CN1336001A - Synchronous pulse generating method and OFDM signal receiving method - Google Patents

Abstract

An OFDM receiver generates a synchronization pulse that defines a fast Fourier transform window by examining the output of a correlator to find the most correlated sub-interval between samples of a symbol separated by the length of the active portion of the symbol. Synchronization pulses are generated between the subintervals (102). The timing adjustment of the synchronization pulse is performed only when the current error is an intentional and continuous error. A signal indicative of the amount of adjustment is used to determine the phase rotation applied to the output of the FFT circuit.

Description

Synchronous pulse generating method and OFDM signal receiving method

(technical field)

The present invention relates to OFDM modulation, and more particularly to the generation of synchronization pulses representing OFDM symbol boundaries used, for example, in Fourier transform demodulation.

(Background technique)

OFDM systems are well known. Various techniques are used in the synchronization of OFDM receivers. Some of these techniques must transmit special synchronization signals. Other techniques rely on standard OFDM signals where the entire symbol consists of an active part and a guard interval, often called a guard interval, period extension, or period taper.

The guard interval is located in front of the valid part of the symbol, including the repetition of data at the end of the valid part. (This is the same as placing a guard interval including the same data as that at the beginning of the valid part after the valid part).

Synchronization techniques that rely on repeated data within guard intervals typically operate by correlating each other using complex samples separated by only the length of the significant part of the symbol. From this, timing pulses for the Fourier transform of the received signal are generated. The timing of the pulses is the timing at which the Fourier transform window only contains data from a single symbol.

When the timing is not suitable, inter-symbol interference (ISI) occurs. However, by using a guard interval, ISI is avoided while allowing for some amount of variation in the timing of the pulses. The guard interval must be longer than the width of the longest predicted delay between signals received via different paths. The guard interval is relatively small compared to the active part of the signal, typically the guard interval includes Nu/32, Nu/16, Nu/8 or Nu/4 samples, where Nu is the number of samples in the active part of the symbol.

There are various techniques for deriving sync pulses from correlation with each other. These may operate appropriately under normal reception conditions, but occur at undesired timing and pulses and cause ISI.

In the absence of noise or multi-path interference, the cross-correlator produces an output equivalent to setting the average value to zero over the significant portion of the symbol being equivalent, outside the guard-interval samples and intervals in which the samples are correlated with each other. The intercorrelator produces a high level output during this interval. This high level output ends with the end of 1 symbol and the beginning of the next signal. One of the original devices, after integrating the output of the correlator, detects the peak of the resulting signal and generates a timing pulse at the end of each symbol.

In the case of multi-path interference where the same signal is received with different delays, to avoid ISI, the synchronization pulse must occur in a window with equal width on the repeated part between the guard intervals of the 2 received signals. However, the mutual correlator produces an intended output by processing the interval of one or both of the samples of the guard interval by the mutual correlator. Depending on the circumstances, the result is timing pulses delivered outside the most appropriate window, which in turn leads to ISI.

European patent application EP0772332 describes other techniques for generating synchronization pulses. One of the techniques it discloses is that the output of the correlator is dependent on the transmission to a phase locked loop (PLL). Even then, in the case of intentional noise or multipath interference, the result is that the synchronization pulses are generated outside the optimum window. Also, there must be a substantial number of symbol periods on the PLL to achieve lock, so the collection time is relatively long.

Other problems with the original arrangement arise from, for example, varying degrees of phase rotation induced by complex values in the frequency bins of the FFT output as the sync pulse is adjusted as a result of changes in signal state. Subsequent channel estimators and collectors can handle these changes, thus further increasing the collection time and requiring considerable processing power.

Accordingly, it would be desirable to provide techniques for generating synchronization pulses that circumvent, or at least alleviate, the above-mentioned problems.

(Summary of Invention)

The aspects of the invention are described within the scope of the appended claims.

That is, the present invention is a method of generating a synchronous pulse representing a symbol boundary of an OFDM signal comprising valid symbol periods separated by guard intervals, comprising the steps of combining data in each guard interval with a part of data in each valid period Correspondingly, a signal is provided representing the degree of correlation between samples of the received signal separated by a period corresponding to the significant part of the symbol, whereby the signal provides an output in each symbol representing the interval of correlation identified as intended, and also includes the decision A step in which a sub-interval of maximum correlation occurs to supply a synchronization pulse within that sub-interval.

In another case, the synchronization pulse is generated by supplying a signal representing the degree of correlation between samples of the received signal separated by a period corresponding to the significant part of the symbol, and then supplying a signal representing the interval between which the intended correlation is confirmed. Outputting, the synchronizing pulse generating method includes the step of determining a sub-interval in which a maximum degree of correlation occurs to supply a synchronizing pulse in the sub-interval.

In the case of multi-path interference, the degree of correlation is judged to be maximum after the width passes through the repetition period corresponding to the guard interval. This is the most appropriate period to generate the sync pulses because each Fourier transform window is guaranteed to contain samples from only a single symbol even when the same symbol is received with different delays. Using the technique of the present invention, the output of the cross-correlator is examined to determine when the most appropriate period described above occurs.

In a preferred embodiment, the output of the mutual correlator is compared to a threshold value, and an optimum sub-interval specifying the period in which the synchronization pulse should be generated is represented by the period in which the output of the mutual correlator exceeds said threshold value. Preferably, the threshold value is changed corresponding to the outputs of the correlators, and more preferably, the threshold value is specified based on the highest level of the correlator output.

As an independent aspect of the invention, the use of thresholds can be seen. In this aspect, it is considered to apply to the level detector the output of the correlator representing the degree of correlation between samples of the received signal separated by a prescribed number of sample intervals, only signals exceeding a prescribed (preferably variable) level Part of the time at which other units should generate sync pulses.

Timing pulses can also be generated at any time within the window where it is desired to represent maximum correlation, for example in the middle of the window. As signal conditions change, the point in time moves, and the sync pulse pairs respond to this change. However, in a preferred embodiment, the timing of the sync pulses is only changed when certain conditions are met. For example, the timing is changed only when it is confirmed that the current timing has an error of only a prescribed number of hours and/or only when the current error exceeds a prescribed amount. As an independent invention, by this technique, the phase rotation of each carrier of the output of the FFT is produced at a different angle for each change, thereby avoiding the Fourier transform that generates repetitive workloads on the original channel estimator The timing of actions varies too much.

In another aspect of the invention, means for applying a different phase rotation to each sample of the FFT output is provided on the output of the FFT, the means corresponding to the signal representing the amount of shift the pulse generator applies to the synchronization pulse, other The amount of phase rotation applied by the cell. Thus, timing variations of the sync pulses can be compensated for very rapidly depending on the situation or instant. The phase rotation is applied by the FFT and a circuit arranged between the channel estimator and the collector, or the phase rotation is applied by the channel estimator and the collector. Preferably, the timing variation of the sync pulses is formed to occur less frequently (according to the above-mentioned conditions), more preferably, only a predetermined amount is usually generated. Thereby, the determination of the appropriate phase rotation to apply to the FFT output is facilitated. These phase rotations may be calculated corresponding to signals representing actual or predicted degrees of timing of the sync pulses. Based on such a signal, it can be derived from an address-specified look-up table.

In the case of the prior art described, the outputs of the correlators are filtered using, for example, a sliding window averaging arrangement which sums the latest Ng samples from each other's correlators. However, this filtering technique changes the shape of the output of the correlator, so in the preferred embodiment of the invention, the outputs of the mutual correlators are filtered by adding the samples of the most recent L1, but L1 is much smaller than Ng.

It is a known technique to take the output of the sliding window averaging means and process it to provide a slightly frequency corrected signal. This technique is preferably also used in the device according to the invention. However, in order to achieve more accurate frequency estimation, in the preferred embodiment of the present invention, a second filter is added to the output of the first filter to produce an output representing the addition and average of the number of samples much more than L1 . For example, the output may be equivalent to that which would result from a single filter summing the latest Ng samples.

(Brief description of the attached picture)

Fig. 1 is a block diagram of an OFDM receiver according to the present invention;

FIG. 2 is an explanatory diagram schematically showing an OFDM signal;

Figure 3 is a block diagram of an existing device for generating synchronization pulses;

Fig. 4 is an explanatory diagram briefly showing the effect of applying multi-path interference on mutual correlation outputs;

Fig. 5 is the block diagram of synchronous circuit of the present invention;

6 is a block diagram of a timing recovery circuit forming part of a synchronization circuit;

FIG. 7 is an explanatory diagram showing a portion of a standard waveform derived from a correlation output, the portion occupying a subinterval within which a sync pulse is most likely to be generated;

FIG. 8 is a graph showing changes in three values of Ns, L1, and L2.

(Detailed description of the invention)

Referring to FIG. 1, an OFDM receiver 2 is configured with an antenna 4 that receives a signal and transmits the signal to a down converter 6 that converts the RF signal into an IF signal. Then it is converted into a baseband signal by the IF baseband converter 8 . The frequency converter produces at its output complex samples of the transmitted OFDM symbols. These complex samples are digitized by an analog-to-digital (A/D) converter 10 and sent to a fast Fourier transform (FFT) circuit 14 via a frequency trimming circuit 12 . FFT circuit 14 changes the sampling from time domain to frequency domain, and sends the output symbol data to phase rotator 15 , channel estimator and collector 16 and decoder 17 .

The techniques in accordance with the present invention facilitate the implementation of positive feedback systems that do not rely on feedback or phase locked loops for adjusting the local oscillator frequency. However, in other constructions, where it is desired to configure the aforementioned feedback, synchronization circuit 18 may be responsive to complex sampling from, for example, analog-to-digital (A/D) converter 10 and/or signals from channel estimator and collector 16 .

The composite samples are sent to symbol synchronization circuit 20 which generates a frequency offset signal for frequency trim circuit 12 and a synchronization pulse used by fast Fourier transform (FFT) circuit 14 . There must be a synchronization pulse on the FFT circuit 14, so that each transformation action is consistent with the start position of the OFDM symbol.

The circuits described above, except for the phase rotator 15, are known in the prior art. The present invention is particularly relevant to the new, inventive technique employed by symbol synchronization circuit 20 .

Referring to Fig. 2(a), it is assumed that an OFDM symbol consists of Nu+Ng samples representing Nu samples in the effective part U of the signal, starting with Ng samples in the guard interval G. The Ng samples within the guard interval G (each symbol is hatched) contain the same data as the last Ng samples of the active part U of the symbol.

Referring to FIG. 3 , in a prior art synchronous circuit, complex samples from the IF baseband converter 8 are fed consecutively to a first-in-first-out (FIFO) register 30 of a mutual correlator 28 . Since this register includes Nu parts, it brings a corresponding Nu sample delay. The output of register 30 is sent to complex conjugate circuit 32 of correlator 28 which transforms each sample to its complex conjugate. Next, each complex conjugate is multiplied by the undelayed samples from the analog-to-digital (A/D) converter 10 by the multiplier 34 of the correlator 28 (or the complex conjugate circuit 32 is inserted in another path of the multiplier 34 ).

The output of the correlator is high each time the complex conjugate of the delayed samples within the guard interval G is multiplied by samples of equal value derived from the end of the subsequent active part U of the symbol. At other times, the output of the correlator uses random values. Figure 2(b) shows the output of the correlator. In fact, the summing and averaging are carried out in a later stage. For better understanding, Fig. 2(b) shows the optimum output after summing and averaging of multiple symbols.

The output of the correlator 28 is fed to a further FIFO register 36 which includes Ng memory locations. Integrator 38 receives the output of FIFO register 36 and directly the output of correlator 28 . The effect of the integrator is to add each new sample to the current integrator output, and the samples received from the previous Ng samples. Thus, the output represents the sum of the latest Ng samples. The output is shown in Figure 2(c). Note that the output gradually increases towards the end of each symbol and then decreases immediately thereafter. A peak detector (not shown) generates a timing signal every time the output of the integrator reaches a peak value (eg, shown at timing t in FIG. 2 ). It is used as a synchronization pulse for FFT 14, and it is suitable to appear at the end of each symbol, that is, the latest Nu samples accepted by FIFO register 36 are used by FFT 14, and it should be noted that it appears in this case.

The FFT operates on Nu samples of the significant part of the signal. From Fig. 2, the synchronization signal t can be supplied at any time within the last Ng samples of the symbol (that is, for the case where the waveform of Fig. 2(b) is at a high level), and, since the guard interval G It means that the previous Nu samples are samples from the same signal, and ISI judges avoidance.

Figure 4 shows one possibility of the effect of multipath interference. Figures 4(a) and 4(b) show the same signal received at different times, with Figure 4(a) showing a weaker signal than the originally received signal in this case.

Figure 4(c) shows the output supplied by the correlator of Figure 3 in the absence of the signal of Figure 4(b), and Figure 4(d) shows the output supplied by the correlator of Figure 3 in the absence of the signal of Figure 4(a). When both signals are present, the correlator produces the output shown in Figure 4(e). (At this time, FIG. 4(c) to FIG. 4(e) show correlator outputs that are summed and averaged over a plurality of symbols).

The waveform of Fig. 4(e) has three parts 101, 102 and 103. These sections collectively represent intervals in which the output of the correlator is high by an intentional correlation between values separated by samples of Nu within one or both signals. Note that section 102 takes the last Ng samples from the symbols of signal 4(a) as the only part of the waveform that occurs concurrently with the last Ng samples of the symbols of signal 4(b). Thus, subinterval 102 is the only interval in which ISI is avoided while a timing signal can be supplied.

However, when the output of the correlator is integrated by the conventional circuit of FIG. 3, the output becomes the output as shown in FIG. 4(f). The peak of this output occurs at the end of section 103, meaning at this position at the latest. In particular, it means that FFT 14 only processes the sample of symbol i from the signal of FIG. 4( b ), and additionally processes the sample of symbol i+1 from the signal of FIG. 4( a ).

Referring to FIG. 5 , the synchronous circuit 20 of the embodiment of the present invention is equipped with a correlator 28 formed by a shift register 30 , a complex conjugate circuit 32 and a multiplier 34 , as in the conventional device of FIG. 3 . The output of correlator 28 is fed to averager 46, which may be the same FIFO register as in the existing arrangement of Fig. 3, but in this case the number of parts is equal to L1 which is much less than Ng. The output of FIFO 46 is sent to a symbol averager 48 which sums the samples corresponding to Ns symbols preceding each sample from FIFO 46 . Therefore, the output of the symbol summing averager is equal to the output of the correlator which takes L1 samples and Ns samples as the summing average. In the case of multi-path interference shown in Figure 4, the output is slightly smoothed due to the averaging of L1, but it is similar to the waveform shown in Figure 4(e).

The output is then sent to a timing recovery circuit 50 . This supplies the synchronization signal to FFT 14 .

The functions performed by the timing recovery circuit 50 are briefly represented by the block diagram of FIG. 6 . The samples output from the signal averager 48 are supplied to an absolute value circuit 52 . Absolute value circuit 52 calculates the absolute value of each sample, i.e. (x ² +y ² ) ^1/2 (where x and y are the in-phase and right-angle phase components of the sample) These are checked by peak detector 54 which determines the sample with the largest value. Window generation circuit 56 responds to the samples from absolute value circuit 52 and the peak value detected by circuit 54 to determine the closest samples on both sides of the peak value as less than a threshold equal to 0.75 times the peak value. Thus, the window generation circuit 56 detects the sampling range from n _min to n _max representing the highest degree of correlation of the signal from the correlator. Figure 7 shows a standard waveform representing the samples from the symbol averager 48 during this period. A timing signal generated with a period from n _min to n _max may be suitable for avoiding ISI.

The synchronization signal generator 58 generates a synchronization pulse once per symbol following the initialization operation described later.

The comparator 60 compares the time at which the timing signal is generated with the sample value range from n _min to n _max determined by the window generating circuit 56 . If there is a large difference, the value stored in the counter circuit 62 is changed. When one of the several values stored on the counter circuit 62 reaches a prescribed threshold, a signal for adjusting the timing of the synchronization signal by only an amount dependent on the range n _min to n _max calculated by the window generation circuit 56 is sent to the signal generation Device 58. The means are configured to generate a timing signal approximately midway between samples n _min and n _max , but to adjust only if there is a continuous and/or substantial error from the current timing signal.

In this embodiment, as shown in FIG. 7 , the comparator 60 divides the range from n _min to n _max into four quadrants, namely q1, q2, q3 and q4, in order to increase the number of samples. The initial "early" register in counter 62 is only incremented by one when comparator 60 determines that the current timing of the sync pulse is at q1. The second "late" register is only incremented by 1 when the timing signal is determined to be at q4. When it is judged that the timing is in q2 or q3, both registers are only decremented by 1, but not lower than 0. At any time when the arbitrary counter reaches 4, the counter circuit 62 shifts the timing pulse generated by the signal generator 58 by only an amount corresponding to the timing for the next symbol (or more time for further processing as described below, For example, (n _max −n _min )/4 is smoothly added to the 4 closest samples of a specified later symbol such as the second or third subsequent signal). The timing pulse moves in the forward or reverse direction in response to whether the early or late counter reaches a value of 4.

The other registers of counter circuit 62 are incremented or decremented corresponding to whether the timing is outside the range n _min to n _max . When this happens for four consecutive cycles, the counter circuit 62 performs an initialization operation.

As a result of retuning to a new station time or following power on, the signal generator 58 is set to generate a timing signal midway between n _min and n _max . The initialization operation may also include changing filter characteristics, as will be described later.

Each time signal generator 58 shifts the timing of the sync pulse, the output of FFT 14 produces a different phase rotation of the carrier. To facilitate this process, the counter circuit 62 of the timing recovery circuit 50 outputs a signal, which is received by the phase rotator 15 , indicative of the amount of change applied on the sync pulse. Phase rotator 15 includes a look-up table that stores pre-calculated phase rotations corresponding to possible values represented by the signal from timing recovery circuit 50 . Therefore, when the signal is received, the appropriate value is derived from the lookup table, and each complex sample within the FFT output is adjusted by the corresponding amount only. As an alternative thereto, the phase rotator 15 may include means for computing a phase rotation in response to a signal from the timing recovery circuit 50 . The reasons for the promotion of phasing are:

(a) The timing recovery circuit 50 generates a signal representing the adjustment amount of the synchronization pulse,

(b) As previously stated, the timing recovery circuit is configured to adjust less frequently,

(c) smoothing the adjustments of the additional sync pulses, thereby reducing the number of different adjustments that may be applied to the timing of the sync pulses,

(d) The timing recovery circuit may pre-specify the initial symbol,

(e) Since the timing adjustment is only performed after the timing recovery circuit 50 detects continuous timing errors of the same nature, it is desirable to perform appropriate phase rotation determination in advance, for example, when it is judged that only 1 or 2 symbols have a timing error. have more time for action,

(f) The applied changes are calculated in advance and stored in a lookup table.

Referring again to FIG. 5, the output of FIFO register 46 is also sent to other FIFO register 64 which functions as a sliding window averager for summing groups of L2 consecutive samples. Divide the sample rate by F1 and add the L2 most recent samples. Preferably L1XL2 is actually equal to Ng. The combination of the two averagers 46 and 64 is functionally equivalent to the existing averager 36 of the prior art circuit of FIG. 3 . The output of the averager 64 is sent to a peak search circuit 66 which detects the sample of the largest size and derives the angle of that sample which yields an estimate of the slight frequency offset. Next, a signal indicating the frequency offset is sent to a frequency correction circuit 12 that corrects the frequency by phase rotation of the received samples.

In this embodiment, the L2 averager 66 adds and averages the continuous values in the symbols, or (the delay is correct to obtain the small frequency estimation value) performs the addition and average on the corresponding values in the consecutive symbols.

When a receiver is powered up or tuned to a new station, it is desirable to lock on to the new signal as quickly as possible. The process preferably begins with the initially received symbol. At this time, the value of Ns, i.e. the number of symbols considered by the symbol summing averager 48, starts at 1, so that it is not permissible to increase said value to a relatively small number ( For example, above 8), it will be incremented for each newly received symbol.

Since Ns starts from a very small value and then increases, it is desirable to vary the values of L1 and L2 in the initial stage. It is better to start with a relatively large value for L1 (of course, it is better to be much smaller than Ng). The reason is that if the value of Ns is small, the output of the L1 averager tends to become excessively irregular. For example, when L1 is set to 64 and Ns is set to 1, a good first estimated timing value is obtained for the synchronization signal of the first symbol. When L1 is initially set to a relatively large value, it is better to set L2 to a small value to compensate.

The table of Fig. 8 shows an example of how these values vary.

The value of the 9th and subsequent symbols retains the value of the 8th symbol.

It can be understood that the present invention is effective not only in the simple multi-path interference described with reference to FIG. 4, but also in other states where signals are received via two or more paths. Under such circumstances, the waveform of Figure 4(e) becomes a more complex trapezoidal waveform. However, in the case where the delay is spread out such that there are guard interval cycles repeated, the technique of the present invention can be used to determine the window corresponding to which the synchronization signal should be generated.

Referring to the period at the beginning of the symbol where there is a repetition of the guard interval, as in the case of the above embodiment, it should be noted that this is not necessarily an appropriate time to generate a synchronization signal. In the above-mentioned embodiments, the overlapping degree of repeated data corresponds to the period of the signal to be sent out. The choice of an appropriate interval is influenced by factors such as whether the guard interval should be considered as the beginning or end of the signal, and (as in the above embodiments) whether a timing signal is used to specify the end of the symbol period rather than the beginning. It should be noted that the above description does not take into account, for example, the delay caused by the FIFO averager 46 . In the above-mentioned embodiment, actually, since this delay is taken into consideration, it is preferable to perform correction corresponding to -(L1)/2 sampling.

In the above-described embodiment, samples at intervals separated by only Nu's sampling period are correlated by multiplying 1 sample by the complex conjugate of other samples. Other structures are also possible. For example, as described in the opposing party's British Patent No. BPA9920446.3 (UK Attorney Docket No. J00041703GB), a correlator operates by taking the difference of the absolute values of samples separated by Nu sampling periods.

The invention has so far been described in the context of an OFDM receiver architecture using synchronization pulses in order to define the window of samples for the fast Fourier transform. However, the present invention is useful in other environments where it is necessary to have sync pulses representing symbol boundaries. For example, such pulses are available in transponders that do not perform full FFT demodulation.

The functional components described here can be implemented with either dedicated hardware or software.

As described above, since the synchronization pulse generating method related to the present invention is a method of generating a synchronization pulse representing a symbol boundary of an OFDM signal containing effective symbol periods separated by guard intervals, it includes the steps of: Corresponding to a portion of the data within each active period, a signal representing the degree of correlation between samples of the received signal separated by a period corresponding to the active portion of the symbol is provided, whereby the signal provides an interval in each symbol representing a correlation identified as intended output, but also includes the step of determining the sub-interval in which the maximum degree of correlation occurs to supply a synchronization pulse within the sub-interval, so that while avoiding inter-symbol interference (ISI), reduction of processing time including collection time and processing power are realized reduce.

Claims (23)

1. the method for the lock-out pulse of a character boundary that produces the ofdm signal that expression comprises the significant character cycle of being separated at interval by protection comprises step:

A part of corresponding the data in the data in each protection at interval and each effective period; the signal of expression by the degree of correlation between the sampling that separates received signal corresponding to the cycle of symbol live part is provided; thus; signal provides the output of representing to confirm as the relevant interval of having a mind in each symbol, but also comprises that judgement the son of maximal correlation degree wherein occurs at interval to supply with the step of lock-out pulse in this son interval.

2. according to the method for claim 1, it is characterized in that being suitable for threshold values by the signal to the expression degree of correlation judges described son at interval.

3. according to the method for claim 2, it is characterized in that described threshold values is variable.

4. according to the method for claim 3, it is characterized in that described threshold values represents to depend on the peaked value of the signal of the expression degree of correlation.

5. according to the method for aforementioned any one claim; it is characterized in that representing signal filtering before utilizing the signal of judging described son interval of the described degree of correlation; this filtering be actually each filtering the filtering of mean value of continuous sampling of output sampled representation defined amount, the number of samples in the described defined amount actual specific protection at interval is few.

6. according to the method for claim 5, it is characterized in that described filtering output be illustrated in the average value of addition on a plurality of symbols.

7. according to the method for claim 6, it is characterized in that described filtering the output number that is added average symbol in collection phase, increase in, adjust filtering so that reduce number of samples continuous in the described collection phase, by filtering the mean value of the described continuous number of samples of each output sampled representation.

8. according to claim 5,6 or 7 method, before the output that it is characterized in that handling described filtering, carry out filtering again, so that the signal of the accurate frequency shift (FS) of expression to be provided.

9. according to the method for aforementioned any one claim, it is characterized in that comprising the step of only when the error overshoot threshold values that current timing is calculated, adjusting the timing of lock-out pulse.

10. according to the method for aforementioned any one claim, it is characterized in that comprising and only on the symbol period of defined amount, judge the step of adjusting the timing of lock-out pulse when having that the symbol period of defined amount is more than 1 because of the error due to the current timing.

11., it is characterized in that the ormal weight of the timing adjustment of lock-out pulse corresponding to a plurality of sampling periods according to the method for aforementioned any one claim.

12. a character boundary that produces the ofdm signal that expression comprises the significant character cycle that is spaced apart by protection the method for lock-out pulse, comprise step:

A part of corresponding the data in the data in each protection at interval and each effective period when the error overshoot threshold values of current timing calculating, adjusted the timing of lock-out pulse in response to this.

13. a character boundary that produces the ofdm signal that expression comprises the significant character cycle that is spaced apart by protection the method for lock-out pulse; comprise step: a part of corresponding the data in the data in each protection at interval and each effective period; on the symbol period of defined amount, judge when having because of the error due to the current timing; adjust the timing of lock-out pulse in response to this; wherein, the symbol period of described defined amount is more than 1.

14. according to the method for claim 13, it is characterized in that when having the error of overshoot threshold values on the symbol period of existing timing at described defined amount, adjust the timing of lock-out pulse in response to this.

15. a character boundary that produces the ofdm signal that expression comprises the significant character cycle that is spaced apart by protection the method for lock-out pulse; comprise step: a part of corresponding the data in the data in each protection at interval and each effective period, only the ormal weight of the timing adjustment of lock-out pulse corresponding to a plurality of sampling periods.

16. according to the method for one of claim 12 to 14, the timing that it is characterized in that lock-out pulse only is adjusted the ormal weight corresponding to a plurality of sampling periods.

17. an ofdm signal method of reseptance is characterized in that comprising step: be the compound sampling that fast fourier transform is applied to derive, use the method and the lock-out pulse of aforementioned any one claim to produce lock-out pulse from ofdm signal.

18., it is characterized in that also comprising step according to the method for claim 17: when the timing of lock-out pulse changes, provide the signal of expression intensity of variation, to conversion sampling apply the phase place rotation of determining by this signal.

19., it is characterized in that the rotation of described phase place determined by the value of the search list of address appointment according to the signal of the intensity of variation of the timing of expression lock-out pulse according to the method for claim 18.

20. in a kind of method that receives ofdm signal, comprise producing pulse and using the step of lock-out pulse from the compound sampling that ofdm signal is derived, but also comprise the step that the intact sampling of conversion is applied the definite phase place rotation of this signal for fast fourier transform is applied to.

21., it is characterized in that phase place rotation determined by the value of the search list of address appointment according to the signal of the intensity of variation of the timing of expression lock-out pulse according to the method for claim 20.

22. a lock-out pulse generating means is characterized in that operating according to the method for one of claim 1 to 16.

23. an OFDM receiver is characterized in that operating according to the method for one of claim 17 to 21.

Images