JP3006382B2 - Frequency offset correction device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital modulation / demodulation used in digital communication, and more particularly to a frequency for estimating and correcting a frequency offset when a modulated wave subjected to digital angle modulation is received and demodulated. The present invention relates to an offset correction device.

[0002]

2. Description of the Related Art In recent years, in the field of digital mobile communication, research for practical use has been actively conducted. In particular, in terms of modulation and demodulation, the use of equalizers is being considered to overcome the environment of mobile communication.However, when using a relatively high modulation rate corresponding to the demand for a high data rate, the moving speed is not very high. For applications that are not high, such as cordless telephones, it is considered that the frequency fluctuation due to fading is not so large. Further, a system is conceivable in which the fixed frequency offset between transmission and reception is not so large (for example, about several percent) with respect to the symbol rate in that the modulation rate is relatively high. When transmission and reception are performed in such a situation, it is considered that the deterioration due to fading is not so large even if synchronous detection is used. However, considering the configuration of the receiving system, a delay is required to realize a simple configuration. Generally, a detection system is used. The delay detection system naturally does not require the synchronization of the reception frequency due to its configuration, but has the disadvantage that the characteristics are rapidly deteriorated with respect to a frequency shift exceeding a certain value. Therefore, practically, frequency offset correction is required.

[0003] A conventional frequency offset correction method will be described below. FIG. 4 shows a configuration of a conventional frequency offset correction method. FIG. 4 shows a configuration of a baseband differential detection method in which a delay is performed in a baseband, and a so-called baseband phase differential detection in which a delay operation (differential operation) is performed based on phase information is adopted. In FIG. 4, reference numeral 10 denotes an asynchronous quadrature demodulator 11 for performing an operation of converting a received IF input signal 100 into a complex baseband signal by a local source asynchronous with a transmission carrier, and angle data on a phase plane from the complex baseband signal. A demodulation unit 20 comprising the complex / angle converter 12 for converting the angle data into a delay signal 21 for receiving 10 angle data and delaying it by one symbol, and a subtractor for calculating the difference between the delayed angle data and the current angle data 22 and a differential operation unit comprising
Is an N-multiplier 31 that performs an N-multiplication operation on the phase data in accordance with the arrangement of the modulation signal, an MOD2π converter 32 that degenerates the N-multiplied data into a range of 0 to 2π,
1 / N arithmetic unit 33 for performing an operation of 1 / N of the multiplied data, and a modulation scheme for forcibly inserting a frequency offset between symbols (for example, π / 4 shift QPSK)
In the case of the above, a phase plane comprising a fixed offset value generator 34 for forcibly offsetting the degenerated phase point so that the angle becomes 0, and a difference circuit 35 for subtracting the fixed offset value from the output 133 of the 1 / N arithmetic unit The degenerating section 70 has an estimating section including a short section averaging section 71 that receives and averages the phase difference data output in symbol units, and an averaging section 72 that performs averaging over a long section thereafter. A phase corrector 41 that corrects with an offset value 170 estimated by receiving the output 120 of the differential operation unit,
The correction / determination unit includes a code determination unit 42 that determines a code from the correction phase and outputs the code as received data 200.

[0004] The operation of the frequency offset correction method configured as described above will be described below. First, the received modulated wave is converted into an IF band signal by appropriate means (100) and input to the demodulation unit 10. In the demodulation unit 10, the signal is decomposed into in-phase and quadrature components and then converted into angle data to become current phase angle data 110. The differential operation unit 30 takes the phase difference from one symbol before in order to perform differential detection on the phase plane and outputs it to the correction determination unit as phase difference data. Is output. In preparation for estimating an offset value on the phase difference plane, the phase degenerating unit 30 performs an N-multiplication operation according to a modulation method to remove a modulation component, or when there is a forced phase offset between one symbol. , By removing the fixed offset,
Adjustment is performed so that the degeneration result when there is no frequency offset represents “0” on the phase difference plane. The degenerate data is averaged by the estimating unit so as to suppress the characteristic deterioration in a noise environment. In the conventional example, considering the actual communication data configuration, it is generally considered that the intermittent operation is performed in terms of power consumption and the like. Therefore, for example, as shown in FIG. In this configuration, the average of one intermittent operation section is calculated, and an operation of taking an average over the intermittent operation section by the long section averager 72 is performed to remove the influence of the fluctuation of the communication channel environment.

FIG. 5 is a block diagram showing another conventional example. FIG. 5 shows a configuration different from FIG. 4 in which the N multiplying operation is performed using a complex baseband signal. The method and the principle of operation are the same, although the method is different. 5, reference numeral 100 denotes a received modulated signal (carrier frequency band or IF frequency band); 81, a local oscillator; 82, a frequency converter; 83, a quadrature demodulator; 84, an N-multiplier; , 86 are low-pass filters, and 200 is judgment data.

The operation of the frequency offset correcting method configured as described above will be described below. This configuration basically relates to a method of correcting a frequency of a demodulator for differential detection, and is not a completely synchronous system with respect to a carrier frequency. The received signal 100 is a local oscillator 81
, And is subjected to quadrature demodulation 83. The resulting complex baseband signal is directly converted to N
The signal is multiplied by 84, becomes an error signal, passes through an LPF 86, is fed back to a local oscillator, and is configured as a PLL.

The points different from the first conventional example are that the N multiplying operation is performed by a complex signal and that the phase error signal obtained after the multiplication is used as the control signal of the local oscillator.

[0008]

However, in the above-mentioned conventional configuration, in the method using the phase error for feedforward correction, when the noise component of the received input modulated wave is small, the estimated value of the frequency offset is constant. Values are distributed near the value, and satisfactory correction operation is possible.
When the noise was large, the estimated value deviated considerably from the original offset value. It is clear that, as shown in FIG. 6, when the noise is small, the phase difference has a distribution (a) having a very sharp peak, and when the average is taken, the phase difference is close to the peak value. However, when the noise is large, the distribution of the phase difference tends to have a tail as shown in (b), and when the offset value is large, the phenomenon occurs that the tail goes to the opposite side (c). This is because when performing asynchronous demodulation, the reference source on the receiving side causes rotation on the phase plane due to an absolute frequency error with the transmitting side, and removes the modulation component at the symbol decision point. This is because, in the case where the N multiplication is performed, data in which a determination error occurs in an adjacent quadrant of the N quadrants is regarded as estimated data having the opposite sign. As a result, there is a problem that a certain degree of deterioration occurs essentially in a place where the frequency offset is large and the noise is large. For this reason, even when performing asynchronous quadrature demodulation, there is a method in which the center frequency of the modulated signal is made variable in the frequency of the local oscillation source and the frequency offset component is removed via the control signal as in synchronous detection. In the second conventional example to which this method is applied, phase rotation information is calculated using a complex baseband signal, and a phase error is fed back to a local oscillator via an LPF. Complex arithmetic is indispensable. In addition, there is a problem that it is difficult to use together with the delay detection, which is generally characterized by a simple configuration, such that oscillation is generated or stable control is required for stable operation depending on the control method.

SUMMARY OF THE INVENTION The present invention solves the above-mentioned problems of the prior art, and provides a frequency offset correction device having a simple configuration that operates stably without being affected by noise even when the frequency offset between transmission and reception is relatively large. The purpose is to:

[0010]

In order to achieve this object, the present invention provides, as a first means, when a feedforward type configuration is used, a difference is calculated for each symbol and degenerated on a phase difference plane. The result is the same as that of the first conventional example up to the point of averaging a predetermined period (for example, one time slot period during intermittent operation), but temporarily stores individual difference data. In addition, the average value is corrected by subtracting the calculated average value from the stored difference data, and the average is calculated again for the corrected data, thereby obtaining a more accurate average value. ing. The calculation of this average value can be performed as many times as the configuration allows. Further, as a second means, even when a feedback type configuration is used, a method of returning a correction value to phase data after asynchronous detection, instead of returning an error signal to the frequency conversion means, is adopted. Specifically, in the first means, the recalculation of the average value is performed on the difference data of the same time slot, and the result is applied to the next time slot. That is, the average value once obtained is stored, the average value is applied to the next time slot, the difference is calculated, and the average value is recalculated for the data.

[0011]

According to the present invention, when the phase rotation due to the frequency offset is estimated under the noise environment, first, for example, when the average of the phase rotation amount is calculated in units of symbols over one time slot, the present invention can be applied simply. By subtracting the averaged calculation value from the original data, it is possible to correct for some degree of distribution bias, and even if the frequency offset is more than a certain value and the phase distribution wraps around due to the influence of noise etc. Deterioration of characteristics can be suppressed. Secondly, by applying and subtracting the average value in the current time slot to the next time slot, it is possible to calculate the average for the result of performing the correction in advance, and update the average value sequentially. By doing so, it becomes possible to make the corrected estimated value closer to the true value.

[0012]

【Example】

Embodiment 1 Hereinafter, a first embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of a frequency offset correcting apparatus according to a first embodiment of the present invention.

In FIG. 1, reference numeral 10 denotes an asynchronous quadrature demodulator 11 for performing an operation of converting a received IF input signal 100 into a complex baseband signal by a local source asynchronous with the transmission carrier, and a phase plane from the complex baseband signal. A demodulation unit 20 comprising a complex / angle converter 12 for converting the angle data into the following angle data: a delay unit 21 for receiving the angle data from the demodulation unit 10 and delaying it by one symbol, and a difference between the delayed angle data and the current angle data. And a subtraction unit 22 for calculating the angle data. An N-multiplier 31 performs an N-multiplication operation on the phase plane in accordance with the arrangement of the modulation signal. MOD2π converter 32 for degenerating to the range of ２2π, the N-multiplied data is 1 / N
In the case of a 1 / N unit 33 that performs an operation of performing the operation and a modulation method (for example, π / 4 shift QPSK) that forcibly inserts a frequency offset between symbols, the degenerated phase point is forcibly set to an angle of 0. A phase plane degenerating unit 40 comprising a fixed offset value generator 34 for offsetting the fixed offset value and a difference circuit 35 for subtracting the fixed offset value from the output 133 of the N-th unit, is estimated by receiving the output 120 of the differential operation unit. A correction / judgment unit including a phase corrector 41 for performing correction based on the offset value 150 and a code determiner 42 for determining the received data 200 from the corrected phase. The above-described configuration is the same as that of the conventional example. is there. Also 50
The average value is obtained from a storage device 51 that receives the phase difference data output in symbol units and retains the data for a certain period, an averaging circuit 52 that averages the same data as the stored data, and a stored data 151. Difference circuit 5 for subtracting 152
3. An estimating unit comprising a reaveraging circuit 54 for averaging the difference data 153 again in the same section, and a long section averaging circuit 55 for holding the reaveraging output 154 in a fixed section and averaging it over a predetermined section. .

The operation of the frequency offset correction device configured as described above will be described with reference to the phase distribution diagram shown in FIG. First, (a) is a reproduction of (c) in FIG. 6. In the case where there is a phase rotation due to a frequency offset, a case where noise wraps around a symbol distribution point (shaded area) occurs. Is shown. The error of the estimated average calculation result from the true average value in this case is represented as ΔE . This error is naturally due to a large phase error due to a misjudgment in the hatched portion, and a fairly large error is calculated as it is. Therefore, the original data is re-averaged at the new coordinates (b) using the estimated average value as the origin (± 0). For simplicity, the average number of sections of the long section averaging circuit 55 is one (no average), but the same applies to the case where the average number of sections is two or more. As a result, the phase wraparound becomes smaller than at the time of the initial averaging, and as a result, the estimated average error becomes ΔE ′ (< ΔE ). This is shown in FIG.
The horizontal axis shows the phase rotation amount (fixed value) due to the frequency offset, and the vertical axis shows the phase error estimated by the averaging operation (simple average: ΔE , re-average: ΔE ′ ). This FIG. 2 (c)
As is clear from the above, according to the frequency offset correction method according to the present embodiment, an excellent effect is obtained in estimating the amount of phase rotation in a noise environment.

As described above, according to the present embodiment, by providing a re-averaging circuit in the average portion of the degenerated phase data on the phase difference plane, the average of the phase data wraparound caused by a symbol decision error due to noise is reduced. It is possible to avoid the disadvantage that the value estimation is largely deviated, and to reduce the phase estimation error.

(Embodiment 2) Hereinafter, a second embodiment of the present invention will be described with reference to the drawings.

FIG. 3 shows a block connection of a frequency offset correction device according to a second embodiment of the present invention. In FIG. 3, a demodulation unit 10, a differential operation unit 20, a phase degeneration unit 30,
The four blocks of the correction determination unit 40 are the same as those in the configuration of FIG. 1 is different from the configuration in FIG. 1 in that the estimating unit 60 calculates an average of the average values 164 of the past one average unit, a difference circuit 61 that subtracts the average value 163 from the phase degenerate data 130, and a difference data 161. And an adding circuit 64 for adding a past average to the averaged data 162.

The operation of the frequency offset correction device configured as described above will be described below.

First, as in the first embodiment, the degenerate phase data on the phase difference plane is input to the estimator. A short section averaging operation is performed over a certain period of time by subtracting the estimated value of the average value from the calculation results so far from this data. Thereafter, the averaged data is added to the original averaged data to sequentially correct the deviation of the averaged value. This situation is shown in FIGS. 2 (d), (e) and (f).
FIG. 3D is a diagram showing FIG. 3C again. Note that, as in the first embodiment, the case where the average section number of the long section averaging circuit 63 is 1 will be described below for simplicity. (E)
Corresponds to FIG. 2 (b) in the first embodiment, but differs in that the distribution to be re-averaged is in a new averaging section (time slot) different from (d). It is. The operation according to this produces a slight difference from the first embodiment, but if the average section is more than a certain level, the same distribution can be assumed and the same discussion can be developed. That is, as the estimation error ΔE of the average value 1, the average value 2
Of the estimated error △ E 'is △ E' can be a <△ E. Estimation error of average value 3 in the next average section (time slot) △
E ″ should also be ΔE ″ < ΔE ′ . In this example, as shown in (f), the estimation error is drawn so that the estimation error can be almost zero in the third average. However, in actuality, the estimation error is estimated around a certain center value due to the dispersion of the distribution of the individual time slots. The operation is such that the values are distributed.

As described above, the estimating section 60 for averaging the degenerated phase data on the phase difference plane is offset by the estimated average value in the previous average section (time slot) while the average section (time slot) is offset. By repeating the reaveraging operation in (slot), the phase rotation amount actually caused by the frequency offset can be sequentially estimated. According to this configuration, if the signal-to-noise ratio is equal to or more than a certain level, an almost true average value can be obtained as the estimated value. In addition, there is an advantage that a storage unit for recalculating the same data within the same average section (time slot) as in the first embodiment is not required.

In the first embodiment, 51, 52,
Although the re-averaging circuit composed of 53 and 54 has only one stage, a configuration in which this portion is cascaded to form multiple stages is also possible. In addition, it goes without saying that the short section averaging circuit 62 in the second embodiment may be replaced by the reaveraging circuit composed of 51, 52, 53 and 54 of the first embodiment, and may be realized by a combination configuration.

[0023]

As described above, according to the present invention, first, when a digital angular modulation wave is received and demodulated, when a frequency offset between transmission and reception is estimated on a phase difference plane, a modulation component is detected at a discrimination timing. In order to remove the phase difference data, the phase difference data degenerated by the N-multiplication operation is stored for a certain average interval (time slot), and the result of the averaging operation is subtracted from the stored data, and re-averaging is performed in the same average interval. An averaging operation unit is provided. Secondly, when averaging the degenerated phase difference data, the average value of a predetermined past average section is subtracted from the input phase difference data, and the average value is re-averaged in the current average section (time slot). Is provided. By using such a configuration, the amount of fixed phase rotation due to the frequency offset is large, and the distribution of the phase at the symbol identification point over a wide range due to disturbance such as noise is suppressed. Deterioration of characteristics can be reduced. Also in terms of the configuration, it is possible to realize an excellent frequency offset correction device that can be realized only by an operation related to phase information without using a frequency conversion unit and an analog element such as a loop filter that needs adjustment.

[Brief description of the drawings]

FIG. 1 is a block diagram of a frequency offset correction device according to a first embodiment of the present invention.

FIG. 2 is a characteristic diagram of the frequency offset correction device according to the first and second embodiments.

FIG. 3 is a block diagram of a frequency offset correction device according to a second embodiment of the present invention.

FIG. 4 is a block diagram of a conventional frequency offset correction device.

FIG. 5 is a block diagram of a conventional frequency offset correction device.

FIG. 6 is a characteristic diagram of a conventional frequency offset correction device.

[Explanation of symbols]

 Reference Signs List 10 demodulation unit 11 asynchronous quadrature demodulator 12 complex baseband signal / phase angle converter 20 differential operation unit 21 1 symbol delay unit 22 difference circuit 30 phase plane degenerating unit 31 N multiplier 32 MOD2π operation unit 33 1 / N circuit 34 Fixed offset value generator 35 Difference circuit 40 Correction / determination unit 41 Phase correction circuit 42 Sign determination unit 50 Estimation unit 1 51 Storage circuit 52 Short section averaging circuit 53 Difference circuit 54 Short section re-averaging circuit 55 Long section averaging circuit Reference Signs List 60 estimation section 2 61 difference circuit 62 short section averaging circuit 63 long section averaging circuit 64 addition circuit 70 estimating section 0 71 short section averaging circuit 72 long section averaging circuit 81 local oscillator 82 frequency converter 83 quadrature demodulator 84 complex baseband N Multiplier 85 Sign decision unit 86 Low-pass filter

────────────────────────────────────────────────── ─── Continuation of front page (56) References JP-A-6-205062 (JP, A) JP-A-4-172040 (JP, A) JP-A-5-48665 (JP, A) JP-A-5-48665 347645 (JP, A) (58) Field surveyed (Int. Cl. ⁷ , DB name) H04L 27/00-27/38

Claims (2)

(57) [Claims] 1. A phase data conversion means for receiving a digital angle modulation signal and converting it into phase data on a phase plane for each symbol discrimination point, and receiving a phase data for each symbol to obtain a phase difference between one symbol. Detecting a modulation component from an output of the phase difference detecting means.
Phase elimination means to be removed, first averaging means for performing an averaging operation over a certain section in response to output data of the phase degeneration means, and first storage means for temporarily storing phase difference data in the section; From the stored phase difference data, the first
Difference calculating means for subtracting the average value obtained by the averaging operation, and re-averaging the difference calculation output over the section
A second averaging means, a third averaging means for averaging the output data of the second averaging means over an arbitrary averaging section, and a third averaging means for receiving the phase difference output from the phase difference detecting means, A frequency offset correction device comprising: a difference calculating means for performing a difference calculation with respect to an output of an averaging means; and a sign determining unit which receives the difference calculation output and determines a sign. 2. A phase data conversion means for receiving a digital angle modulation signal and converting it into phase data on a phase plane for each symbol discrimination point, and receiving a phase data for each symbol to obtain a phase difference between one symbol. Detecting a modulation component from an output of the phase difference detecting means.
Phase degenerating means to be removed, first difference calculating means for receiving the output data of the phase degenerating means and calculating a difference from a second average output, and a first difference calculating means for averaging the result of the difference calculation over a certain section. Averaging means, addition means for performing addition between the average data and a second average output described later,
A second averaging means for averaging the output data of the arithmetic operator over an arbitrary averaging section and sending out a second average output; and a second averaging means for receiving the phase difference output from the phase difference detecting means. A frequency offset correction apparatus comprising: a difference calculating means for performing a difference calculation between the output and a sign calculating unit which receives the difference calculation output and determines a sign.