DE102013224664B4 - Compensation device and method - Google Patents

Abstract

An equalizer (100) comprising: a first Fourier transform unit (110) for transforming a receive signal into a frequency domain; a channel characteristic calculation unit (120) for calculating a first channel characteristic concerning the pilot carrier, the first channel characteristic being expressible in relation to a frequency deviation axis and a delay time axis; a channel characteristic partitioning unit (130) for partitioning the first channel characteristic into components arranged in a plurality of regions corresponding to a plurality of first intervals on the frequency deviation axis and a plurality of second intervals on the delay time axis; a desired component extraction unit (140, 150) for extracting desired components from the components detected by the channel characteristic partitioning unit (130) by identifying the desired components from distribution symmetry of the components defined by the channel characteristic partitioning unit (130 ) with respect to an axis of symmetry extending parallel to the delay time axis at a predetermined frequency deviation, or by identifying unwanted components from the distribution symmetry of the components passing through the channel characteristic partitioning unit (130) with respect to the axis of symmetry extending parallel to the delay time axis at a predetermined frequency deviation, and suppressing the unwanted components; a channel characteristic combining unit (160) for generating a second channel characteristic by combining the desired components extracted by the desired component extraction unit (140, 150); and a compensation unit (170) for using the second channel characteristic to compensate for channel distortion in the received signal after transformation into the frequency domain by the first Fourier transform unit (110), wherein the plurality of regions are arranged symmetrically about the symmetry axis. ..

Description

BACKGROUND OF THE INVENTION

1. Field of the invention

This invention relates to a balancing device and a balancing method.

2. Description of Related Art

Orthogonal frequency division multiplexing (OFDM), in which a plurality of orthogonal carriers are multiplexed in each symbol period, is used in many wireless communication systems and in terrestrial digital broadcasting. When an OFDM signal is received in a moving object such as a car, the signal may be disturbed by reflection, scattering and diffraction of obstacles such as buildings and by variations in the signal transmission environment with movement of the object.

To compensate for such a disturbance, the receiver estimates the characteristic of the signal transmission environment and multiplies the received signal by the inverse of the estimated characteristic. The estimation of the signal transmission environment is commonly referred to as channel estimation. As an aid to channel estimation, in addition to the transmitted data, the transmitter inserts a known signal into the transmitted signal, and the receiver computes a channel interference of the known signal, and then uses interpolation between carriers and between symbols to estimate the signal transmission environment in the data intervals. Improved channel estimation techniques are critical for accurate reception under adverse channel conditions with pronounced variations.

In the Japanese patent JP 3802031 B2 Abe describes a channel estimation method in which the channel characteristic is calculated from a scattered pilot (SP) as a pilot carrier signal and a two-dimensional discrete Fourier transform is performed to estimate the channel. In this prior art technique, a two-dimensional discrete Fourier transform is first performed with respect to carrier frequency and symbol time to produce a channel characteristic that includes delay time and frequency deviation. 1 schematically illustrates the result of the two-dimensional discrete Fourier transform for a Rice channel including a direct wave with no time deviation and a delayed Rayleigh wave. The desired channel characteristic 700 goes with unwanted repetitions 800 as shown by hatching in 1 displayed. From a predetermined first filter extraction region 900 in the signals transformed by the two-dimensional discrete Fourier transform become second filter extraction regions 901 which higher power components of the first filter extraction region 900 contained, and it will be the unwanted repetitive components 800 suppressed what the desired channel characteristic 700 leaves. A two-dimensional inverse Fourier transform is then performed on the result to estimate the channel characteristic. In this technique, the suppression of repeat components results 800 also for interpolation of the signal characteristic at non-SP data intervals in the frequency and symbol time directions.

By filtering the results of the two-dimensional Fourier transform, the Abe technique is capable of suppressing both unwanted repetition components and Gaussian noise effects, but if there is a large frequency deviation, the unwanted components may invade the predetermined region, in which case an exact channel characteristic is not obtained. An example will be in 2 shown in which due to large time deviation in a two-signal Rice channel an undesirable component 801 in the region 902 extends, from which the desired channel characteristic 701 is extracted, which causes an undesirable region 904 in addition to the desired regions 903 is extracted.

US Pat. No. 8,064,507 B1 describes an OFDM receiver. A second set of channel estimates is generated based on a Wiener filter interpolation of a first set of channel estimates. The coefficients of the Wiener interpolation filter are based on a channel response estimate that includes two or more narrow ranges of non-zero amplitude separated by two or more wide ranges of substantially vanishing amplitude.

Other equalizers with channel estimators performing aliasing suppression are off EP 1 564 946 A2 . US Pat. No. 7,940,848 B2 . US Pat. No. 8,520,783 B2 . US 2011/0206146 A1 . US 2012/0020427 A1 . US 2009/0135923 A1 . US 2011/0158334 A1 . EP 2 387 840 B1 and US 2008/0175140 A1 known.

SUMMARY OF THE INVENTION

An object of the present invention is to perform a channel estimation in a manner that adapts to large frequency deviations in the channel environment.

The invention provides a compensating device with the features according to claim 1 ready.

The dependent claims describe preferred embodiments.

The invention also provides a compensation method having the features according to claim 17.

The above apparatus and method can perform channel estimation in a manner that adapts to large frequency variations in the channel environment.

BRIEF DESCRIPTION OF THE DRAWINGS

In the accompanying drawings:

1 schematically illustrates the result of a two-dimensional Fourier transformation for a Rice channel with two signals and small time deviation;

2 schematically illustrates the result of a two-dimensional Fourier transform for a Rice channel with two signals and a large temporal deviation;

3 Fig. 11 illustrates a block diagram schematically illustrating the configuration of a balancing device according to a first embodiment of the invention;

4 schematically illustrates the structure of an OFDM symbol in the first embodiment;

5 schematically illustrates the placement of a dispersed pilot signal in the first embodiment;

6 schematically illustrates an exemplary subdivision of the channel characteristic in the first embodiment;

7 schematically illustrates exemplary filter passbands in the first embodiment;

8th schematically illustrates a partitioning of the frequency deviation axis in the first embodiment;

9 schematically illustrates the power distribution of the scattered pilot in a two-signal Rice channel in a space having a frequency deviation axis and a delay time axis in the first embodiment;

10 schematically illustrates the power distribution of the scattered pilot in a three-signal Rice channel in a space having a frequency deviation axis and a delay time axis in the first embodiment;

11 schematically illustrates a region for estimating noise power in the first embodiment;

12 schematically illustrates an exemplary comparison of channel characteristic power symmetry and channel characteristic suppression in the first embodiment;

13 schematically illustrates the suppression of repetition components in carriers in which no scattered pilot can be present in the first embodiment;

14 FIG. 10 is a flowchart illustrating the equalizing method used by the equalizer in the first embodiment; FIG.

15 Fig. 10 is a block diagram schematically illustrating the configuration of a balancer according to a second embodiment;

16 Fig. 10 is a block diagram schematically illustrating the configuration of the channel characteristic partitioning unit in the second embodiment;

17 Fig. 10 is a block diagram schematically illustrating the configuration of the channel characteristic combining unit in the second embodiment;

18 Fig. 12 is a block diagram schematically illustrating the configuration of a balancing device according to a third embodiment;

19 schematically illustrates the power distribution of the channel characteristics for frequency deviation and delay time axes in the third embodiment;

20 Fig. 10 is a block diagram schematically illustrating the configuration of the channel characteristic partitioning unit according to a variation of the third embodiment;

21 Fig. 10 is a block diagram schematically illustrating the configuration of the channel characteristic combining unit according to a variation of the third embodiment;

22 Fig. 10 is a block diagram schematically illustrating the configuration of the channel characteristic partitioning unit according to a fourth embodiment;

23 Fig. 10 is a block diagram schematically illustrating the configuration of a channel characteristic partitioning unit according to the fourth embodiment;

24 schematically illustrates shifts in the frequency deviation axis direction and delay time axis direction in the first embodiment;

25 schematically illustrates components of regions calculated from shifts in the frequency deviation axis direction and delay time axis direction in the fourth embodiment;

26 Fig. 12 is a block diagram schematically illustrating the configuration of a carrier inverse Fourier transform unit according to a fifth embodiment;

27 schematically illustrates an exemplary partitioning of the channel characteristic at N-point intervals on the carrier frequency axis in the fifth embodiment;

28 schematically illustrates a scattered pilot channel characteristic calculated without partitioning in the fifth embodiment;

29 schematically illustrates a scattered pilot channel characteristic calculated by partitioning in the fifth embodiment;

30 Fig. 12 is a block diagram schematically illustrating the configuration of the carrier Fourier transform unit according to the fifth embodiment;

31 schematically illustrates a window function in the fifth embodiment; and

32 FIG. 12 is a block diagram schematically illustrating the configuration of a variation of the carrier inverse Fourier transform unit according to the fifth embodiment. FIG.

DETAILED DESCRIPTION OF THE INVENTION

Embodiments of the invention will now be described with reference to the accompanying drawings, in which like elements are indicated by like reference numerals.

First embodiment

Referring to 3 includes the balancing device 100 in the first embodiment, a Fourier transform unit 110 , a scattering pilot (SP) channel characteristic calculation unit 120 , a channel characteristic partitioning unit 130 , a repetition appraisal unit 140 , a repetition suppression unit 150 , a channel characteristic combining unit 160 and a compensation unit 170 , The repetition estimation unit 140 and the repetition suppression unit 150 form a desired component extraction unit.

The channel characteristic partitioning unit 130 includes first to Mth symbol filter units 131-1 to 131-M and first to Mth carrier inverse Fourier transform units 132-1 to 132-M where M is an integer equal to or greater than two. The channel characteristic combiner unit 160 includes an adder unit 161 and a carrier Fourier transform unit 162 ,

The balancing device 100 forms part of a receiver (not shown), the OFDM symbols SBL of the in 4 type, including a guard interval (GI) preceding the valid symbol part. The total symbol duration Ts in 4 consists of the guide interval period Tgi and the valid symbol duration Tu. The OFDM symbol SBL is generated by an inverse discrete Fourier transform with F_DFT points; in the OFDM symbol SBL, the scattered pilots become like in 5 shown set. The scattered pilot SP will be like in 5 shown at intervals of four symbols in a time direction and twelve carriers in a frequency direction, with a shift of three carrier positions per symbol. Accordingly, the number of carriers in which the scattered pilot can be inserted is F_DFT / 3. This signal configuration is also used in the following embodiments.

Again referring to 3 performs the Fourier transform unit 110 a discrete Fourier transform with F_DFT points to determine the individual carrier signals. This transformation converts the received signal into the frequency domain. The Fourier transformation unit 110 is also referred to as the first Fourier transform unit.

The scatter pilot channel characteristic calculation unit 120 calculates a pilot characteristic that touches the pilot carriers contained in the received signal. The scattered pilot channel characteristic calculation unit 120 calculated channel characteristic is also referred to as the first channel characteristic or the scattered pilot channel characteristic. For example, the scattering pilot channel characteristic calculation unit extracts 120 the scattered pilot signal from the through the Fourier transform unit 110 determined individual carrier signals and determines the channel characteristic, which relates to the scattered pilot, by dividing the extracted scattered pilot values by known values. The scatter pilot channel characteristic calculation unit 120 outputs a channel characteristic for every third carrier, that is to say for the carriers into which the scattered pilot can be inserted, and outputs zero values for the carriers into which the scattered pilot can not be inserted. The scatter pilot channel characteristic can be considered as a complex function with carrier frequency and symbol time as variables. An SP channel characteristic with F_DFT / 3 points per symbol is applied to the channel characteristic partitioning unit 130 applied.

The channel characteristic is referred to as a channel characteristic independent of various timeline and frequency axis transformations. The dimensions of the discussed channel estimation are noted as needed. For example, the output from the scatter pilot channel characteristic calculation unit 120 are expressed as a function of carrier frequency and symbol time, but the channel characteristic in 1 and in most of the discussion below, has been expressed on a delay time axis by performing an inverse Fourier transform with respect to the carrier frequency and on a frequency deviation axis by performing a Fourier transform with respect to symbol time.

The channel characteristic partitioning unit 130 divide that by the scatter pilot channel characteristic calculation unit 120 calculated channel characteristic in components occurring in a plurality of partitions corresponding to a plurality of first intervals on the frequency deviation axis with a plurality of second intervals on the delay time axis. The channel characteristic partitioning unit 130 for example, divides the one from the scatter pilot channel characteristic calculation unit 120 determined scattered pilot channel characteristic in the N intervals on the delay time axis and M intervals on the frequency deviation axis, as in 6 shown. The partitions are preferably placed symmetrically with respect to the delay time axis. In the first embodiment, because the scattered pilot channel Characteristic F_DFT / 3-points per symbol, N is equal to F_DFT / 3. reference numeral 702 in 6 denote a desired component and reference numerals 802 denote unwanted components.

Next, a specific method of partitioning the scattered pilot channel characteristic will be described.

Each of the first to Mth filter units 131-1 to 131-M wid is configured as a finite impulse response (FIR) filter or an infinite impulse response (IIR) filter and each has its own filter characteristic. The filters that contain the first to Mth symbol filter units 131-1 to 131-M form, are also referred to as the first filter. The mth symbol filter unit 131-m (1 ≦ m ≦ M) performs a filtering process in the symbol direction by performing on N carriers with an interspersed scattered symbol in the scattered pilot channel characteristic as shown in the following equation (1).

In equation (1), h (•) denotes the scattered pilot channel characteristic, k is an index indicating the carrier frequency, 1 is an index denoting the symbol time, is a_m a filter coefficient for the m-th symbol filter unit 131-m and g_m (•) denotes the filter result.

Exemplary filter passbands are in 7 shown. In 7 the margin between the maximum frequency deviation fd_max and the minimum frequency deviation fd_min is divided into M intervals. After suppression of repetition in a subsequent stage, the individual filter outputs are replaced by those in 3 shown adding unit 161 so that the filter components are preferably set to be the sum of the symbol filter characteristics of the first to Mth symbol filter units 131-1 to 131-M is constant. If such filter characteristics are used, the outputs correspond to the first to Mth symbol filter units 131-1 to 131-M M regions determined by partitioning the plane defined by the delay time and frequency deviation axes in M sections on the frequency deviation axis. In 8th denote reference characteristics 703 a desired component and designate reference characters 803 unwanted components.

Back on 3 Referring to Figure 1, each of the first through Mth carries carrier inverse Fourier transform units 132-1 to 132-M an inverse Fourier transform on the output of the corresponding one of the first through Mth symbol filter units 131-1 to 131-M representing the part of the scattered pilot channel characteristic in an interval of the frequency deviation axis. This transformation is equivalent to partitioning the characteristic on the delay time axis. Each of the first to Mth carrier inverse Fourier transform units 132-1 to 132-M For example, the channel characteristic may be partitioned into N components on the delay time axis by an inverse Fourier transform with N points performed on a filtered, scattered pilot channel characteristic, as in equation (2). In the output from the channel characteristic partitioning unit 130 For example, the scattered pilot channel characteristic is divided into M × N components according to frequency deviation and delay time axes.

In equation (2), f_m (•) denotes the output from the m-th Fourier transform unit 132-m and τ is an index indicating the delay time.

The value of M (number of frequency partitions) is set according to the channel. Increasing the number of partitions allows for finer granular frequency band control, that is, finer granular separation of desired components of the channel characteristic from undesired components such as repetitive components and Gaussian noise and fine granular suppression of the unwanted components. The values of fd_max and fd_min can be set according to the amount of time deviation in the channel. Moreover, although the channel characteristic need not be partitioned into equal intervals, the partitions are preferably symmetrical in positive and negative directions of the frequency deviation axis, in other words symmetric with respect to the delay time axis.

The positions of the first through Mth symbol filter units 131-1 to 131-M and the first to M-th Fourier transform units 132-1 to 132-M can be exchanged. In other words, the scattered pilot channel characteristic may first pass through each of the first to Mth carrier inverse Fourier transform units 132-1 to 132-M and then through each of the first to Mth symbol filter units 131-1 to 131-M ,

The desired component extraction unit containing the repetition estimation unit 140 and the repetition suppression unit 150 includes extracts desired components. For example, the desired component extraction unit may extract desired components from the components defined by the channel characteristic partitioning unit 130 by identifying the desired components from the distribution symmetry of the components passing through the channel characteristic partitioning unit 130 with respect to an axis of symmetry extending parallel to the delay time axis at a predetermined frequency deviation. Alternatively, the desired component extraction unit may desire components by identifying unwanted components from the distribution symmetry of the components passing through the channel characteristic partitioning unit 130 with respect to this symmetry axis, and suppressing the unwanted component, extract.

The repetition estimation unit 140 identifies unwanted components from the symmetry of the components passing through the channel characteristic partitioning unit 130 with respect to an axis of symmetry extending in the delay time direction. For example, the repetition estimation unit identifies 140 unwanted components from a distribution symmetry of power values in the frequency deviation and delay time regions in the M × N partitions of the scattered pilot channel characteristic resulting from the channel characteristic partitioning unit 130 with respect to an axis of symmetry extending parallel to the delay time axis at a predetermined frequency deviation. If the components in two regions have significant asymmetric power values in mutually symmetrical position with respect to the axis of symmetry, the repetition estimation unit identifies 140 include the components in at least one of the two regions, such as the region of higher power, as the undesired component. The repetition estimation unit 140 shares the repetition suppression unit 150 their decision on whether or not there is an unwanted component in each partitioned region (these messages are also referred to as initial decision information).

Alternatively, the repetition estimation unit 140 desired components from the symmetry of the components through the channel characteristic partitioning unit 130 be detected with respect to an axis of symmetry extending in the delay time direction. For example, if the components in two regions have sufficiently symmetrical power values in mutually symmetrical positions with respect to the symmetry axis, the repetition estimation unit identifies 140 containing the two regions as desirable components. The repetition estimation unit 140 shares the repetition suppression unit 150 their decision on whether or not there is a desired component in each partitioned region (these messages are also referred to as second decision information).

The repetition suppression unit 150 suppresses those of the components passing through the channel characteristic partitioning unit 130 which are the repetition estimation unit 140 identified as undesirable. For example, the repetition suppression unit suppresses 150 by the repetition appraisal unit 140 identified repetition components from the partitioned regions of the scattered pilot channel characteristic based on decisions (first decision information) generated by the repetition estimation unit 140 be reported and pick them up. If the repetition estimation unit 140 decides that it is a repetition component in a region multiplies the repetition suppression unit 150 the component of the scattered pilot channel characteristic in this region with zero; if the decision is that there is no repetition component in a region, the repetition suppression unit multiplies 150 the component of the scattered pilot channel characteristic in this region is one.

Alternatively, the repetition suppression unit 150 by the repetition appraisal unit 140 identified desired components from those through the channel characteristic partitioning unit 130 Extract ascertained components. For example, if the decision (second decision information) provided by the repetition estimation unit 140 is reported that there is a desired component in a region multiplies the repetition suppression unit 150 the component of the scattered pilot channel characteristic in the region of unity; if the decision is that there is no desired component in a region, the repetition suppression unit multiplies 150 the component of the scattered pilot channel characteristic in this region is zero.

Assuming that the desired components in the with reference numerals 701 in 2 have the desired components in the part of the channel characteristic with positive frequency deviation and the part of the channel characteristic with negative frequency deviation to each other symmetrical power distributions. The desired components can then be identified from their symmetry with respect to the delay time axis at zero frequency deviation. Because the positions of the unwanted repetition components 801 are characteristically shifted in carrier frequency and symbol timing in the scattered pilot channel characteristic, and repeated with delay time and frequency deviation shifts, lack the power distributions of the undesired repetition components 801 this positive-negative frequency deviation symmetry. A method of extracting desired components in the channel characteristic using this feature will be described below.

First, a two-signal Rice channel with a large time deviation is considered. Referring to 9 For example, a pair of regions labeled a and b and another pair of regions labeled c and d will be compared. The regions a and b are in mutually symmetrical positions and the regions c and d are also in mutually symmetrical positions in the parts of the channel characteristic plane with positive and negative frequency deviation values. The regions a and b contain parts of a desired component 704 , which expresses the channel characteristic of the delayed wave, and therefore have equal power values in the channel characteristic. However, regions c and d have an asymmetric power distribution: only region c contains a repeated component 804 , so that region c has a higher power value.

Referring to 3 calculates the repetition estimation unit using this feature 140 the symmetry of the power distribution from the power in partitions that are at mutually symmetrical positions on the frequency deviation axis, where one partition has positive frequency deviation, and the other has negative frequency deviation. If the power distribution is symmetric, the repetition estimation unit decides 140 in that a desired component is present; if the power distribution is asymmetrical, the repetition estimation unit decides 140 in that there is a repeat component. More specifically, the repetition estimation unit 140 calculate the symmetry of two mutually symmetrical regions by calculating a power difference, a power ratio, or both a power difference and a power ratio, and comparing the calculated value or values to a predetermined threshold or predetermined thresholds. For example, if the absolute value of the power difference is used, if this value is higher than the predetermined threshold, the repetition estimation unit may 140 decide that the current distribution is asymmetrical. The threshold may be determined in advance on the assumption that the power distribution of the desired components may be skewed by an assumed amount of noise and an assumed preselection of the angle of arrival. When the power distribution of the desired components is biased by a known factor, such as antenna directivity, the repetition estimation unit may 140 first correct the current distribution to cancel the preset and then calculate the symmetry.

In calculating the performance, the repetition estimation unit 140 use the following equation (3) to take an average of the scattered pilot channel characteristic over a plurality of symbols. Also, amplitude can be used instead of power.

In Equation (3), Q denotes the number of symbols and q denotes an index indicating a particular symbol.

Using the by the repetition estimation unit 140 The decisions taken are extracted by the repetition suppression unit 150 desired components from the partitioned scattered pilot channel characteristic and suppresses unwanted components.

Another method for identifying and suppressing repeat components by using the power distribution symmetry is described below.

First, the 10 shown a three-signal Rice channel including a direct wave with no time deviation and a delayed Rayleigh wave. In 9 become the desired components 704 from the unwanted, repeated components 804 separated. In 10 however, a desirable component will overlap 704 and an undesirable component 804 each other in region e. If a threshold decision for symmetry is made as described above, because the regions e and f have an asymmetric power distribution, all components in the regions e and f are suppressed. Because the regions e and f are also a desirable component 705 however, if all components in regions e and f are suppressed, channel estimation is degraded.

To avoid this kind of degradation of channel estimation, the repetition estimation unit may 140 identify an inclusion ratio of repeating components from the power in paired regions in mutually symmetrical positions, and may be the repetition rejection unit 150 undesired components by setting suppression ratios for the individual regions according to that by the repetition estimation unit 140 identified repeat component inclusion ratio, and then suppress multiplying the components in each region by the suppression ratio set for the region. For example, the repetition suppression unit 150 Unwanted unwanted components by reducing the suppression ratio with increasing, unwanted component containment ratio.

As motion path, the repetition estimation unit 140 determine an inclusion ratio of repetitive components from the following equations (4) to (8). Pe = Pa + Pb (4) Pf = Pa (5)

In equations (4) and (5), Pe denotes power in the region e, Pf denotes power in the region f, Pa denotes power of a desired component, and Pb denotes power of an unwanted component. The basis for equations (4) and (5) is the assumption that a desired component in the channel characteristic has a symmetric power distribution. Pa therefore occurs equally in regions e and f. Equations (4) and (5) can be rewritten as equations (6) and (7). Pb = Pe - Pa (6) Pa = Pf (7)

The repetition estimation unit 140 can calculate the inclusion ratio Re in the region e as Equation (8). Re = Pb / Pa (8)

In equation (8), the inclusion ratio Re is zero if there is no repeat component. In a region where only one repeat component is present, the inclusion ratio may be given by, for example, a "zero" value.

The repetition estimation unit 140 calculates an inclusion ratio in each region as described above and tells the repetition suppression unit 150 the calculated inclusion ratios with.

The repetition suppression unit 150 can maximum suppression of the unwanted component in region e in 10 without suppressing the desired component in region f by calculating a suppression ratio for region e such that regions e and f have equal channel characteristic performance, and by using this ratio to suppress the undesired component in region e. More specifically, the repetition suppression unit calculates 150 a suppression ratio Rf in each region by the repetition estimation unit 140 reported inclusion ratios using, for example, the following equation (9). In Equation (9), the value of the suppression ratio decreases as the inclusion ratio increases. The Repetitionsunterdrückungseinheit 150 can then suppress the unwanted components by multiplying the component of the channel characteristic in each region by the suppression ratio Rf calculated for the region. Rf = Pa / Pa + Pb (9)

Next, a case in which an undesirable noise component is added to the scattered pilot channel characteristic will be described. The noise component extends over both the frequency deviation and the delay time axes. The repetition suppression unit 150 can suppress the noise component by suppressing components of divisions in which power values do not exceed a predetermined threshold.

Similarly, the repetition estimation unit 140 which works by directly using the power of the channel characteristic power in the above description, the noise power in a region such as region g in 11 , to subtract the noise power from the channel characteristic power values, and then to perform repeat estimation. For example, the repetition estimation unit 140 calculate the noise power in region g and subtract the calculated noise power from the power values in two regions to be compared. As region g, the repetition estimation unit 140 select a subdivision region in which there is neither a desired component nor an undesired component, and the channel characteristic power value is less than a predetermined threshold, or simply select the minimum channel characteristic power value.

The scattered pilot channel characteristic in a Rice channel is in 12 schematically illustrated. Because the main wave has a small frequency deviation, it appears as a valid component only in the subdivisions with a frequency deviation of zero. The desired component of the delayed wave, which has a frequency deviation with positive-negative symmetry, is also valid. Although unwanted repetitions of the delayed wave component penetrate into the region under consideration due to their large temporal variation, these components are suppressed because the frequency deviation of their power distribution lacks positive-negative symmetry. Low power regions are suppressed as noise components. Unwanted components 806 are thereby suppressed in the cross-hatched regions; desired components 706 are extracted. The desired channel characteristic can be obtained by suppressing unwanted components and low power components as described above.

Again referring to 3 calculates the channel characteristic combiner unit 160 the channel characteristic in the carrier frequency domain for use in balancing by combining the desired components extracted by the desired component extraction unit. For example, the channel characteristic combiner unit calculates 160 the channel characteristic in a carrier frequency domain for use in compensation in combining the partitioned scattered pilot channel characteristic in which unwanted repetition components are suppressed. In other words, the channel characteristic combining unit generates 160 a channel characteristic by combining components after undesired repeat components by the repetition suppression unit 150 are suppressed, and by converting the combined components into the frequency domain. The through the channel characteristic combining unit 160 Channel characteristic is also referred to as the second channel characteristic.

The adding unit 161 is configured as an adder which stores the outputs of the repetition suppression unit 150 in M regions for each delay time, thereby uniting the divided sections of the channel characteristic on the frequency deviation axis. The adder unit 161 forming adder is also referred to as the first adder.

Next, the carrier Fourier transform unit adds 162 Zero values into the part of the summed subdivided channel characteristic on the delay time axis with delays outside the interval from 0 to Tu / 3, thereby increasing the number of points from N to F_DFT. The carrier Fourier transform unit 162 converts the channel characteristic from the delay time component to the carrier frequency domain by performing a discrete Fourier transform with F_DFT points. The discrete Fourier transform is equivalent to combining the divided sections of the channel characteristic in the delay time domain. Introducing zero values into the channel characteristic is equivalent to suppressing the repetitive components in the cross-hatched regions, where τ <0 and Tu / 3 <τ in 13 is. The repetition components occur because if all F_DFT points are on the Carrier frequency domain, there are carriers in which no scattered pilot is inserted in any symbol. The repetition suppression unit 150 determines a channel estimate, in which the one by the scatter pilot channel characteristic calculation unit 120 calculated scattered pilot channel characteristic is interpolated into all data intervals, as described above.

Referring again 3 finally uses the equalization unit 170 the channel characteristic to the carrier frequency domain, which by the channel characteristic combining unit 160 is determined to compensate for the channel distortion in the Fourier-transformed received signal.

To put the repeat components in the in 13 to suppress the shown crosshatched regions, the repetition components are first suppressed by an inverse Fourier transform on the carrier frequency axis with F_DFT / 3 points, where F_DFT / 3 is the number of carriers into which the scattered pilot can be inserted, and then zero Values are inserted in the delay time regions, where τ <0 and Tu / 3 <τ to perform a Fourier transform with F_DFT points. Instead of this process, however, the carrier Fourier transform unit 162 first perform a Fourier transform with F_DFT / 3 points and then insert zero values into carriers in which no scattered pilot is present, and a filter unit configured as a carrier direction filter with a passband defined by 0 ≤ τ ≤ Tu / 3 can be the channel Interpolate characteristic. Alternatively, the scatter pilot channel characteristic calculation unit 120 Inserting zero values into all carriers in which there is no scattered pilot can be any of the first to Mth carrier inverse Fourier transform units 132-1 to 132-M perform a Fourier transform with F_DFT points and can use the repetition suppression unit 150 Suppress recurrence components in the regions where τ <0 and Tu / 3 <τ.

In the above discussion, the scattered pilot in F_DFT / 3 of the carrier may be inserted so that the total number of carriers is three times the number of carriers including scattered pilot signals, but this is not a necessary condition; the total number of carriers may be a different multiple of the number of carriers containing scattered pilot signals. Alternatively, each of the first to Mth carrier inverse Fourier transform units 132-1 to 132-M and the carrier Fourier transform unit 162 perform a Fourier transform with a number of points greater than the number of carriers with scattered pilot signals. the signal points for use in the Fourier transform can be generated by extrapolating the scattered point channel estimate in the carrier direction. An exemplary extrapolation method is in Japanese Patent 3802031 disclosed.

The operation of the balancing device 100 in the first embodiment will be described below.

Referring to 14 The compensation method includes a Fourier transform step S10, a scattered pilot channel characteristic calculating step S11, a channel characteristic dividing step S12, a repetition estimating step S15, a repetition suppressing step S16, a channel characteristic combining step S17 and a compensating step S20.

The channel characteristic dividing step S12 includes first through M-th symbol filtering steps S13 and first through M-th carrier inverse Fourier transforming steps S14.

The repetition estimation step S15 and the repetition suppression step S16 constitute a desired component extraction step.

The channel characteristic combining step S17 includes an adding step S18 and a carrier Fourier transforming step S19.

In the Fourier transform step S10, the Fourier transform unit performs 110 a discrete Fourier transform with F_DFT points on the received signal to determine the individual carrier signals.

In the scattered pilot channel characteristic calculation step S11, the scattered pilot channel characteristic calculation unit extracts 120 the scattered pilot signal from the through the Fourier transform unit 110 determined individual carrier signals and determines the channel characteristic affecting the scattered pilot by dividing the extracted scattered pilot by known values. The scatter pilot channel characteristic calculation unit 120 outputs a channel characteristic for every third carrier, that is, the carriers in which the scattered pilot is inserted, and outputs zero values for the carriers into which the scattered pilot is not inserted. The scatter pilot channel characteristic calculation unit 120 provides a scattered pilot channel characteristic with F_DFT / 3 dots per symbol to the channel characteristic partitioning unit 130 , The output from the scattering pilot channel characteristic calculation unit 120 In the scattered pilot channel characteristic calculation step S11, for example, a function of carrier frequency and symbol time is displayed.

In the channel characteristic dividing step S12, the scattered pilot channel characteristic detected in the scattered pilot channel characteristic calculating step S11 is divided into N intervals on the delay time axis and M intervals on the frequency deviation axis, as in FIG 6 through the first through m th symbol filtering steps S13 and first through mth carrier inverse Fourier transform steps S14.

In the first to Mth symbol filtering steps S13, a FIR or IIR filtering process is performed. Each of these steps has a different filter characteristic. The mth symbol filtering step S13-m uses the mth symbol filter unit 131-m to perform a filtering process in the symbol direction, centered on one of the N carriers with scattered pilot signals in the scattered pilot channel characteristic as in equation (1).

In the first to Mth carrier inverse Fourier transform steps S14, the first to the red carrier inverse Fourier transform units partition 132-1 to 132-M the scattered pilot channel characteristic at intervals on the delay time axis by performing inverse Fourier transforms on the outputs of the first through Mth symbol filter units 131-1 to 131-M in the above steps S13, which divides the scattered pilot channel characteristic into intervals on the frequency deviation axis. Each of the first to Mth carrier inverse Fourier transform units 132-1 to 132-M For example, one of these channel characteristic outputs may be partitioned into N components on the delay time axis by performing an N-point inverse Fourier transform on a filtered carrier frequency characteristic in the scattered pilot channel characteristic as in equation (2). In the output of the channel characteristic dividing step S12, the spread pilot channel characteristic is divided into M × N components corresponding to the frequency deviation and delay time axes.

The desired components are extracted in the desired component extraction step consisting of the repetition estimation step S15 and the repetition suppression step S16. The desired components may then be extracted from the components determined in the channel characteristic divider step S12, for example, by identifying the desired components from the symmetry of their distribution with respect to a delay time axis having a predetermined frequency deviation value. Alternatively, the desired components can be extracted by identifying unwanted components from the distribution symmetry of the components detected in the channel characteristic dividing step S12 with respect to this symmetry axis and suppressing the undesired components.

For example, in the repetition estimation step S15, the repetition estimation unit 140 Identify unwanted components from a distribution symmetry of power values in the frequency deviation and delay time regions in the MxN partitions of the scattered pilot channel characteristic output in the channel characteristic dividing step S12 with respect to an axis of symmetry that extends parallel to the delay time axis at a predetermined frequency deviation ,

In the repetition suppression step S16, the repetition suppression unit uses 150 the results of decisions made by the Repetition Assessment Unit 140 in step S15 have been made by the repetition estimation unit 140 in the partitions of the scattered pilot channel characteristic identified repetitive components to suppress and cancel.

In the channel characteristic combining step S17, the channel characteristic combining unit calculates 160 a channel characteristic in the carrier frequency domain for use in compensation by combining the divided scattered pilot channel characteristics in which the unwanted repetition components have been suppressed. The channel characteristic combining step S17 is performed by the adding step S18 and carrier Fourier transforming step S19.

In addition step S18, the adder unit adds 161 the outputs from the repetition suppression step S16 in M regions for each delay time, thereby uniting the divided portions of the channel characteristic on the frequency deviation axis.

Next, in the carrier Fourier transform step S19, the carrier Fourier transform unit adds 162 Zero values in the part of the summed, partitioned channel characteristic on the delay time axis with delays outside the interval from 0 to Tu / 3, thereby increasing the number of points from N to F_DFT, and converts the channel characteristic from the delay time domain into the carrier frequency domain by performing a discrete Fourier transform with F_DFT points.

Finally, in compensation step S20, the compensation unit is used 170 the frequency domain-channel characteristic determined in step S19 for compensating channel distortion in the Fourier-transformed received signal.

As described above, the balancing device 100 in the first embodiment, determine a desired signal by dividing the scattered pilot channel characteristic into the frequency deviation and delay time axes, and using the channel characteristic parts, the power distributions symmetrical to each other with respect to an axis of symmetry parallel to the delay time axis at a predetermined one Frequency deviation extending, for identifying desired components, or for identifying unwanted components.

Second embodiment

The second embodiment uses methods of dividing and combining the channel characteristics different from those performed by the channel characteristic partitioning unit 130 and the channel characteristic combining unit 160 different methods used in the first embodiment. While a filtering process for partitioning the scattered pilot channel characteristic on the frequency deviation axis has been used in the first embodiment, a Fourier transform is used in the second embodiment. In other words, the channel estimation is performed by a two-dimensional Fourier transform, as in FIG Japanese Patent 3802031 ,

Referring to 15 includes the balancing device 200 a balancing device 100 , a scattering pilot channel characteristic calculation unit 120 , a channel characteristic partitioning unit 230 , a repetition appraisal unit 140 , a repetition suppression unit 150 , a channel characteristic combining unit 260 and a compensation unit 170 , The balancing device 200 in the second embodiment differs from the equalizer 100 in the first embodiment, regarding the channel characteristic partitioning unit 230 and the channel characteristic combining unit 260 ,

Referring to 16 includes the channel characteristic partitioning unit 230 first to Nth Fourier transform units 231-1 to 231-N and first to Mth carrier inverse Fourier transform units 232-1 to 232 M ,

The channel characteristic partitioning unit 230 performs a two-dimensional Fourier transform on the scattered pilot channel characteristic using the first to Nth symbol Fourier transform units 231-1 to 231-N and the first to Mth carrier inverse Fourier transform units 232-1 to 232 M through, like in the Japanese Patent 3802031 , The two-dimensional Fourier transform is equivalent to dividing the scattered pilot channel characteristic into N intervals on the frequency deviation axis and M interval on the delay time axis.

The repetition estimation unit 140 identifies repetition components using the distribution of the scattered pilot channel characteristic. The repetition suppression unit 150 extracts the desired components and suppresses the unwanted components in the channel characteristic. The repetition estimation unit 140 and the repetition suppression unit 150 are also referred to as the desired component extraction unit.

Referring to 17 includes the channel characteristic combining unit 260 first to Nth symbol inverse Fourier transform units 261-1 to 261-N , first to Mth carrier Fourier transform units 262-1 to 262 M and a channel characteristic output unit 263 ,

The first to Nth symbol inverse Fourier transform units 261-1 to 261-N in the channel characteristic combiner unit 260 convert the desired components of the channel characteristic from the frequency deviation domain to the symbol time domain by performing inverse Fourier transforms in the frequency deviation direction; the first to Mth carrier Fourier transform units 262-1 to 262 M convert desirable components of channel characteristic from Delay time domain in the carrier frequency domain by performing Fourier transforms in the delay time direction. The inverse Fourier transforms in the frequency deviation direction are equivalent to combining the partitioned channel characteristic on the frequency deviation axis; the Fourier transforms in the delay time direction are equivalent to combining the partitioned channel characteristic on the delay time axis. The channel characteristic output unit 263 delivers the compensation unit 170 the parts of the channel characteristic consisting of the first to Mth carrier Fourier transform units 262-1 to 262 M are output according to the symbols of the Fourier transform unit 110 to the compensation unit 170 supplied received signal.

The dimensionality of the partitioned scattered pilot channel characteristic representing the output from the channel characteristic partitioning unit 230 is different from the dimensionality in the first embodiment, because instead of a filtering process, a Fourier transform is used in the sub-division in the frequency deviation direction. Because the dimensionality of the partitioned scattered pilot channel characteristic has the same dimensionality as in the first embodiment in the channel characteristic combining unit 260 however, repetition estimation and repetition suppression can be performed as in the first embodiment.

Instead of combining the channel characteristic first in the frequency deviation domain and then in the delay time domain, by first performing the inverse Fourier transformation with respect to the frequency deviation, the channel characteristic combining unit 260 operate in the reverse order by first performing the Fourier transform on the delay time axis.

Similarly, the positions of the first to Nth Fourier transform units 231-1 to 231-N and the first to Mth carrier inverse Fourier transform units 232-1 to 232 M be replaced. In other words, first, the scattered pilot channel characteristic may be transmitted through the first through the M-th carrier inverse Fourier transform units 232-1 to 232 M and then through the first to Nth Fourier transform units 231-1 to 231-N ,

In addition to using a Fourier transform instead of a FIR or IIR filtering process to partition the scattered pilot channel characteristic on the frequency deviation axis, the second embodiment may also use a filtering process performed by a plurality of filters instead of a Fourier Transformation to partition the scattered pilot channel characteristic on the delay time axis. The filters used in this case may also be referred to as second filters. The channel characteristic combining unit then uses an adder (also referred to as the second adder) in combining the scattered pilot channel characteristic on the delay time axis.

Generally speaking, if the number of frequency partitions is small and a Fourier transform is large, the use of a filtering process tends to reduce the circuit size and memory usage if the number of frequency partitions is large. The use of a Fourier transform for partitioning the frequency deviation axis as described in the second embodiment accordingly permits fine granular suppression of undesired components and extraction of desired components.

Third embodiment

Referring to 18 includes the balancing device 300 in the third embodiment, a Fourier transform unit 110 , a scattering pilot channel characteristic calculation unit 120 , a channel-channel characteristic partitioning unit 330 , a repetition appraisal unit 340 , a repetition suppression unit 350 , a channel characteristic combining unit 360 , a carrier filter unit 380 and a compensation unit 170 , The balancing device 300 in the third embodiment differs from the equalizer 100 in the first embodiment, by the channel characteristic partitioning unit 330 , the repetition appraisal unit 340 , the repetition suppression unit 350 and the channel characteristic combining unit 360 running processes and by including a carrier filter unit 380 ,

The channel characteristic partitioning unit 330 includes first to Mth symbol filter units 331-1 to 331-M and first to (M-1) th carrier inverse Fourier transform units 332-1 to 332- (M - 1) ,

The repetition estimation unit 340 and repetition suppression unit 350 are also referred to as the desired component extraction unit.

The channel characteristic combiner unit 360 includes a first adding unit 361 , a carrier Fourier transform unit 362 and a second adding unit 363 ,

The first to Mth symbol filter units 331-1 to 331-M are equivalent to the first to Mth symbol filter units described in the first embodiment 131-1 to 131-M , The first to (M-1) th carrier inverse Fourier transformation units 332-1 to 332- (M - 1) are equivalent to the first to (M-1) th carrier inverse Fourier transform units 132-1 to 132- (M - 1) from the first to Mth carrier inverse Fourier transform units described in the first embodiment 132-1 to 132-M , The first adding unit 361 is equivalent to the adder unit described in the first embodiment 161 , The carrier Fourier transform unit 362 is equivalent to the carrier Fourier transform unit described in the first embodiment 162 ,

The second adding unit 363 is a feature of the third embodiment.

The first to Mth symbol filter units 331-1 to 331-M partition through the scatter pilot channel characteristic calculation unit 120 on the frequency deviation axis determined channel characteristics, as the first to Mth symbol filter units 131-1 to 131-M in the first embodiment. The Mth symbol filter unit 131-M which also divides the scattered pilot channel characteristic on the frequency deviation axis has a passband region in which the frequency deviation becomes zero.

Like the first to Mth carrier inverse Fourier transform units 132-1 to 132-M In the first embodiment, the first to (M-1) th carrier inverse Fourier transform units partition 332-1 to 332- (M - 1) the channel characteristic at intervals on the delay time axis by performing inverse Fourier transforms on the outputs of the first to (M-1) th symbol filter units 331-1 to 331- (M - 1) which subdivided the scattered pilot channel characteristic into intervals on the frequency deviation axis. The partitioned region having the minimum absolute frequency deviation value in the partitioned channel characteristic, that is, the output from the Nth symbol filter unit 331-M however, it is not partitioned on the delay time axis.

19 Fig. 10 shows how the power distribution of the channel characteristic is partitioned on the frequency deviation and delay time axes in the third embodiment. In contrast to 12 In the first embodiment, in the region of a small absolute value on the frequency deviation axis, the channel characteristic is not partitioned on the delay time axis.

The repetition estimation unit 340 and the repetition suppression unit 350 Also referred to as the desired component extraction unit, the desired components, such as the repetition estimation unit, extract the desired components 140 and the repetition suppression unit 150 in the first embodiment. For example, use the repetition estimation unit 340 and the repetition suppression unit 350 the parts of the channel characteristic having mutually symmetric power distributions with respect to an axis of symmetry extending in parallel with the delay time axis at a predetermined frequency deviation to identify repetitive components and suppressing the repetition components. The repetition estimation unit 340 and the repetition suppression unit 350 however, do not identify and suppress repetition components in the undivided region extending along the delay time axis, ie in the output from the Mth symbol filter unit 331-M ,

The repetition components in the delay time direction described with reference to 13 in the first embodiment, also occur in the third embodiment. When these delay-time-direction repetition components in the outputs of the first to (M-1) th carrier inverse Fourier transform units 332-1 to 332- (M - 1) may occur, the carrier Fourier transform unit 362 suppress them by inserting zero values into the long distortion components in the partitioned regions on the delay time axis as in the first embodiment.

Because no Fourier transform on the output from the Mth symbol filter unit 331-M is executed, however, the repetition components in the output from the M-th symbol filter unit 331-M is suppressed by a carrier-device filtering process subsequent to the control suppression unit 350 or the second adding unit 363 is executed using a filter with a pass band at the delay time. If this filtering process in the carrier filter unit 380 that after the second adding 363 are then executed, for example, all in the delay time direction in the by the first to M-th symbol filter units 333-1 to 331-M partitioned scattered pilot channel characteristic occurring repetition components by the data processing section 30 be suppressed. For this reason, the carrier direction filtering process is preferably carried out at a stage of that of the second adding unit 363 instead of inserting zero values and performing F_DFT point Fourier transform in the carrier Fourier transform unit 362 ,

The first adding unit 361 combines the partitioned sections of the channel characteristic on the frequency deviation axis and the carrier Fourier transformation unit 362 converts the channel characteristic from the delay time domain into the carrier frequency domain, as does the adder unit 161 and the carrier Fourier transform unit 162 in the first embodiment. The second adding unit 363 then determines the desired channel characteristic for use in equalizing by adding and thus combining the output from the carrier Fourier transform unit 362 for each carrier for output from the Mth symbol filter unit 331-M which is the channel characteristic in the partitioned region of small absolute value on the frequency deviation axis.

There are few channel environments in which repeated components in the region with a small frequency deviation are present in the channel characteristic. The calculation amount can be reduced accordingly by, as in the third embodiment, the part of the channel characteristic with small frequency deviation is not partitioned into different regions on the delay time axis.

The exemplary configuration of the third embodiment in the above description is based on the configuration of the first embodiment, but the third embodiment may also be configured based on the configuration of the second embodiment. In this variation of the third embodiment, the channel characteristic is partitioned as in FIG 20 is shown, and is the channel characteristic combining unit as in 21 shown configured.

The channel characteristic partitioning unit 330 # in 20 includes first to Nth symbol Fourier transform units 333-1 to 333-N and first to (M-1) th carrier inverse transform units 334-1 to 334- (M - 1) ,

The first to Nth symbol Fourier transformation units 333-1 to 333-N are equivalent to the first to Nth Fourier transform units 231-1 to 231-N which are described in the second embodiment. The first to (M-1) th carrier inverse transformation units 334-1 to 334- (M - 1) are equivalent to the first to (M-1) th carrier inverse Fourier transform units 132-1 to 132- (M - 1) together with the first to Mth carrier inverse transformation units 132-1 to 132-M , or the first to (M-1) th carrier inverse Fourier transform units 232-1 to 232- (M - 1) from the first to Mth carrier inverse Fourier transform units 232-1 to 232 M which are described in the second embodiment.

The channel characteristic partitioning unit 330 # partitions the controlled pilot channel characteristic by performing a two-dimensional Fourier transform on the scattered pilot channel characteristic using the first through the N-th Fourier transform units 333-1 to 33-N and the first to (M-1) th carrier inverse Fourier transform units 334-1 to 334- (M - 1) , like in Japanese Patent 3802031 except that the frequency deviation partition with minimum absolute frequency deviation value remains unpartitioned on the delay time axis.

Referring to 21 The channel characteristic combining unit 360 # includes first through (M-1) th carrier inverse Fourier transform units 364-1 to 364- (M - 1) first to Nth carrier inverse Fourier transform units 365-1 to 365-N and a channel characteristic output unit 366 ,

The first to (M-1) th carrier inverse Fourier transformation units 364-1 to 364- (M - 1) are equivalent to the first to (M-1) th carrier inverse Fourier transform units 262-1 to 262- (M - 1) from the first to M-th Fourier transformation units described in the second embodiment 262-1 to 262 M , The first to Nth symbol inverse Fourier transform units 365-1 to 365-N are equivalent to the first to Nth symbol inverse Fourier transform units 261-1 to 261-N which are described in the second embodiment. The channel characteristic output unit 366 is equivalent to the channel characteristic output unit 263 described in the second embodiment.

The first to (M-1) th carrier inverse Fourier transformation units 364-1 to 364- (M - 1) and first to Nth symbol inverse Fourier transform units 365-1 to 365-N in the channel characteristic combining unit 360 #, the partitioned channel characteristic returns to the symbol time and carrier frequency axes, and the channel characteristic output unit 366 in the channel characteristic combining unit 360 # combines the parts of this channel characteristic and gives the channel characteristic to the equalizing unit 170 out. The first to Nth symbol inverse Fourier transform units 365-1 to 365-N also combine the region with the minimum absolute frequency deviation value in the partitioned channel characteristic.

The above variation of the third embodiment based on the configuration of the second embodiment leaves the single partitioned region unpartitioned with minimum absolute frequency deviation values on the delay time axis. In another possible variation, a plurality of regions having small absolute values on the frequency deviation axis on the delay time axis are left undivided.

Fourth embodiment

The fourth embodiment uses the characteristics of repetition components to perform the calculations, which has partitioned the channel characteristic on the frequency deviation axis in the first to third embodiments, in a simpler manner. This allows a considerable reduction in the necessary number of symbol filter units and carrier inverse Fourier transform units as compared with the channel characteristic partitioning units 130 . 330 in the first and third embodiments.

Referring to 22 includes the channel characteristic partitioning unit 430 in the fourth embodiment, first to Zth characteristic partitioning units 431-1 to 431-Z , The parameter Z is a positive integer satisfying the following equation (10).

In Equation (10), L_z + 1 is a number of sets of outputs from the zth characteristic partitioning unit 431-z , The first to Zth characteristic partitioning units 431-1 to 431-Z are described by describing the zth characteristic partitioning unit 431-z (where z is a positive integer such that 1≤z≤Z).

Referring to 23 includes the zth characteristic partitioning unit 431-z a symbol filter unit 432 , a carrier inverse Fourier transform unit 433 , and first to (L_z) th delay time shift symbol time rotation units 434-1 to 434 (L_z) ,

First, partition the symbol filter unit 432 and the carrier inverse Fourier transform unit 433 the scattered pilot channel characteristic on the frequency deviation and delay time axes by executing the same processes as those performed in the first embodiment. More specifically, the symbol filter unit shares 432 the scattered pilot channel characteristic in components included in a plurality of first intervals on the frequency deviation axis. The Carrier Inverse Fourier Transformation Unit 433 divides the components that pass through the symbol filter unit 432 are determined in components at a plurality of second intervals on the delay time axis.

Next, the first to (L_z) th delay time shift symbol time rotation units shift 434-1 to 434 (L_z) the partitioned components in the carrier inverse Fourier transform unit 433 output a channel characteristic by a predetermined first amount on the frequency deviation axis and perform recursive shifts by a predetermined second amount on the delay time axis. For example, the first to (L_z) th delay time shift symbol time rotation units 434-1 to 434 (L_z) recursive shifts by an integer multiple of Tu / 12 on the delay time axis, followed by a phase rotation around exp (j2πm / 4) (where m is an index denoting the mth symbol) at the partitioned components in the one of Vehicle inverse Fourier transform unit 433 output channel characteristic. Below is a detailed description given.

As in 24 Because the positions of the repetitive components are typically shifted on the frequency deviation and delay time axes in the scattered pilot channel characteristic, the repetition components with shifts occur by (1/4) Ts on the frequency deviation axis and Tu / 12 on the time delay axis. Because of this property, the channel characteristic in region B becomes in 25 is determined by shifting region A by (1/4) Ts on the frequency deviation axis and can be calculated by performing a recursive shift on the channel characteristic in region A by Tu / 12 on the delay time axis.

Because the first embodiment uses a plurality of symbol filters to partition the scattered pilot channel characteristic on the frequency deviation axis, it is incapable of making shifts directly on the frequency deviation axis. In contrast, the fourth embodiment performs a phase rotation on the symbol time axis on the partitioned component of the channel characteristic in a frequency deviation band to calculate the channel characteristic in another frequency deviation band.

As is well known, a shift performed on the frequency axis after an inverse Fourier transform becomes a phase rotation in a time domain. The channel characteristic on the symbol time axis t1 (1 is an index indicating the first symbol) obtained by performing an inverse Fourier transform on the frequency deviation fd is therefore rotated by exp (j2πt1 (1/4) Ts). Since t1 is an integer multiple of Ts, exp (j2πt1 (1/4) Ts) equals exp (j2πm / 4). Moreover, because the phase rotation factor exp (j2πm / 4) takes only four values (1, j, -1, -j), the calculation can be performed in a simple manner, which makes it possible to reduce the circuit size.

The channel characteristic in the partitioned region B on the delay time axis will be determined by taking the channel characteristic in region A passing through the symbol filter unit 432 and using the carrier inverse Fourier transform unit to partition it on the delay time axis, then perform a recursive shift on the partitioned channel characteristic by Tu / 12 in the delay time direction and multiplying the resulting channel characteristic by the phase rotation factor exp (2πm / 4) determined.

The partitioned scattered pilot channel characteristic may similarly be for other repeated components at distances equal to integer multiples of (1/4) Ts from the partitioned channel characteristic in region A by performing recursive shifts by integer multiple of Tu / 12 on the delay time axis, and multiplied by an integer multiple of the phase rotation factor exp (j2πm / 4) on the partitioned channel characteristic in region A.

Combining the above calculations allows for a reduction by substantially half in symbol filtering and carrier inverse Fourier transform processing as compared to M frequency partitions. For example, the symbol filter unit 432 in each of the first to Zth characteristic partitioning units 431-1 to 431-Z Partition a region equal to half the total region to be partitioned into M / 2 partitions.

As described above, the fourth embodiment uses the characteristics that repetitive components occur uniformly in the channel characteristic to add simple calculations that allow a large reduction in the necessary amount of filtering and inverse Fourier transformation processing used to drive the channel characteristics. To partition characteristics.

Fifth embodiment

The fifth embodiment includes a carrier inverse Fourier transform unit and a carrier Fourier transform unit that reduce the number (N) of partition on the delay time axis by reducing the number of points in the Fourier transform.

The carrier inverse Fourier transformation unit in the fifth embodiment may include the first to Mth carrier inverse Fourier transformation units described in the first embodiment 132-1 to 132-M , the first to Mth carrier inverse Fourier transform units described in the second embodiment 232-1 to 232 M , the first to (M-1) th carrier inverse Fourier transform units described in the third embodiment 332-1 to 332- (M - 1) or the carrier inverse Fourier transform unit 433 in the fourth embodiment.

The carrier Fourier transform unit in the fifth embodiment may be the carrier Fourier transform unit described in the first embodiment 162 , the first to Mth carrier Fourier transform units described in the second embodiment 262-1 to 262 M or the carrier Fourier transform unit described in the third embodiment 362 replace.

Referring to 26 , includes the carrier inverse Fourier transform unit 532 in the fifth embodiment, a carrier partitioning unit 532a and an inverse Fourier transform unit 532b ,

As in 27 shown partitioning the carrier partitioning unit 532a the channel characteristic on the carrier frequency axis at N-point intervals. For example, the carrier partitioning unit partitions 532a the scattered pilot channel characteristic in components at a plurality of predetermined intervals on the frequency deviation axis. The by the carrier partitioning unit 532a Partitioned intervals generated are also referred to as third intervals.

The inverse Fourier transformation unit 532b partitions the N-point components of the partitioned channel characteristic in the delay time axis direction by performing an inverse Fourier transform, which converts them into components on the delay time axis. While the first, second, third, and fourth embodiments performed F_DFT / 3-point inverse Fourier transforms, by partitioning the channel characteristic into N-point intervals, the fifth embodiment reduces the number of points used in the inverse Fourier transform which reduces the amount of computation and memory usage.

The result of an N-point inverse Fourier transform will be described with reference to FIG 27 to 29 described.

If an inverse Fourier transform to in 27 If the scattered pilot channel characteristic shown is executed without partitioning, the in 28 shown channel characteristic on the delay time axis determined. The maximum detectable delay time is determined as the inverse of the interval between carriers into which the scattered pilot is inserted and a channel characteristic for the number of points at which the inverse Fourier transform was performed.

The contrasting result of an N-point inverse Fourier transform performed on the partitioned channel characteristic after partitioning will be explained with reference to FIG 29 described. The maximum detectable delay time is again the inverse of the interval between carriers in which the scattered pilot is inserted, and a channel characteristic with the number of points in the inverse Fourier transform is determined. The N-point inverse Fourier transform unit 532b Effectively partitioned the channel characteristic into N points on the delay time axis.

The fifth embodiment also includes a repetition estimation unit 140 , a repetition suppression unit 150 , an adding unit 161 (in 26 not shown) that perform the same processes as in the first embodiment. However, in order to identify repetition components, the repetition estimation unit may 140 the power distribution in the channel characteristic using all one-symbol units as a result of the inverse Fourier transform unit 532b calculate the executed inverse Fourier transform.

Referring to 30 includes the Fourier transform unit 562 in the fifth embodiment, a Fourier transform unit 562a and a carrier combination unit 562b ,

By performing a Fourier transform on the channel characteristic components in each of the N-point partitioned processing units after suppressing unwanted components in the pre-stage, thereby converting the components on the delay time axis into components on the carrier frequency axis, the Fourier transform unit 562a the channel characteristic in the delay time axis direction.

The carrier combination unit 562b combines the partitioned channel characteristic by returning the partitioned channel characteristic on the delay time axis to a channel characteristic on the carrier frequency axis. In other words, the carrier combination unit decreases 562b the combined components produced by the Fourier transform unit 562a are generated and combined in the frequency deviation axis direction.

By partitioning the channel characteristic on the carrier frequency axis as described above, prior to performing the inverse Fourier transform, the equalizer in the fifth embodiment can reduce the amount of computation and memory usage.

The carrier partitioning unit 532a can partition the channel characteristic into overlapping components on the carrier frequency axis, which are then multiplied by window functions, as in FIG 31 shown before at the inverse Fourier transformation unit 532b be entered. A window function multiplier unit 632c which outputs from the carrier partitioning unit 532a multiplied by window functions, is then used between the carrier partitioning unit 532a and the inverse Fourier transform unit 532b in the carrier inverse Fourier transform unit 632 inserted as in 32 shown. Window multiplication reduces distortion due to discontinuities at the ends of the partitioned intervals. The window function shape is preferably selected so that the sum of the window functions is constant at all points on the frequency deviation axis. Trigonometric window functions are an example, but the present invention is not limited to trigonometric window functions.

Alternatively, both the carrier partitioning unit 532a as well as the carrier combination unit 562b Multiply the channel characteristic in each interval by the square root of the window function. The by the carrier combination unit 562b executed window function multiplication can then correct the distortion due to suppression and extraction in the partitioned regions by the repetition suppression unit 150 to reduce.

In the discussion above, the inverse Fourier transform unit result 532b , the repetition suppression unit 150 , the adding unit 161 and the Fourier transform unit 562a Time division processing in each of the partitioned intervals provided by the carrier partitioning unit 532a but these intervals can be processed in parallel.

As described above, when an inverse Fourier transform is used to partition the channel characteristic on the delay time axis, the method of the fifth embodiment reduces the number of partitions on the delay time axis by partitioning the channel characteristic on the carrier frequency axis. The amount of computation and memory usage involved in the Fourier transform process can be reduced accordingly. Distortion by partitioning the channel characteristic and extracting desired components of the channel characteristic can be reduced by performing window function multiplication for partitioning the channel characteristic on the carrier frequency axis.

Those skilled in the art will recognize that other variations are possible within the scope of the invention, which is defined by the appended claims.

Claims (17)

Balancing device ( 100 ), comprising: a first Fourier transform unit ( 110 ) for transforming a received signal into a frequency domain; a channel characteristic calculation unit ( 120 ) for calculating a first channel characteristic concerning the pilot carrier, the first channel characteristic being expressible in relation to a frequency deviation axis and a delay time axis; a channel characteristic partitioning unit ( 130 ) for partitioning the first channel characteristic into components arranged in a plurality of regions corresponding to a plurality of first intervals on the frequency deviation axis and a plurality of second intervals on the delay time axis; a desired component extraction unit ( 140 . 150 ) for extracting desired components from the channel characteristic partitioning unit (s) ( 130 ) components by identifying the desired components from a distribution symmetry of the components defined by the channel characteristic partitioning unit (FIG. 130 ) with respect to an axis of symmetry extending parallel to the delay time axis at a predetermined frequency deviation, or by identifying unwanted components from the distribution symmetry of the components detected by the channel characteristic partitioning unit (FIG. 130 ) is detected with respect to the axis of symmetry that extends parallel to the delay time axis at a predetermined frequency deviation, and suppressing the unwanted components; a channel characteristic combiner unit ( 160 ) for generating a second channel characteristic by combining by the desired component extraction unit ( 140 . 150 ) extracted desired components; and a compensation unit ( 170 ) for using the second channel characteristic to detect channel distortion in the received signal after transformation into the frequency domain by the first Fourier transform unit ( 110 ), wherein the plurality of regions are arranged symmetrically about the axis of symmetry that extends parallel to the delay time axis at the predetermined frequency deviation, and when the components in two regions in mutually symmetrical positions with respect to the axis of symmetry have significantly asymmetric power values that are desired Component Extraction Unit ( 140 . 150 ) identifies the components in at least one of the two regions as including an undesired component. Balancing device ( 100 ) according to claim 1, wherein the desired component extraction unit ( 140 . 150 ) determines that the components in the two regions in symmetrical positions with respect to the axis of symmetry have significantly asymmetric power values when their power values differ by more than a predetermined threshold. Balancing device ( 100 ) according to claim 1 or 2, wherein the desired component extraction unit ( 140 . 150 ) identifying one of the two regions having a higher power value than an undesired component. Balancing device ( 100 ) according to one of claims 1 to 3, wherein the desired component extraction unit ( 140 . 150 ) reduces the components in the region identified as containing an undesired component to zero. Balancing device ( 100 ) according to one of claims 1 to 3, wherein the desired component extraction unit ( 140 . 150 ) calculates an inclusion ratio of unwanted components based on component power in the two regions in mutually symmetric positions with respect to the axis of symmetry, calculates suppression ratios for the regions into which the channel characteristic partitioning unit (16) 130 ) divides the first channel characteristic, the suppression ratios having values decreasing as the inclusion ratio increases, and multiplying the components detected by the channel characteristic partitioning unit by the suppression ratios of their regions. Balancing device ( 100 ) according to any one of claims 1 to 5, wherein the desired component extraction unit ( 140 . 150 ) calculates a noise power from the components detected by the channel characteristic partitioning unit ( 130 ), and subtracts the noise power from the power values of the components in the two regions in mutually symmetrical positions with respect to the axis of symmetry, before determining whether the components have substantially asymmetrical power values. Balancing device ( 100 ) according to one of claims 1 to 6, wherein the channel characteristic partitioning unit ( 130 ) a plurality of first filters ( 131-1 to 131-M ) is used to separate the components at different first intervals on the frequency deviation axis, and the channel characteristic combining unit ( 160 ) a first adder ( 161 ) is used to combine the desired components on the frequency deviation axis. Balancing device ( 200 ) according to one of claims 1 to 6, wherein the channel characteristic partitioning unit ( 230 ) uses a Fourier transform to partition the first channel characteristic into the components on the frequency deviation axis and the channel characteristic combining unit (FIG. 260 ) uses a Fourier transform to combine the desired components on the frequency deviation axis. Balancing device ( 100 ) according to one of claims 1 to 8, wherein the channel characteristic partitioning unit ( 130 ) uses an inverse Fourier transform to introduce the first channel characteristic into the Partition components on the delay time axis, and the channel characteristic combiner unit ( 160 ) uses a Fourier transform to combine the desired components on the delay time axis. Balancing device ( 100 ) according to one of claims 1 to 8, wherein the channel characteristic partitioning unit ( 130 ) uses a plurality of second filters to partition the first channel characteristic into the components on the delay time axis, and the channel characteristic combining unit ( 160 ) uses a second adder to combine the desired components on the delay time axis. Balancing device ( 100 ) according to one of claims 1 to 8, wherein: the channel characteristic partitioning unit ( 130 ) further comprises: a carrier partitioning unit ( 532a ) for partitioning by the channel characteristic calculation unit ( 120 ) is detected in components of a plurality of predetermined third intervals on the frequency deviation axis, and an inverse Fourier transform unit ( 532b ) for performing inverse Fourier transforms on the carrier partitioning unit ( 532a ) components to partition the components on the delay time axis; and the channel characteristic combining unit ( 160 ) further comprises: a second Fourier transformation unit ( 562a ) for combining the components on the delay time axis after the desired component extraction unit ( 140 . 150 ) has suppressed the unwanted components, and a carrier combination unit ( 562b ) for performing an addition operation for combining, on the frequency deviation axis, the components combined by the second Fourier transformation unit ( 562a ). Balancing device ( 100 ) according to claim 11, wherein: the carrier partitioning unit ( 532a ) determined by the channel characteristic calculation unit ( 120 ) determined first channel characteristic divided into mutually overlapping components; the channel characteristic partitioning unit ( 130 ) a window function multiplying unit ( 632c ) for multiplying the components into which the carrier partitioning unit ( 532a ) divides the first channel characteristic with a predetermined window function; the inverse Fourier transformation unit ( 532b ) performs the inverse Fourier transforms on the components after the window function multiplication unit ( 632c ) multiplied by the predetermined window function; and the carrier combination unit ( 562b ) generated by the second Fourier transformation unit ( 562a ) overlaps combined components in combining the components combined by the second Fourier transform unit. Balancing device ( 100 ) according to one of claims 1 to 9, 11 or 12, wherein the channel characteristic partitioning unit ( 130 ) a plurality of characteristic partitioning units ( 431-1 to 431-Z ), each characteristic partitioning unit comprising: a filter unit ( 432 ) for partitioning by the channel characteristic calculation unit ( 120 ) calculated first channel characteristics in components included in the first intervals on the frequency deviation axis; a second inverse Fourier transformation unit ( 433 ) for performing an inverse Fourier transform, which further comprises passing through the filter unit ( 432 ) components partitioned into components included at a plurality of second intervals on the delay time axis; and a plurality of rotation units ( 431-1 to 431-L_z ) for performing displacements by a predetermined first amount on the frequency deviation axis at each of the components detected by the second inverse Fourier transform unit (10). 432 at the second intervals, and performing a recursive shift by a predetermined second amount on the components in a given interval on the frequency deviation axis, thereby shifting the components to different intervals to determine the components in each of the second intervals. Balancing device ( 100 ) of one of claims 1 to 13, wherein the characteristic partitioning unit ( 130 ) Components of the first, through the channel characteristic calculation unit ( 120 ) channel characteristics having absolute values smaller than a predetermined threshold on the frequency deviation axis, prevented by the desired component extraction unit ( 140 . 150 ) to be processed. Balancing device ( 100 ) according to one of claims 1 to 14, wherein the channel characteristic calculation unit ( 120 ) Components with carrier frequencies that do not contain the pilot carrier are reduced to zero. Balancing device ( 100 ) according to one of claims 1 to 14, wherein the channel characteristic combining unit ( 160 ) converts the combined desired components to the frequency domain, then reduces components at carrier frequencies not including the pilot carrier to zero to produce the second channel characteristic. Compensation method comprising: Transforming a receive signal containing a pilot carrier into the frequency domain; Calculating a first channel characteristic concerning the pilot carrier, the first channel characteristic being expressible with respect to a frequency deviation axis and a delay time axis; Partitioning the first channel characteristic into components in a plurality of regions corresponding to a plurality of first intervals on the frequency deviation axis and a plurality of second intervals on the delay time axis; Extracting desired components from the components into which the first channel characteristic is divided by identifying the desired components from distribution symmetry of the components with respect to an axis of symmetry that extends parallel to the delay time axis at a predetermined frequency deviation, or by identifying unwanted components the distribution symmetry of the components with respect to the symmetry axis parallel to the delay time axis at the predetermined frequency deviation and suppressing the unwanted components; Generating a second channel characteristic by combining the desired components; and Using the second channel characteristic to compensate for channel distortion in the received signal as transformed into the frequency domain, wherein the plurality of regions are arranged symmetrically about the axis of symmetry that extends parallel to the delay time axis at the predetermined frequency deviation, and when the components in two regions at symmetrical positions with respect to the axis of symmetry have significantly asymmetric power values, the components in at least one of two regions are identified as containing an undesired component.

Images