th path of the transmission channel at time k.

Furthermore, implementations of the invention may also include one or more of the following features. The channel estimation module is configured to calculate an inter-channel interference (ICI) value related to the transmission channel in the frequency domain. The channel estimation module is configured to perform channel estimation using this ICI value. The channel estimation module is configured to perform MMSE equalization after subtracting this ICI value from the received OFDM DVB signal. The channel estimation module is configured to calculate a channel estimate by estimating a channel correlation value according to:

wherein K is satisfied (K)₂-k₁)_MThe number of cases M and M is the total number of subcarriers present in this OFDM DVB signal. The estimated channel correlation values are averaged according to:

where β is a predetermined known parameter. The channel estimation module is configured to use a Viterbi (Viterbi) decoder.

In general, in another aspect, the invention provides a method for calculating a channel estimate for a transmission channel used to transmit an Orthogonal Frequency Division Multiplexing (OFDM) Digital Video Broadcasting (DVB) signal comprising a set of OFDM symbols including data symbols and pilot symbols transmitted using a number of subcarriers, the method comprising: receiving an OFDM DVB signal via a transmission channel at a receiver; a channel estimate for a transmission channel is calculated in the frequency domain by performing a fourier transform on a set of OFDM symbols to produce a set of transformed symbols in the frequency domain, and by performing Minimum Mean Square Error (MMSE) equalization on these set of transformed symbols using a subset of pilot symbols in this OFDM DVB signal.

Implementations of the invention may include one or more of the following features. Calculating the channel estimate comprises calculating the channel estimate by performing a single fourier transform on each received set of OFDM symbols. The method also includes selecting, as the subset of pilot symbols, the N closest pilot symbols transmitted using subcarriers having a lower frequency than a frequency used to transmit the data symbol being viewed and the N closest pilot symbols transmitted using subcarriers having a higher frequency than the frequency used to transmit the data symbol being viewed. The method also includes selecting, as the subset of pilot symbols, the N closest pilot symbols transmitted using subcarriers having a frequency lower than the frequency used to transmit the data symbol under examination. The method also includes selecting, as the subset of pilot symbols, the N closest pilot symbols transmitted using subcarriers having a frequency higher than a frequency used to transmit the data symbol under examination. Computing the channel estimate comprises computing channel estimates for more than one data symbol in parallel.

Furthermore, implementations of the invention may also include one or more of the following features. Calculating the channel estimate comprises calculating the channel estimate for the data symbol of interest as N equals 2 and according to:

wherein c is_pAre the coefficients of the respective filters and are,is the channel estimation for the respective pilot symbol in the subset of pilot symbols, and p is the index value for the respective pilot symbol in the subset of pilot symbols. Calculating the channel estimate comprises calculating filter coefficients according to:

c(m)＝[c₁(m) c₂(m) c₃(m) c₄(m)]＝

[R((m-m₁)_M) R((m-m₂)_M) R((m-m₃)_M) R((m-m₄)_M)]R^-1

wherein

And is

Furthermore, implementations of the invention may also include one or more of the following features. Calculating the channel estimate comprises calculating a channel estimate corresponding to each respective subcarrier used to transmit the OFDM DVB signal. The method also includes calculating an inter-channel interference (ICI) value associated with the transmission channel in the frequency domain. Calculating the channel estimate includes calculating a channel estimate for the transmission channel using the ICI value. Calculating the channel estimate comprises performing MMSE equalization after subtracting this ICI value from the OFDM DVB signal. Calculating the channel estimate further comprises calculating the channel estimate by estimating a channel correlation value according to:

wherein K is satisfied (K)₂-k₁)_MThe number of cases M and M is the total number of subcarriers present in this OFDM DVB signal. Calculating the channel estimate further comprises averaging the estimated channel correlation values according to:

where β is a predetermined known parameter. Calculating the channel estimate further includes performing viterbi decoding.

Various aspects of the invention may provide one or more of the following capabilities. The calculations performed by the DTV channel estimation module can be reduced compared to the prior art. The implementation cost of the DTV channel estimation module can be reduced compared to the prior art. Channel estimation may be performed by using a single fourier transform for each symbol received via each channel of the OFDM DTV signal. One-dimensional channel estimation may be performed in the frequency domain. The minimum mean square error estimation may be performed using a subset of pilot symbols provided via the OFDM DTV signal. A subset (e.g., 4) of the pilot symbols may be used to calculate a channel estimate for the selected channel. Using every third symbol (starting at the second lowest frequency subcarrier) in combination with a pilot symbol can result in similar performance in an 8K DVB-T/H system as using all subcarriers. These and other capabilities of the present invention will be more fully understood along with the invention itself after a review of the following figures, detailed description, and claims.

Brief Description of Drawings

Fig. 1 is a schematic diagram of a transmission channel.

Fig. 2 is a block diagram of a baseband OFDM system including a transmitter and a receiver.

Fig. 3 is a diagram of an OFDM frame transmitted by the transmitter shown in fig. 2.

Fig. 4 is a block diagram of functional elements of the receiver shown in fig. 2.

Fig. 5 is a flowchart of a procedure for performing channel estimation in the OFDM transmission system shown in fig. 2.

DETAILED DESCRIPTIONS

Embodiments of the present invention provide techniques for transmission channel estimation and equalization for Digital Video Broadcast (DVB) transmissions by combining Minimum Mean Square Error (MMSE) equalization using a subset of the total available pilot symbols with inter-channel interference (ICI) estimation in the frequency domain. For example, a DTV system includes a transmitter and a receiver. The transmitter generates an OFDM signal including pilot symbols. The transmitter broadcasts the OFDM signal to the receiver. The receiver uses frequency domain MMSE equalization to compute the channel estimate. MMSE equalization is performed using the channel correlation information and a subset of the pilots contained in the OFDM signal. These channel estimates are used by the receiver to estimate the signal transmitted by the transmitter. The receiver outputs a resulting signal that is substantially similar to the signal transmitted by the transmitter to the receiver. Other embodiments are also within the scope of the invention.

Referring to fig. 2, the OFDM transmission system 100 includes a transmitter 105 and a receiver 110. The system 100 includes appropriate hardware, firmware, and/or software (including computer-readable instructions, preferably computer-executable instructions) to implement the functions described below. The transmitter 105 and the receiver 110 may be configured to communicate various types of information. Here, by way of example only and not limitation, transmitter 105 is a transmitter for DTV signals and receiver 110 is a DTV receiver such as a digital television or a combination of a set-top box and a digital television. For example, system 100 is configured to transmit and receive terrestrial DTV signals according to the DVB-T/H standard via an antenna (not shown in FIG. 2). The transmitter 105 and the receiver 110 are linked by a transmission channel 115. The transmission channel 115 is a propagation medium such as the atmosphere (in the case of terrestrial broadcast), although other propagation media are possible (e.g., cable if appropriate transmitters and receivers are used). The transmitter 105 is configured to receive an input signal 120 and broadcast an OFDM signal 125 to the receiver 110. The transmission channel 115 affects the signal 125 and transforms it into a signal 130. Receiver 110 is configured to receive signal 130 and output an output signal 135 that is preferably substantially equal to signal 125.

The relationship of signals 125 and 130 may be defined in the time domain as:

where y (k) is signal 130, l is the index of the channel path (of transmission channel 115) representing the channel taps in the time domain, x (k-l) is signal 125, h_l(k) Is the Channel Impulse Response (CIR) of the l-th path of the transmission channel 115 at time k, n (k) (0 ≦ k ≦ M-1) is that both the real and imaginary components in the time domain have zero mean and variance σ²Independent complex-valued random gaussian variable of (a)E.g., Additive White Gaussian Noise (AWGN)), L is a number representing the delay spread of the channel 115, and M is the number of subcarriers used by the transmitter 105. CIRH_l(k) (0 ≦ L ≦ L-1) may be an independent complex-valued random variable with Gaussian distribution corresponding to different paths, which may represent a channel with frequency selective Rayleigh (Rayleigh) fading, where

The variation of the transmission channel 115 over time may be defined by the value f_dT_sIs characterized in that f_dIs the Doppler frequency and T_sIs the OFDM symbol set duration. Furthermore, the power spectrum of the rayleigh fading process (e.g., the time spread of the signal (or signal dispersion) and/or the time-varying behavior of the transmission channel 115) can be defined as:

where the correlation of CIR taps in the time domain (e.g., CIRs of a single path in a multi-path signal) can be characterized as:

wherein J₀(. is a first class of zeroth order Bessel function:

transmitter 105 includes a modulation unit 140, a serial-to-parallel (S/P) converter 150, processors 155 and 160, and a parallel-to-serial (P/S) converter 165. The modulation unit 140 is configured to receive a frequency domain signal 120 comprising information, e.g., information representative of a video image, and to modulate the signal 120 using one (or more) of several modulation schemes. For example, the modulation unit 140 may map the signal 120 to a constellation using a modulation scheme such as Quadrature Phase Shift Keying (QPSK), or Quadrature Amplitude Modulation (QAM) (e.g., 16QAM or 64QAM), although other modulation schemes are possible. The modulation scheme may be different for each subcarrier (e.g., one subcarrier may use 16QAM modulation and another QPSK modulation), or all subcarriers may use the same modulation scheme. Modulation unit 140 is configured to output a modulated signal 145 (e.g., X (0),..,. X (M-1), where M is the number of subcarriers (e.g., the number of samples in one OFDM symbol.) modulation unit 140 is configured to provide modulated signal 145 to S/P converter 150. S/P converter 150 is configured to convert signal 145 into parallel information signals 152 that can be provided to processor 155 on parallel streams the number of parallel streams may depend on the type of DVB system implemented (e.g., a 2K, 4K, or 8K system.) processor 155 is configured to perform an Inverse Fast Fourier Transform (IFFT) on signals 152 to convert signal 145 into time domain signals 157 that include output symbols X (0),..,. X (M-1). processor 160 is configured to add a cyclic prefix to the beginning of each output symbol generated by processor 155 and generate a cyclic prefix for frame X162 The prefix may duplicate a portion of frame x-1 and may be used to reduce the effects of inter-symbol interference (ISI). The P/S converter 165 is configured to convert the signal 162 into a serial transmission signal 125. The transmitter 105 is configured to broadcast a signal 125 via an antenna (not shown in fig. 2). Although the transmitter 105 has been described as including multiple pieces of processor and hardware, the functions provided by the transmitter 105 may be combined in a single chip, for example, having multiple software modules that perform respective tasks.

Referring again to fig. 3, signal 125 includes an OFDM frame 205 that includes a number of OFDM symbol sets 210. Each of these OFDM symbol sets 210 includes a plurality of symbols 215 modulated by a particular subcarrier frequency. Each OFDM symbol set 210 includes pilot symbols 220, data symbols 225, and TPS symbols 227, although other configurations are possible. Pilot symbols 220 include consecutive pilot symbols 221 modulated by the same subcarrier frequency in each of the OFDM symbol sets 210, and scattered pilot symbols 222 disposed at different subcarrier locations in different symbol sets 210. The scattered pilot symbols 222 are spaced 12 symbol positions apart such that every 12 th symbol position in the symbol set is occupied by a scattered pilot symbol. The occurrence of the continual pilot symbols 221 and scattered pilot symbols 222 may coincide (e.g., pilot symbols 223). Although OFDM frame 205 is shown in fig. 3 as including 12 OFDM symbol sets 210, other numbers of symbol sets 210 are possible. Further, the placement of the pilot symbols 220, data symbols 225, and TPS symbols 227 may be different from the arrangement shown in fig. 3.

Receiver 110 includes an S/P converter 170, processors 175, 180, 185, 190, and 191, and a demodulation unit 200. Receiver 110 is configured to receive signal 130 and output signal 135. Signal 130 includes time domain information (e.g., symbol y (0),.., y (M-1), where M is the number of subcarriers used to transmit signal 130). S/P converter 170 is configured to receive signal 130 and convert it to parallel signal 172. S/P converter 170 is configured such that parallel paths 172 of signal 172₁To 172_nEach of which includes symbols transmitted using a different subcarrier frequency. The number of parallel streams in signal 172 is equal to the number of symbols 215 in each OFDM symbol set 210, although other configurations are possible. The S/P converter 170 is providedIs arranged to provide a signal 172 to the processor 175. Processor 175 is configured to receive signal 172 and remove the cyclic prefix added by processor 160. Processor 175 is configured to provide parallel time domain signal 177 to processor 180. The processor 180 is configured to perform a Fast Fourier Transform (FFT) to convert the time domain signal 177 into a frequency domain signal 182. Processor 191 is configured to provide symbol synchronization such that a starting symbol of a set of OFDM time domain OFDM symbols may be obtained for use in a fourier transform.

The FFT operation performed by processor 180 may be defined as:

where Y (m) is signal 182, X (n) is signal 152, N (m) is AWGN depicted in the frequency domain, and H_l(m-n) is provided by the formula:

equation (6) can be rewritten in vector form as:

Y＝HX+N，(8)

wherein the elements in the mth row and nth column of the channel matrix H are:

processors 185 and 190 are each configured to receive signal 182 provided by processor 180. Processor 185 is configured to estimate the positions of the samples in the constellation of the selected modulation method using information provided by processor 190 (e.g., a final channel estimate as described below). For example, each of the parallel information streams in signal 182 includes modulation symbols (e.g., 16QAM, 64QAM, and/or QPSK) and processor 185 may map the modulation symbols to corresponding constellation points using the final channel estimates provided by processor 190. Although receiver 110 has been described as including multiple processors and other hardware, the functions provided by receiver 110 may be combined, for example, in a single chip having multiple software modules providing the respective functions.

For each set of OFDM symbols 210 received by receiver 110 from transmitter 105 via signal 130, processor 190 is configured to:

precomputation (or retrieval from memory)The IFFT of (1);

computing MMSE filter coefficients using the initial channel correlation estimate, interpolation, and a subset of the pilot symbols present in signal 182;

performing MMSE channel estimation using the calculated MMSE filter coefficients;

the result of the MMSE channel estimation and a single tap equalizer (contained in processor 190) are used to estimate signal 125.

Calculating the frequency domain variation of the transmission channel 115;

performing an ICI cancellation scheme to re-estimate the transmitted signal 125;

MMSE equalization is carried out on the version of signal 130 from which the estimated ICI has been subtracted using previously calculated MMSE coefficients;

using the final estimate of the signal 125 and the signal 130 to obtain an estimate of the transfer function of the transmission channel 115;

estimating and updating the channel correlation in the frequency domain;

update MMSE filter coefficients; and

provide the final channel estimate (e.g., transfer function) to processor 185.

Referring to fig. 2 and 3, processor 190 is configured to use a priori known pilot symbols as known anchor points for estimating characteristics of transmission channel 115. The receiver 110 is configured to expect pilot symbols at certain frequencies and/or times in each OFDM frame 205. Processor 190 is configured to search signal 130 for the pilot symbols that have been affected by transmission channel 115 (e.g., according to a transfer function as shown in equation (1)). Processor 190 is configured to calculate a channel estimate for transmission channel 115 using a subset (e.g., 4) of the scattered pilot symbols 222 received in each set 210 of OFDM symbols, preferably those surrounding and closest to the symbol 215 to be estimated. For example, to perform channel estimation on symbol 245, pilot symbols 246, 247, 248, and 249 may be used. Channel estimates for symbol 215 occurring near the lowest and highest frequency subcarriers (e.g., symbol 250 and symbol 255, respectively) may be calculated using less than 4 pilot symbols. For example, to calculate a channel estimate for symbol 253, pilot symbols 246 and 247 may be used, or pilot symbols 246, 247, 248, and 249 may be used. Although the subset of pilot symbols has been described as two and/or four scattered pilot symbols 222, other numbers are possible. Using a subset of the scattered pilot symbols 222, MMSE equalization can be applied in the frequency domain to provide a channel estimation algorithm that can be used for rapidly varying transmission channels and/or long transmission channels.

Referring again to FIG. 4, processor 190 includes modules 300 and 305, bus 310, and memory 315. Modules 300 and 305 may be, for example, software functions running on a processor, although other configurations are possible (e.g., separate pieces of hardware). Processor 190 is configured to receive a copy of each symbol 215 in each OFDM symbol set 210 received by receiver 110 and to output a final channel estimate to processor 185. The memory 315 is configured to store a copy of the symbol 215 received from the processor 180. Although processor 190 is shown to include bus 310, other topologies are possible (e.g., point-to-point connections). The processor 191 is configured to calculate a desired fourier transform window based on the output of the S/P converter 170. For example, the processor 191 is configured to exploit the property that the cyclic prefix of an OFDM symbol set is a repetition of the tail end of a preceding OFDM symbol set.

Module 300 is configured to perform initial channel correlation using pilot symbols 222 and interpolation. Correlation is a measure of how well the transfer function of the first subcarrier can be predicted based on a measurement of the transfer function of the second subcarrier. A higher correlation value between two subcarriers indicates a higher probability that the behavior (e.g., transfer function) of the first subcarrier can be predicted based on observations of the second subcarrier. Thus, a larger correlation value can indicate that the subcarrier frequencies have very similar transfer functions to each other, thereby permitting a higher degree of predictability between the two subcarriers. A smaller correlation value can indicate that the channels are independent, thereby reducing the degree of predictability between the two subcarriers. The module 300 is configured to calculate an initial correlation between two subcarriers using the assumed correlation according to:

where M is the total number of subcarriers, L is a number representing the delay spread of the channel 115, and L is the index of the channel path (of the transmission channel 115) representing the channel taps in the time domain. Equation (10) is preferably used when the receiver 110 has not received the previous set of OFDM symbols 210 (e.g., at startup), however equation (10) may also be used when the previous set of OFDM symbols 210 has been received.

For each symbol 215 in each OFDM symbol set 210, the module 300 is configured to determine the location of the 4 nearest pilots 222. To select the four pilots nearest to the subcarrier of interest, module 300 is configured to select the two closest scattered pilot symbols 222 with subcarrier indices higher than the subcarrier of interest and the two closest scattered pilot symbols 222 with subcarrier indices lower than the subcarrier of interest. For example, to calculate a channel estimate for symbol 245, module 300 is configured to select pilot symbols 246 and 247 (the two closest pilot symbols 222 with indices lower than the subcarrier of interest) and to select pilot symbols 248 and 249 (the two closest pilots 222 with indices higher than the subcarrier of interest).

For "edge" subcarriers (e.g., the subcarriers used to transmit symbol 215 between pilot 246 and pilot 247), module 300 is configured to use less than 4 pilots in the estimation process. For example, only one pilot symbol 222 is transmitted using a lower frequency subcarrier than symbol 253 (e.g., pilot symbol 246). Likewise, only one pilot symbol 222 is transmitted using a higher frequency subcarrier than symbol 254, i.e., pilot symbol 256 (assuming pilot symbol 256 is considered a "scattered" pilot symbol). The pilots used by module 300 to calculate channel estimates for the edge subcarriers may be symmetric or asymmetric. For example, to calculate a channel estimate for symbol 253, pilot symbols 246 and 247 may be used; pilot symbols 246 and 248; pilot symbols 247 and 248; or pilot symbols 248 and 249. Alternatively, module 300 may be configured to locate four (or more) pilots to be used to estimate the edge subcarriers. For example, module 300 may be configured to select pilot symbols 246, 247, 248, and 249 to calculate a channel estimate for symbol 253. Using two fewer pilot symbols than four to estimate the edge subcarriers may reduce the number of computations used in the estimation process.

Module 300 is configured to compute at least 12 sets of MMSE filter coefficients for each set of OFDM symbols 210 received by receiver 110. Each MMSE filter coefficient set corresponds to one of those symbols 215 that reside between pilot symbols 247 and 248, and one coefficient set corresponds to pilot symbol 248. Further, each MMSE filtering coefficient set corresponds to a respective symbol position (relative to the surrounding scattered pilot symbols 222). For example, a first set of MMSE coefficients corresponds to data symbols 225 with indices 1 higher than the indices of scattered pilot symbols 222 (e.g., the first set of MMSE coefficients corresponds to symbols 260, 270, 280,..) a second set of MMSE coefficients corresponds to data symbols 225 with indices 2 higher than the indices of scattered pilot symbols 222 (e.g., the second set of MMSE filtering coefficients corresponds to symbols 261, 271, 281,..), a third set of MMSE coefficients corresponds to data symbols 225 with indices 3 higher than the indices of scattered pilot symbols 222 (e.g., the third set of MMSE coefficients corresponds to symbols 262, 272, 282,.), and so on. One of these sets of MMSE coefficients (e.g., the twelfth set) corresponds to scattered pilot symbols 222. Module 300 is configured to calculate the twelve sets of MMSE filtering coefficients according to:

c(m)＝[c₁(m) c₂(m) c₃(m) c₄(m)]＝

[R((m-m₁)_M) R((m-m₂)_M) R((m-m₃)_M) R((m-m₄)_M)]R^-1，(11)

wherein (·)_MIs a modulo operation and

where R is preferably calculated each time receiver 110 receives one set 210 of OFDM symbols. However, R may also be computed multiple times for each received OFDM symbol set 210.

Module 300 is configured to calculate supplemental MMSE filter coefficients corresponding to subcarriers adjacent to the highest and lowest subcarrier frequencies used to transmit signal 125. The number of supplemental MMSE coefficients calculated by module 300 depends on the pattern of pilot symbols 220. The set of OFDM symbols 210 shown in fig. 3 has four different configurations for the scattered pilot symbols 222. For example, the configuration of scattered pilot symbols 222 is identical in OFDM frames 211 and 216, in OFDM frames 212 and 217, in OFDM frames 213 and 218, and in OFDM frames 214 and 219. The number of supplemental MMSE coefficient sets computed by processor 190 can depend on the number of data symbols 225 between the lowest frequency subcarriers and the lowest frequency scattered pilot symbols 222 (or the second lowest frequency scattered pilot symbols 222 when the lowest frequency continual pilot symbols 221 coincide with the lowest frequency scattered pilot symbols 222). For example, if the lowest frequency pilot symbol 222 was transmitted using the fourth lowest subcarrier, then four additional sets of MMSE coefficients are calculated. If the lowest frequency pilot symbol 222 was transmitted using the seventh lowest subcarrier, then seven additional sets of MMSE coefficients are calculated. Likewise, the number of supplemental MMSE coefficient sets computed by the processor depends on the number of data symbols 225 between the highest frequency subcarrier and the highest frequency scattered pilot symbol 222 (or the second highest frequency scattered pilot symbol 222 when the highest frequency continual pilot symbol 221 coincides with the highest frequency scattered pilot symbol 222). Equation 12 may vary depending on the pattern of the pilot symbols 220. For example, if two pilot symbols 220 are used to calculate the channel estimate for the edge subcarrier, equation 12 may be rewritten as:

where S is the spacing of the pilot symbols 220 used to calculate the channel estimate. For example, for OFDM frame 213, equation 12 may be rewritten as:

for OFDM frame 214, equation 12 may be rewritten as:

for OFDM frame 216, equation 12 may be rewritten as:

for OFDM frame 217, equation 12 may be rewritten as:

although module 300 has been described as calculating 4, 7, 10, or 13 supplemental MMSE coefficients, other numbers of coefficients may be calculated.

Using the above filter coefficients, the module 300 is configured to calculate a channel estimate for the subcarrier of interest according to:

whereinIs the channel estimate for the subcarriers used to transmit the selected subset of pilot symbols 220 (selected as described above):

and c is_p(m) is the corresponding MMSE filter coefficient.

Module 300 is configured to estimate a signal (e.g., signal 125) transmitted by transmitter 105 using the initial channel estimate obtained in equation (13). The module 300 is configured to estimate the signal transmitted by the transmitter 105 using the channel estimates for all the subcarriers used to transmit the signal 125. Module 300 is configured to calculate an estimate of signal 125 using single tap equalization according to:

whereinIs an estimated version of the transmitted signal (here signal 125), y (m) is the fourier transform of signal 130, andis the channel estimate obtained using equation (13).

The module 300 is configured to replace the received pilot symbols with the actual transmitted pilot symbols 220 and replace the received TPS symbols 227 with the estimated TPS symbols. Replacing the received pilot symbols 220 with actual pilot symbols and the received TPS symbols 227 with estimated TPS symbols may increase the accuracy of the channel estimates calculated by processor 190. For example, scattered pilot symbols 247 are affected by the transmission channel 115 (as described above). However, the value of the pilot symbol 247 is known a priori by the receiver 110. Thus, module 300 may replace the received version of pilot symbols 247 with the actual version of pilot symbols 247. The module 300 is further configured to replace the received TPS symbols 227 with estimated TPS symbols to establish an appended known reference point. To replace the received TPS symbols 227 with estimated TPS symbols, module 300 is configured to estimate the TPS symbols 227 by taking the average of all TPS symbols in a single set of OFDM symbols 210 according to:

the accuracy of the channel estimation process can be increased by increasing the accuracy of the channel correlation calculation using the substituted pilot symbols 220 and the averaged TPS symbols 227. Accordingly, the module 300 is configured to modify the initial channel correlation value calculated using equation (10). Module 300 is configured to correct the initial channel correlation by calculating another iteration of the channel estimate according to:

whereinNow an estimated version of the signal 125 with the substituted pilot symbols and the substituted TPS symbols. Module 300 is configured to calculate an estimate of the channel correlation (thereby replacing the assumption made in equation (10)) using:

wherein K is satisfied (K)₂-k₁)_MThe number of cases m. The channel correlation module 300 is configured to correlate the channelIs stored in the memory 315. The pilot symbols 220 and TPS symbols 227 are preferably used to calculate channel correlation because the channel estimates obtained using these symbols are more reliable than using data symbols. Processor 190 can calculate a more accurate channel correlation given a priori knowledge of pilot symbols 220. For example, processor 190 can calculate a relatively accurate channel estimate (at a particular time and frequency for each respective pilot symbol 220) by comparing the actual transmitted pilot symbol to the received version of the transmitted pilot symbol. Using a substantially accurate channel estimate for pilot symbols 220, processor 190 calculates a more accurate channel correlation valueThe probability is higher. Other channels (e.g., non-pilot or TPS symbols) may be chosen as reference points using, for example, signal-to-noise ratio as a selection criterion.

The channel correlation module 300 is configured to obtain an estimated averaged correlation of the channel 115 in the time domain according to:

where β is a predetermined known parameter (e.g., 1/16, 1/8) that is used to replace a portion of a previous correlation estimate. Increasing the number of symbol sets 210 used to calculate the average channel correlation results in a smoother and/or closer to actual channel correlation. The channel correlation module 300 is configured to correlate the channelStored in the memory 305.

To perform channel estimation for a rapidly varying transmission channel, receiver 110 should incorporate ICI in the channel estimation calculations. Inter-channel interference may occur when the transmission channel 115 is not constant within a single set of OFDM symbols 210. ICI can be accounted for in MMSE calculations by rewriting equation (6) to:

Y(m)＝H(m)X(m)+ICI(m)+N(m)，0≤m≤M-1，(19)

wherein

It is an FFT of the average CIR within a particular OFDM symbol, and

this is interference caused by time variations in the transmission channel 115.

The module 305 is configured to account for rapidly varying channels by accounting for ICI present in the signal 130. The module 305 is configured to estimate the frequency-domain ICI present in the signal 130 and to derive it from the estimated signal (e.g.,) And (4) deducting. Module 305 is configured to assume that the time domain variation of transmission channel 115 between two OFDM symbol sets 210 is linear, although other assumptions may be made. Using this assumption, module 305 is configured to calculate the difference between the channel estimate for the currently receiving set of OFDM symbols 210 and the channel estimate for the previously received set of OFDM symbols 210 according to:

whereinIs the channel estimate for the current OFDM symbol set 210,is the channel estimate of the previous OFDM symbol set 210 (e.g., retrieved from memory 315), and G is the length of the guard interval, which is known a priori.Andmay be calculated as described herein.

To understand the relationship between channel slope and ICI terms, it is helpful to reorganize equation (6) into the following equation:

whereinThe remainder including N (m) and | Q | > Q, where Q is the index of the subcarrier. Using equation (23), the module 305 is configured to represent the impulse response of the transmission channel 115 according to:

wherein

And alpha is_lIs the slope of the channel variation. Therefore, equation (22) can be simplified as:

wherein

(i.e., the amount of the acid,IFFT) that may be pre-computed (e.g., stored in memory during manufacture of receiver 110), and

(i.e., the amount of the acid,is the channel slope estimate in the frequency domain). The module 305 is configured to estimate the CIR slope according to:

whereinIs expressed in the frequency domainModule 305 is configured to provide the calculated information to module 300.

Module 300 is configured to estimate a transmitted signal (e.g., signal 125) using information (e.g., ICI information) provided by module 305. Module 300 is configured to estimate a transmitted signal (e.g., signal 125) by subtracting ICI information provided by module 305 from a received signal (e.g., signal 130) according to:

where Q is a small integer (e.g., 1 or 2, although other integers are possible).

Module 300 is configured to calculate further iterations of equation (30), which may improve the performance of the final channel estimate. Module 300 is configured to perform frequency domain MMSE channel estimation on the ICI subtracted signal according to:

additional iterations of ICI cancellation and/or additional MMSE equalization may be performed by blocks 300 and 305, if desired. In addition, the channel correlation calculated by the module 300 using equation (9) may be updated using equation (12). Module 300 is configured to calculate the final channel estimate using equations (30) and (31) according to the following equation:

the module 300 is configured toIs stored in the memory 315. For each subcarrier of interest, memory 315 can be configured to provide a final channel estimate to processor 185 using equation (32). Alternatively, memory 315 may be configured such that the final channel estimate obtained from equation (32) is stored in memory and retrieved by processor 185 as needed. Using information provided by processor 190, processor 185 is configured to compute a final estimate of signal 130 by mapping received individual symbols 215 to corresponding constellation points (e.g., QPSK, 16QAM, or 64QAM constellations).

A viterbi decoder may be used in the channel estimation process to increase the accuracy of the computed channel estimates and reduce inter-subcarrier ICI. For example, module 300 is configured to analyze the output of equation 23 using a viterbi algorithm. Module 300 is also configured to encode the viterbi-decoded bits according to, for example, the DVB-T/H standard to obtain an estimated transmitted signal (e.g., signal 125) represented by the constellation points. Module 300 is also configured to perform ICI cancellation using channel slope estimates obtained by analyzing, for example, hard and/or soft decisions with a viterbi algorithm. When using a viterbi decoder, the module 300 is configured to perform MMSE using the signal with ICI subtracted to obtain a supplemental channel estimate. The MMSE step using the ICI-subtracted signal can be omitted to reduce the complexity of the channel estimation process.

In operation, referring to fig. 5, and with further reference to fig. 2-4, a process 400 for performing channel estimation using system 100 includes the stages shown. The process 400 is, however, exemplary only and not limiting. The process 400 may be altered, for example, by adding, removing, or rearranging stages.

In stage 405, the receiver 110 receives a signal 125 (e.g., an OFDM DVB-T/H signal that has been converted to a signal 130) transmitted by the transmitter 105 via the transmission channel 115. For each OFDM symbol set 210 received by receiver 110, S/P converter 170 converts the OFDM symbol set 210 into parallel signal 172. Each stream (e.g., 172)₁To 172_n) Corresponding to serving to viaThe transmission channel 115 conveys different subcarriers of the signal 125. Processor 175 receives signal 172 from S/P converter 170. Processor 175 removes the cyclic prefix added by processor 160. Processor 175 provides signal 177 to processor 180. Processor 180 performs an FFT on signal 177 to transform it into frequency domain signal 182. Processor 182 provides each set of OFDM symbols 210 to processors 185 and 190.

At stage 410, processor 190 performs an initial channel correlation estimate. Module 300 calculates R for the received set of OFDM symbols 210 and stores R in memory 315. Module 300 calculates at least 12 sets of MMSE filter coefficients using equation (11), where each set of MMSE coefficients corresponds to a respective position of data symbol 225 relative to surrounding pilot symbols 320. Module 300 stores the MMSE filter coefficients in memory 315. Module 300 calculates the supplemental MMSE filter coefficients corresponding to the subcarriers adjacent to the lowest and highest frequencies used to transmit the signal 125. For frame 205, module 300 calculates 4, 7, 10, or 13 supplemental coefficients using equation (11), depending on the configuration of scattered pilot symbols 222 in symbol set 210.

At stage 415, module 300 determines a subset of pilots 220 to be used for calculating the channel estimate. For each of these data symbols 225 and TPS symbols 227, module 300 determines whether there are at least two scattered pilot symbols 222 between the subcarrier used to transmit the symbol 215 under view and the lowest and highest frequency subcarriers used to transmit the signal 125. If there are at least two scattered pilot symbols 222 between the subcarrier used to transmit the symbol 215 being viewed and the highest and lowest subcarriers, the module 300 selects four pilot symbols 222 for use in MMSE equalization. For example, module 300 selects the closest two of the scattered pilot symbols 222 having a lower subcarrier frequency than the symbol 215 being viewed and the closest two of the scattered pilot symbols 222 having a higher subcarrier frequency than the symbol 215 being viewed. For example, in calculating the channel estimate for symbol 245, module 300 selects pilot symbols 246, 247, 248, and 249. If the symbol 215 being examined has fewer than two scattered pilot symbols 222 between the subcarrier used to transmit the symbol being examined and the lowest or highest frequency subcarrier used to transmit the signal 125, then the module 300 selects the two closest pilots 222 (of the subcarriers having the lower or higher frequencies). For example, this module selects pilot symbols 246 and 247 when deciding which subset of pilots 222 to use in calculating the channel estimate for symbol 253. Module 300 performs MMSE equalization on each symbol 215 being viewed using equation (13) to obtain an initial channel estimate for each subcarrier being viewed. More or less than two pilots 222 may be used to calculate estimates of the edge channel.

At stage 420, module 300 computes an estimate of signal 125 using equation (14) and the initial channel estimate computed in stage 415. The module 300 replaces the estimated pilot symbols with the actual pilot symbols. Module 300 also improves the accuracy of the TPS symbol estimates by averaging all TPS symbols 227 in the set of OFDM symbols 210 being viewed using equation (15). Module 300 calculates a channel estimate for the transmission channel 115 using the estimate of the signal 125 (with the substituted pilot and TPS symbols) and equation (16). Module 300 calculates an updated channel correlation using equation (17), which is stored in memory 315. Module 300 also calculates the estimated average correlation using equation (18).

At stage 425, module 305 estimates the frequency-domain ICI present in signal 130 and extracts it from the estimated signal (e.g.,) And (4) deducting. Module 305 assumes that the time-variation between two OFDM symbol sets 210 is linear and calculates the difference between the channel estimate of the OFDM symbol set 210 being received and the channel estimate of the previously received OFDM symbol set 210. Module 305 calculates the difference between the current channel estimate and the previous channel estimate using equation (22). Module 300 performs MMSE channel estimation using ICI information provided by module 305 (e.g., either directly via bus 310 or indirectly through memory 315) with the ICI-subtracted estimation signal in equation (31). Additional iterations of ICI cancellation and/or additional MMSE equalization may be performed by blocks 300 and 305, if desired.

At stage 430, module 300 computes the final channel estimate for the transmission channel 115 and updates the channel correlation. Module 300 calculates the final channel estimate for the transmission channel 115 using equation (32).

At stage 435, process 190 determines whether another set of OFDM symbols 210 has been received by receiver 110. If so, process 400 returns to stage 405. Otherwise, process 400 terminates.

Other embodiments are within the scope and spirit of the invention. For example, due to the nature of software, the functions described above may be implemented using software, hardware, firmware, hardwiring, or a combination of any of these. Features implementing functions may also be physically located in various places, including being distributed such that portions of functions are implemented in different physical locations. For example, although fig. 4 shows several modules, each of which has been described as performing a particular function, the functionality provided by each module may be combined into a single module and/or separated into supplemental modules. The receiver 110 may be configured to process multiple channel estimates substantially simultaneously using parallel processing. Although memory 305 is shown as part of processor 190, other configurations are possible (e.g., memory 305 may be located in processor 185 or may be remotely located via a network connection).

Referring to fig. 5, process 400 may be modified to accommodate transmission channels in which the power distribution changes over time. For example, process 400 may include a stage where correlation is based on a current channel estimate and the supplemental MMSE is carried out before ICI cancellation is applied. Furthermore, the MMSE performed during stage 420 may be performed during stage 410, and the channel correlation estimation may be moved from stage 430 to stage 410.

Although the present invention has been described in the context of the DVB-T/H standard, in which the subcarriers used to transmit the scattered pilot symbols 222 are spaced 12 apart, other configurations are possible. For example, the present invention can be adaptively modified to work with a transmission standard in which the spacing of the subcarriers used to transmit the scattered pilot symbols 222 is 3.

Although processor 190 has been described as using scattered pilot symbols 222, other pilot symbols may be used. For example, processor 190 can employ continual pilots 221 even if a selected one of the continual pilots 221 does not coincide with scattered pilot symbols 222. For example, referring to fig. 3, symbol 250 may be used to calculate a channel estimate for symbol 251. However, using continuous pilot symbols 221 that do not coincide with scattered pilot symbols 222 increases the complexity of the calculations performed by processor 190 by, for example, adding supplemental terms to equations (11) and (12).

Although certain values and/or quantities have been described as being calculated using one or more equations, other configurations are possible. For example, rather than calculating a particular value, the value may be pre-calculated and retrieved from memory 315.

Although the present invention has been described in the context of digital television broadcasting, the present invention may also be used in other mobile or wireless channels such as cellular communications, satellite radio broadcasting, terrestrial radio broadcasting, wireless networking (e.g., WiFi), and the like.

Furthermore, while the description relates to the invention, the description may include more than one invention.

Claims

1. A receiver for receiving an Orthogonal Frequency Division Multiplexing (OFDM) Digital Video Broadcasting (DVB) signal comprising a set of OFDM symbols including data symbols and pilot symbols transmitted using a plurality of subcarriers, the OFDM DVB signal being transmitted to the receiver via a transmission channel, the receiver comprising:

an input module configured to receive the OFDM DVB signal via the transmission channel; and

a channel estimation module coupled to the input module configured to compute a channel estimate for the transmission channel by performing a Fourier transform on the sets of OFDM symbols to produce sets of transformed symbols in the frequency domain and by performing Minimum Mean Square Error (MMSE) equalization on the sets of transformed symbols using a subset of the pilot symbols in the OFDM DVB signal.

2. The receiver of claim 1, wherein the channel estimation module performs a single fourier transform on each received set of OFDM symbols.

3. The receiver of claim 1 wherein the subset of pilot symbols is substantially less than all of the pilot symbols in the set of OFDM symbols.

4. The receiver of claim 3, wherein the channel estimation module is configured to select the subset of pilot symbols to be the N closest pilot symbols transmitted using subcarriers having a frequency lower than a frequency used to transmit the viewed data symbol and the N closest pilot symbols transmitted using subcarriers having a frequency higher than a frequency used to transmit the viewed data symbol.

5. The receiver of claim 3, wherein the channel estimation module is configured to select the N closest pilot symbols transmitted using subcarriers having a frequency higher than a frequency used to transmit the viewed data symbols as the subset of pilot symbols.

6. The receiver of claim 3, wherein the channel estimation module is configured to select the N closest pilot symbols transmitted using subcarriers having a frequency lower than a frequency used to transmit the viewed data symbols as the subset of pilot symbols.

7. The receiver of claim 1, wherein the channel estimation module computes a channel estimate corresponding to each respective subcarrier used to transmit each of the sets of OFDM symbols.

8. The receiver of claim 1, wherein the channel estimation module is configured to compute channel estimates for more than one of the data symbols in parallel.

9. The receiver of claim 4, wherein N is equal to 2 and the channel estimation module is configured to perform the channel estimation for each of the data symbols according to:

wherein c is_pAre the coefficients of the respective filters and are,is a channel estimate for a respective pilot symbol in said subset of said pilot symbols, and p is an index value for a respective pilot symbol in said subset of said pilot symbols.

10. The receiver of claim 9, wherein the channel estimation module is configured to calculate filter coefficients according to:

c(m)＝[c₁(m) c₂(m) c₃(m) c₄(m)]＝

[R((m-m₁)_M) R((m-m₂)_M) R((m-m₃)_M) R((m-m₄)_M)]R^-1

wherein

And is

0≤m≤M-1，

Where M is the total number of subcarriers present in the OFDM DVB signal, L is an index representing each respective path of a multipath OFDM DVB signal, L is a number representing the delay spread of the channel 115, M is an integer value representing the number of intervening channels between two channels selected for their correlation, and

wherein h is_l(k) Is the channel impulse response of the ith path of the transmission channel at time k.

11. The receiver of claim 1, wherein the channel estimation module is configured to calculate an inter-channel interference (ICI) value related to the transmission channel in a frequency domain.

12. The receiver of claim 11 wherein the channel estimation module is configured to perform the channel estimation using the ICI value.

13. The receiver of claim 12 wherein the channel estimation module is configured to perform MMSE equalization after subtracting the ICI value from the received OFDM DVB signal.

14. The receiver of claim 1, wherein the channel estimation module is configured to calculate the channel estimate by estimating a channel correlation value according to:

wherein K is satisfied (K)₂-k₁) M-the number of cases of M and M is the total number of subcarriers present in the OFDM DVB signal.

15. The receiver of claim 14, wherein the estimated channel correlation value is averaged according to:

where β is a predetermined known parameter.

16. The receiver of claim 1, wherein the channel estimation module is configured to use a viterbi decoder.

17. A method for calculating a channel estimate for a transmission channel used to transmit an Orthogonal Frequency Division Multiplexing (OFDM) Digital Video Broadcasting (DVB) signal comprising a set of OFDM symbols including data symbols and pilot symbols transmitted using a plurality of subcarriers, the method comprising:

receiving the OFDM DVB signal at a receiver via a transmission channel;

calculating a channel estimate of the transmission channel in the frequency domain by

Performing a Fourier transform on the set of OFDM symbols to produce a set of transformed symbols in the frequency domain, an

Performing Minimum Mean Square Error (MMSE) equalization on the set of transformed symbols using a subset of the pilot symbols in the OFDM DVB signal.

18. The method of claim 17, wherein computing the channel estimate comprises computing the channel estimate by performing a single fourier transform on each received set of OFDM symbols.

19. The method of claim 17, further comprising selecting as the subset of pilot symbols the N closest pilot symbols transmitted using subcarriers having a frequency lower than a frequency used to transmit the viewed data symbol and the N closest pilot symbols transmitted using subcarriers having a frequency higher than a frequency used to transmit the viewed data symbol.

20. The method of claim 18, further comprising selecting as the subset of pilot symbols the N closest pilot symbols transmitted using subcarriers having a frequency lower than a frequency used to transmit the data symbol under view.

21. The method of claim 18, further comprising selecting as the subset of pilot symbols the N closest pilot symbols transmitted using subcarriers having a frequency higher than a frequency used to transmit the data symbol under view.

22. The method of claim 17, wherein computing channel estimates comprises computing channel estimates for more than one of the data symbols in parallel.

23. The method of claim 17, wherein computing the channel estimate comprises computing a channel estimate for a data symbol of interest as N equals 2 and according to the following equation:

24. The method of claim 23, wherein computing the channel estimate comprises computing filter coefficients according to:

c(m)＝[c₁(m) c₂(m) c₃(m) c₄(m)]＝

[R((m-m₁)_M) R((m-m₂)_M) R((m-m₃)_M) R((m-m₄)_M)]R^-1

wherein

And is

0≤m≤M-1，

25. The method of claim 11, wherein computing channel estimates comprises computing channel estimates corresponding to each respective subcarrier used to transmit the OFDM DVB signal.

26. The method of claim 11, further comprising calculating an inter-channel interference (ICI) value related to the transmission channel in a frequency domain.

27. The method of claim 26 wherein computing the channel estimate comprises computing a channel estimate for the transmission channel using the ICI value.

28. The method of claim 27 wherein computing the channel estimate comprises performing MMSE equalization after subtracting the ICI value from the OFDM DVB signal.

29. The method of claim 17, wherein computing the channel estimate further comprises computing the channel estimate by estimating a channel correlation value according to:

wherein K is satisfied (K)₂-k₁)_MThe number of cases M and M is the total number of subcarriers present in the OFDM DVB signal.

30. The method of claim 29, wherein computing a channel estimate further comprises averaging the estimated channel correlation values according to:

where β is a predetermined known parameter.

31. The method of claim 17, wherein computing the channel estimate further comprises performing viterbi decoding.