MXPA01001706A - Sampling offset correction in an orthogonal frequency division multiplexing system - Google Patents

Abstract

An Orthogonal Frequency Division Multiplexing (OFDM) receiver detects and corrects sampling offsets in the time domain. The OFDM receiver oversamples (62) a training sequence or symbol in a received OFDM signal, correlates (64) the oversampled training sequence with a stored copy (66) of a truncated version of the training sequence, locates a correlation peak (70), and derives a sampling offset (72) by calculating a difference in magnitude of correlation samples in the vicinity of the correlation peak.

Description

CORRECTION OF SAMPLING DEVIATION IN AN ORTOGONAL FREQUENCY DIVISION MULTIPLEXING SYSTEM DESCRIPTION OF THE INVENTION The present invention relates to the processing of orthogonal frequency division multiplexed signals (OFDM, according to its acronym in English). Orthogonal frequency division multiplexing (OFDM) is a robust technique to efficiently transmit data over a channel. The technique uses a plurality of subcarrier frequencies (sub-carriers) within a channel bandwidth to transmit the data. These sub-carriers are arranged for optimum bandwidth efficiency compared to more conventional transmission aspects, such as frequency division multiplexing (FDM), which spends large portions of the bandwidth of channel in order to separate and isolate sub-carrier frequency spectra and thus avoid inter-carrier interference (ICI). In contrast, although the frequency spectra of the OFDM subcarriers overlap significantly within the bandwidth of the OFDM channel, the OFDM however allows the resolution and retrieval of the information that has been modulated on each subcarrier. The transmission of data through a channel through OFDM signals provides several advantages over more conventional transmission techniques. An advantage is a tolerance to the spread of multiple path delay. This tolerance is due to the relatively long symbol interval, Ts, compared to the typical time duration of the channel impulse response. These long symbol intervals prevent intersymbol interference (ISI). Another advantage is a tolerance to selective frequency fading. Including redundancy in the OFDM signal, the data encoded on fading sub-bearers can be reconstructed from the data retrieved from the other sub-bearers. Another advantage is the efficient use of the spectrum. Since the OFDM sub-bearers are placed very close to each other without the need to leave a frequency space unused between them, the OFDM can efficiently fill a channel. One more advantage is the simplified sub-channel equalization. OFDM shifts the equalization of the time domain channel (as in single carrier transmission systems) to the frequency domain, where a bank of single equalizers of a bypass can individually adjust the phase and amplitude distortion of each sub-channel . Another advantage is the good interference properties. It is possible to modify the OFDM spectrum to represent the energy distribution of an interference signal. Also, it is possible to reduce out-of-band interference by avoiding the use of OFDM sub-bearers near the edges of the channel bandwidth.

Although the OFDM exhibits these advantages, the prior art implementations of OFDM also exhibit several difficulties and practical limitations. One difficulty is the emission of synchronizing the sample rate of the transmitter with the sample rate of the receiver to eliminate the sampling velocity deviation. Any mismatch between these two sampling rates results in a rotation of the constellation of the ari 2m sub symbol from symbol to symbol in a frame for smaller frequency deviations. However, for higher frequency deviations, the result is a contraction or expansion of the frequency spectrum of the received signal. Both points contribute to elevated BER. One cause of the sampling rate deviation is the presence of a sampling frequency deviation. A sampling frequency deviation occurs when the receiver samples the received signal at a frequency that is both higher and lower than the sample rate used in the transmitter. Another cause of the sampling rate deviation is the presence of a sampling phase deviation. A sampling phase deviation occurs when the receiver samples the received signal at a phase deviation from the transmitter's sample rate. Both the sampling frequency and the sample phase deviations can be dangerous for receiver operation, and must be corrected in order for the receiver to be properly synchronized. The present invention is directed to the correction of this problem. An orthogonal frequency division multiplexing (OFDM) receiver detects and corrects sampling deviations in the time domain. The OFDM receiver oversamples a training sequence or symbol in the received OFDM signal, correlates the oversampled training sequence with a stored copy of a truncated version of the training sequence, locates a correlation peak, and derives a sampling deviation calculating a difference in magnitude of correlation samples near the correlation peak.

BRIEF DESCRIPTION OF THE DRAWINGS In the drawings: Figure 1 is a block diagram of a conventional OFDM receiver; Figure 2 illustrates a typical arrangement of OFDM symbols and their corresponding security intervals within a data frame; Figure 3 is a block diagram of a sample deviation correction system illustrative of the present invention; Figure 4 is an illustration of a correlation energy peak when there is a sampling deviation (phase and / or frequency); Figure 5 is a block diagram illustrating the present invention as integrated with the conventional OFDM receiver of Figure 1; Figure 6 is a diagram of an illustrative training sequence in the frequency domain; and Figure 7 is a time domain representation of the training sequence of Figure 6. The features and advantages of the present invention will be more apparent from the following description, given by way of example. Referring to Figure 1, the first element of a typical OFDM receiver 10 is an RF receiver 12. Many variations of the RF receiver exist and are well known in the art, but typically, the RF receiver 12 includes an antenna 14, a low noise amplifier (LNA) 16, an RF bandpass filter 18, an automatic gain control circuit (AGC) 20, an RF mixer 22, a RF carrier frequency local oscillator 24, and a pass filter band IF 26. Through antenna 14, RF receiver 12 couples the OFDM modulated carrier, RF, after it passes through the channel. Then, by mixing it with a frequency receiver carrier fcr generated by the local oscillator RF 24, the RF receiver 12 converts the OFDM modulated carrier, RF, to obtain a received OFDM IF signal. The difference in frequency between the receiving carrier and the transmitting carrier contributes to the deviation of bearer frequency, delta fe.

This received OFDM IF signal is coupled to the mixer 28, and the mixer 30 will be mixed with an IF signal in phase and an IF signal phase shifted to 90 ° (quadrature), respectively, to produce OFDM signals in phase and quadrature, respectively. The IF signal in phase that is fed to the mixer 28 is produced by a local oscillator IF 32. The signal IF of phase shifted to 90 ° which is fed to the mixer 30 is derived from the IF signal in phase of the local oscillator IF 32 passing the IF signal in phase through a 90 ° phase shifter 34 before it is supplied to the mixer 30. The phase and quadrature OFDM signals then pass to analog to digital converters (ADCs) 36- and 38, respectively, in where they are digitized at a sampling rate fck_r as determined by a clock circuit 40. Analog to digital converters 36 and 38 produce digital samples that form a phase OFDM signal and quadrature discrete time, respectively. The difference between the sampling rates of the receiver and those of the transmitter is the sampling velocity deviation, delta fck = fck_r-fck_t. OFDM signals not filtered in phase and discrete-time quadrature of the analog-to-digital converters 36 and 38 then pass through digital low-pass filters 42 and 44, respectively. The output of the low pass digital filters 42 and 44 is filtered in phase and quadrature samples, respectively, of the received OFDM signal. In this way, the received OFDM signal is converted to phase (qi) and quadrature (pi) samples that represent the real and imaginary value components, respectively, of the complex value OFDM signal, ri = qi + jpi . These phase and quadrature samples (of real value and imaginary value) of the OFDM signal received afterwards are sent by the DSP 46. Note that in some conventional implementations of the receiver 10, the analog to digital conversion is done before the process of mixed IF. In such implementation, the mixing process involves the use of digital mixers and a digital frequency synthesizer. Also note that in many conventional implementations of receiver 10, digital-to-analog conversion is performed after filtering. DSP 46 performs a variety of operations on the phase and quadrature samples of the received OFDM signal. These operations may include: a) synchronizing the receiver 10 for time control of the symbols and data frames within the received OFDM signal, b) removing the cyclic prefixes from the received OFDM signal, c) calculating the discrete transformation Fourier (DFT) or preferably Fast Fourier Transform (FFT) of the received OFDM signal in order to recover the frequency domain subsymbol sequences that were used to modulate the subcarriers during each interval of the OFDM symbol; ) performing any channel equalization required on the subcarriers, e) calculating a sequence of subsymbols of frequency domain, and k, of each symbol of the OFDM signal by demodulating the subcarriers of the OFDM signal through the calculation of FFT. DSP 46 then sends these subsymbol sequences to a decoder 48. The decoder 48 retrieves the transmitted data bits from the subsymbol sequences of sequence domains that are delivered to it from the DSP 46. This retrieval is performed by decoding the domain subsymbols. of frequency to obtain a stream of data bits, which ideally must match the data bit stream that was fed to the OFDM transmitter. This decoding process may include soft decoding of Viterbi and / or Reed-Solomon decoding, for example, to retrieve convolutionally encoded block data and / or subsymbols. In a typical OFDM data transmission system, such as one for implementing digital television or a wireless local area network (WLAN), the data is transmitted in the OFDM signal in groups of symbols known as data frames. This concept is shown in Figure 2, where a data frame 50 includes M consecutive symbols 52a, 52b,. . ., 52M, each of which includes a security interval, Tg, as well as the OFDM symbol interval, Ts. Therefore, each symbol has a total duration of Tg + Ts seconds. Depending on the application, the data frames can be transmitted continuously, such as in digital television broadcasting, or the data frames can be transmitted in random time in bursts, such as in the implementation of a WLAN. Referring now to Figure 3, an illustrative embodiment of the present invention is shown. The arrangement of Figure 3 can be employed in the receiver of Figure 1, as illustrated in Figure 5. However, the present invention is illustrated as a different sampling deviation correction loop for clarity, ease of reference, and to facilitate an understanding of the present invention. The present invention operates on a receiver conforming to the ETSI-BRAN wireless LAN standards HIPERLAN / 2 (Europe) and IEEE 802.11a (E.U.A.), incorporated herein by reference. However, it is considered within the skill in the art to implement the teachings of the present invention in other OFDM systems. The previously identified wireless LAN standards propose the use of a training sequence for the detection of OFDM transmissions. The training sequence (e.g., training sequence A or B) includes a series of short OFDM training symbols (having known amplitudes and phases) that are transmitted over a predetermined number of subcarriers or pilot tanks (e.g. 12 pilot sub-carriers). All other sub-carriers (for example, 52 sub-carriers) remain at zero during the transmission of the training sequence. Although the use of the training sequence of LAN standards previously identified is discussed, the use of alternative training sequences and symbols is considered within the scope of the invention as defined by the appended claims. The frequency domain and time domain representations of the illustrative training sequence B of HIPERLAN / 2 are shown in Figures 5 and 6. As illustrated in Figure 6, the training sequence has a block of 16 samples that they are repeated four times per training symbol. This repetitive block or period of time is used by the present invention, as discussed in detail below. Referring now to Figure 3, a sampling deviation correction system 60 is shown. It should be noted that the system 60 may be modalized in software, hardware, or some combination thereof. A pair of samplers (e.g., ADCs) 62 and 78 sample a received OFDM signal. As discussed above, the received OFDM signal contains portions in phase (qi) and in quadrature (pi) representing the components of real value and imaginary value, respectively, of the complex value OFDM signal, ri = qi + jpi . The sampler 78 samples the OFDM signal at a given sampling rate (selected to be close to the transmitter sampling rate) and passes the sampled OFDM signal through a sampling rate converter 76 for downstream processing ( for example, FFT, and the like), as discussed in detail later. The sampler 62 samples in ascending or sub-sampling the received OFDM signal by a predetermined factor (e.g., a factor of 2) and passes the sampled signal upwardly to a correlator module 64. Oversampling of the received OFDM signal provides a resolution of the OFDM signal that is necessary to derive an important error, as discussed in detail later. It should be noted that sampler 78 and sampler 62 can be interconnected in a number of different ways, as is well known to those skilled in the art. For example, the sampler 78 and the sampler 62 can be driven by a clock circuit (not shown) that activates both samplers 78 and 62 to oversample the OFDM signal by a factor of two. In this case, the sampler 62 can pass each sample to a correlator module 64, and the sampler 78 can pass each sample to the sampling rate converter 76. The correlator module 64 correlates the sampled signal received in ascending form from the sampler 62 with samples time domain of the training sequence (e.g., training sequence B of the aforementioned wireless standards) stored in a local memory 66. Each sample in the illustrative training sequence has a value of sqrt (13/6) * [(1 + j) or (-1-j)]. The allocated memory for storing each sample value will depend on the design of a particular OFDM receiver. The stored version of the training sequence, preferably, is a truncated version of the training sequence corresponding to one of the repetitive blocks of samples (eg, 16 samples) of training sequence B. More specifically, the stored version of the truncated training sequence preferably corresponds to an over sampled version (eg, 32 samples) of the repetitive block that is oversampled by the same predetermined factor (e.g., a factor of two) as used in the sampler 62. Only by storing a rather oversampled, truncated version of the training sequence, memory space is efficiently used in local memory 66, since the entire training sequence (ie, 64 samples if the training sequence is not oversampled) is not stored in local memory 66. There will be a maximum correlation between the OFDM signal on samples processed and the truncated version of the training sequence when the stored training sequence coincides with a training sequence contained in the OFDM signal. In this way, a peak in the energy of the correlation output can be used to determine when the received signal matches the stored training sequence. The output of the correlator module 64 is a complex signal, since the inputs (i.e., the stored training sequence and the OFDM signal) are complex. The energy module 68 can calculate the energy or magnitude of each sample of the correlated signal in one of two ways according to the design of a particular OFDM receiver. First, the energy module 68 can calculate the square magnitude (i.e., energy) of each complex sample of the correlated signal to generate a real number indicating the energy of the correlated signal. Secondly, the energy module 68 can obtain the magnitude (opposite to the square magnitude) of each complex sample of the correlated signal. A peak locator module 70 searches for the output of the correlation energy sequence of the energy module 68 in order to locate the sample in the correlation energy sequence having the largest energy or magnitude value. Once the largest value is identified, the peak location locator module 70 outputs the peak location index to an error computation module 72. The index is used by the error computation module 72 as a point of reference. As discussed above, the oversampling of the OFDM signal increases the number of correlation samples so that the error computation module 72 can derive a significant sampling error. For example, Figure 4 shows a main correlation peak 80 and a pair of smaller correlation peaks 82 and 84 on both sides of the main correlation peak 80. If the OFDM signal was not oversampled by the sampler 62, probably only the main correlation peak 80 may be present and an error computation module 72 may not be able to determine a sampling error 86 derived from the magnitude of correlation peaks near the main peak 80, as discussed in detail below. When the main peak of the correlation samples is detected, the error computation module 72 analyzes the correlation samples 82 and 84 on both sides of the main peak 80. When there is no sampling deviation, the frequency correlation samples 82 and 84 will have the same magnitude (not shown). However, if there is a sampling deviation, the correlation samples 82 and 84 will have different magnitudes, as shown in Figure 4. The computation module 72 calculates an error value by calculating the difference in magnitude between the correlation samples. and 84 on either side of the correlation peak 80. The difference in magnitude can be positive or negative. The magnitude of the difference indicates the degree to which the stored training sequence and the received training sequence are out of sync. The sign of the difference indicates whether the sampling frequency is increased or decreased. For a given sampling deviation, the magnitude of the sample to the left of a main correlation peak (for example, main peak index-1) minus the sample value to the right of the main correlation peak (for example, index main peak +1) will produce the error value. Alternatively, the error value can be calculated as the difference between the right sample and the left sample depending on the requirements of a particular system.

Returning to Figure 3, the error computation module 72 outputs the calculated error value to a second order loop filter 74 that adjusts the sampling rate, so that the sampling error is directed toward zero and the velocity Sampling of the receiver is synchronized with the sampling rate of the transmitter. More specifically, the second order loop filter 74 adjusts the sampling rate of a sampler 78 through a conventional sampling rate converter 76 or, in an alternative, can adjust the sampling rate of the sampler 78 and associated up sampler 62. Referring now to Figure 5, an integration of the present invention and the conventional OFDM receiver 10 of Figure 1 is shown. More specifically, the sampling deviation correction system 60 can be coupled to the outputs of the mixers 28 and 30 and the DSP inputs 46. With this arrangement, the sampling deviation correction system 60 receives the phase and quadrature OFDM signals from the mixers 28 and 30, digitizes the received signals at a corrected sampling rate which matches the sampling rate of the transmitter, and outputs the digitized signals to DSP 46 for further processing. It should be noted that LPF 42 and LPF 44 of Figure 1 can be coupled to the outputs of the sampling deviation correction system 60 and to the inputs of the DSP 46 for filtering the digitized OFDM signals, although said arrangement is not shown in FIG. Figure 5 Thus, in accordance with the principle of the present invention, a method for correcting a sampling deviation in an OFDM receiver is provided. The method includes sampling a received OFDM signal, the OFDM signal containing a reference symbol, correlating the sampled OFDM signal with a stored symbol, locating a correlation peak, calculating a difference in magnitude of correlation samples over any of the correlation peaks, and deriving a sampling deviation error from the calculated difference. Although the present deviation has been described with reference to preferred embodiments, it is clear that various changes can be made in the embodiments without departing from the spirit and scope of the invention, as defined by the appended claims.

Claims (20)

1. - A method for correcting a sampling deviation in an orthogonal frequency division multiplexing (OFDM) receiver, the method comprising - the steps of: sampling a received OFDM signal containing a reference symbol; correlate the sample OFDM signal sampled with a stored symbol; locate a correlation peak; calculate a difference in magnitude of correlation samples on either side of the correlation peak; and correct a sampling deviation in response to the calculated difference.

2 - The method according to claim 1, wherein the step of sampling includes oversampling the received OFDM signal through a predetermined factor.

3. The method according to claim 2, wherein the step of correlating includes correlating the oversampled OFDM signal with a stored symbol that is oversampled by the predetermined factor.

4. The method according to claim 3, wherein the predetermined factor is a factor of 2.

5. The method according to claim 1, wherein the stored symbol is identical to the reference symbol.

6. - The method according to claim 1, wherein the stored symbol corresponds to a segment of the reference symbol.

7. The method according to claim 6, wherein the stored symbol is a segment that is periodically repeated within the reference symbol.

8. The method according to claim 1, wherein the sampling and correlation steps occur in the time domain.

9. - The method according to claim 1, wherein the step of correlating includes the steps of: outputting a sequence of correlation samples representing the correlation of the stored symbol with the OFDM signal; and determine the energy of each correlation sample in the sequence.

10. The method according to claim 9, wherein the step of locating a correlation peak includes the step of determining the correlation peak index by placing a correlation sample in the sequence of correlation samples having an energy value maximum.

11. The method according to claim 9, wherein the step of determining the energy of each correlation sample includes the step of calculating a square quantity of each correlation sample.

12. The method according to claim 9, wherein the step of determining the energy of each correlation sample includes the step of obtaining a magnitude of each correlation sample.

13. ~ An orthogonal frequency division multiplexing (OFDM) receiver for receiving an OFDM signal having a training symbol, the OFDM receiver characterized by having: an analog-to-digital converter (ADC) sampling a received OFDM signal at a sampling rate to generate OFDM samples including a plurality of training symbol samples; a training symbol detector coupled to the ADC, the training symbol detector detecting the location of the training symbol samples within the OFDM samples; a sampling deviation unit coupled to the training symbol detector, the sampling deviation unit generating a sampling deviation error based on a comparison of predetermined training symbol samples of the plurality of training symbol samples.

14. The OFDM receiver according to claim 13, further comprising: a correction unit coupled to the ADC and the sampling deviation unit, the correction unit adjusting the sampling rate of the ADC in response to the output of sampling deviation error of the sampling deviation unit.

15. The OFDM receiver according to claim 13, wherein the training symbol detector outputs an index of a training symbol located and the sampling deviation unit calculates a difference in magnitude of symbol samples of training. training on either side of the index.

16. The OFDM receiver according to claim 13, wherein the training symbol detector comprises: a correlator coupled to the ADC, the correlator correlating the OFDM sample output of the ADC with a stored copy of the training symbol for generate a plurality of correlation samples; and a peak correlation detector coupled to the correlator, the peak correlation detector outputting an index of a correlation peak in response to the detection of a correlation peak in the plurality of correlation samples.

17. The OFDM receiver according to claim 16, wherein the stored copy of the training symbol is a training symbol segment of the OFDM signal that is periodically repeated within the training symbol of the OFDM signal.

18. The OFDM receiver according to claim 16, wherein the ADC oversamples the received OFDM signal by a predetermined factor, so that a predetermined resolution of correlation samples is generated by the correlator.

19. A system for identifying a sampling deviation in an orthogonal frequency division multiplexing (OFDM) receiver, the system comprising: means for oversampling an OFDM signal received by a predetermined factor, the OFDM signal containing a symbol of reference; means for correlating the oversampled OFDM signal with a stored symbol to generate a sequence of correlation samples, the stored symbol being sampled by the predetermined factor before storage and corresponding to a segment of the reference symbol; means for locating a correlation peak within the sequence of correlation samples; means for calculating a difference in magnitude of predetermined correlation samples near the correlation peak; and means for deriving a sampling deviation error from the calculated difference.

20. The system according to claim 18, wherein the system is incorporated into an OFDM receiver operating in a wireless LAN.