CN1193501C - Ghost eliminating equalizer - Google Patents

Ghost eliminating equalizer Download PDF

Info

Publication number
CN1193501C
CN1193501C CNB008195420A CN00819542A CN1193501C CN 1193501 C CN1193501 C CN 1193501C CN B008195420 A CNB008195420 A CN B008195420A CN 00819542 A CN00819542 A CN 00819542A CN 1193501 C CN1193501 C CN 1193501C
Authority
CN
China
Prior art keywords
coefficient
described
multiply
ghost image
step
Prior art date
Application number
CNB008195420A
Other languages
Chinese (zh)
Other versions
CN1452809A (en
Inventor
R·W·西塔
S·M·洛普莱斯托
夏劲松
Original Assignee
真尼诗电子有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 真尼诗电子有限公司 filed Critical 真尼诗电子有限公司
Priority to PCT/US2000/013620 priority Critical patent/WO2001089087A1/en
Publication of CN1452809A publication Critical patent/CN1452809A/en
Application granted granted Critical
Publication of CN1193501C publication Critical patent/CN1193501C/en

Links

Classifications

    • HELECTRICITY
    • H03BASIC ELECTRONIC CIRCUITRY
    • H03HIMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
    • H03H17/00Networks using digital techniques
    • H03H17/02Frequency selective networks
    • H03H17/06Non-recursive filters

Abstract

The present invention relates to an equalizer which essentially eliminates ghost images of received main signals through the following steps: 1, a coefficient b (b can be equal to 1) is applied to the received main signals and the ghost images so that the received main signals and the ghost images can be modulated, and thus, the received main signals are not equal to the ghost images; 2, a coefficient a is applied to the modulated received main signals and the modulated ghost images so that the ghost images can be essentially eliminated; 3, according to a window function, a coefficient c is applied to the received main signals containing no ghost image essentially so that the modulation which is applied to the received main signals by the coefficient b can be removed.

Description

The method of equalization channel

Technical field of the present invention

The present invention is directed to eliminate in fact up to and comprise the equalizer of the signal ghost image of 100% ghost image.

Background of invention

Usually because signal arrives receiver by different transmission channels, in described receiver, produce ghost image.For example, in having the system of individual transmitter, because the multipath transmission of signal may take place in signal reflex.That is to say that receiver receives the one or more reflections that send signal and described transmission signal.And for example, having the use same carrier in the system of a plurality of transmitters of receiver emission same signal, the multipath transmission of signal may take place.The network of supporting such transmission is typically called Single Frequency Network.

When signal arrives receiver by two or more different transmission channels, produce interference pattern.In frequency domain, by showing this interference pattern along the variable signal amplitude of frequency axis.The interference pattern of generation has been shown among Fig. 1 when ghost image is 100%.This interference pattern has 0 amplitude at some frequency place or near 0 amplitude.Therefore, any information that is comprised near the main signal that is received in these frequencies might be lost, because near the signal to noise ratio these frequencies is lower than available threshold value.

Design various systems and handled the problem that causes by ghost image.For example, spread spectrum system spreads all on sizable bandwidth by the data expansion that will be sent, with the problem of enough processing 100% ghost images.Therefore,, 100% ghost image may lose some information, but because it is expanded to the high probability of the frequency that does not correspond to described amplitude 0 even meaning near corresponding to the frequency of amplitude 0, but restore data element still.Unfortunately, for many application, the data rate that is associated with spread spectrum system is typically too low.(described data rate is defined as the number of the data bit of every hertz of channel width.)

It also is known using matched filter so that handle the ghost image problem in receiver.In this method, data are transmitted as data vector.Described matched filter makes received data with relevant corresponding to the reference vector of the possible data vector that can be transmitted.Received main signal is relevant with described reference vector corresponding to the data vector that is sent, has produced big peak value, and received main signal with other may reference vector relevant, produced small leak.Therefore, in receiver, can easily detect the data vector that is sent.Unfortunately, for many application, the data rate that is associated with the use of described matched filter is still typically too low.

When requiring high data rate, use equalizer in the receiver of being everlasting, so that reduce the ghost image of main signal.A classical example of time-domain equalizer is the FIR filter.The FIR filter carries out convolution with general shown its response h (t) among Fig. 2 with received signal.The signal that is received comprises the ghost image of described main signal and described main signal.Described FIR filter produces the output of the big peak value with the described main signal of expression.The ghost image of main signal has small component in the output of described FIR filter.Yet, as shown in Figure 2, the values of tap a of FIR filter 1, a 2, a 3... depend on the value of a, and in order to use the FIR filter ideally to eliminate 100% ghost image, the value a of described FIR filter response must be near 1.When described value a near 1 the time, the value of the tap of described FIR filter is not progressively to reduce towards 0.Therefore, if eliminate 100% ghost image, the FIR filter endless that becomes makes the FIR filter can not eliminate 100% ghost image.

The example of frequency-domain equalizer 10 has been shown among Fig. 3.This frequency-domain equalizer 10 comprises fast Fourier transform (FFT) module 12, and this module is carried out fast Fourier transform on the signal that is received, so that received conversion of signals is arrived frequency domain.Multiplier 14 usefulness comprise a row coefficient A iThe compensation vector frequency domain output of multiply by FFT module 12.Inverted-F FT module 16 is carried out inverted-F FT on the multiplication result that is produced by described multiplier 14, so that described multiplication result is transformed into time domain.

Fig. 4 has illustrated can be by one group of exemplary coefficient A of described frequency-domain equalizer 10 uses iSelect coefficient A like this i, make when the FFT of they and received signal being multiplied each other described coefficient A by multiplier 14 iEliminate the ghost image in the received signal, only stayed main signal.Should be noted that described coefficient A iShould have unlimited amplitude at the frequency place that interference pattern has 0 amplitude.Yet, in fact can not make described coefficient A iInfinitely.Therefore, cut off described coefficient A at these frequency places i, this means the information of having lost at the frequency place of described cut-out in the main signal that is received, making the output of inverted-F FT module 16 become only is send data approximate.

And using empty protection between the vector of the frequency-domain equalizer 10 that is used for Fig. 3 is known at interval.Figure 5 illustrates described protection at interval, and provide described protection so at interval, make that the ghost image of vector that is received and the vector that is received is not overlapping, because so overlapping intersymbol interference that causes.Thereby described protection at interval should be the same long with the ghost image of expection at least.Use the circulation continuation of vector, periodic surface is arranged so that give received main signal.Therefore, the Fourier transform of the fast Fourier transform of received signal and received signal looks and is equal to.

The invention that is disclosed in the U. S. application 09/158,730 of application on September 22nd, 1998 is at the equalizer that overcomes one or more the problems referred to above.According to this invention, vector field equalizer 20 as shown in Figure 6 relies on vector that the transmission data are distributed in time and the frequency, makes to come down at random at vector described in time domain and the frequency domain.Therefore, in the channel of high ghost image, can recover all data and only have little noise and strengthen, and the noise of existing any enhancing approaches white noise.

Described vector field equalizer 20 reverse vector territory conversion 22 and vector field conversion 24, they by channel 26 separately.Therefore, described reverse vector territory conversion 22 can be the part of transmitter, and the part that described vector field conversion 24 can be a receiver.Matrix multiplication is carried out in reverse vector territory conversion 22 between input block and transformation matrix.Described input block can comprise any amount of data element of arranging with row.These data elements can be bit, symbol or any other suitable data entity.Described transformation matrix comprises a plurality of vectors of arranging with row, and each vector of transformation matrix preferably has and a considerable amount of data elements of input block.Similarly, the number of the vector of described transformation matrix is preferably suitable with the number of the data element in the described input block.Therefore, if 256 data elements are arranged in the input block, then transformation matrix should preferably have 256 each have the vector of 256 elements.The output of described reverse vector territory conversion 22 is the output blocks that have with a considerable amount of data elements of data element of input block.Thereby if 256 data elements are arranged in input block, then output block has 256 data elements.

Described vector field conversion 24 is in the main signal that is received and a plurality of receiver vector V RBetween carry out matrix multiplication.For example, press row vector and receive the data that send by channel 26.During matrix multiplication, described vector field conversion 24 receiver vector V RThe 1st row in respective components multiply by each component of the row vector that is received, and, produce vector r with output in described vector field conversion 24 to the multiplication result summation iThe 1st component r 1Then, described vector field conversion 24 receiver vector V RThe 2nd row in respective components multiply by each component of the row vector that is received, and to the multiplication result summation, to produce output vector r iThe 2nd component r 2, the rest may be inferred.

Suppose not have the channel distortion such as may causing, and the matrix multiplication generation original input data piece that has described vector field conversion 24 to carry out is used and the identical vector of reverse vector territory conversion 22 employed vectors in hypothesis vector field conversion 24 by channel disturbance.Yet if there is channel distortion, the actual output block that is produced by described vector field conversion 24 will be not equal to the original input data piece.Therefore, call the training session, adjust the vector of described vector field conversion 24 therein according to channel distortion, make to exist under the situation of channel distortion, recover the data of original input data piece.

The invention of U. S. application 09/158,730 can reasonably well be worked.Yet the present invention has produced similar result with less calculating.

Summary of the invention

According to an aspect of the present invention, a kind of equalizer that is used for process data block comprises finite filter and preprocessor.Described finite filter has output, and described finite filter is arranged to eliminate ghost image in fact from received signal, so that the signal that does not have ghost image in fact is provided in output place.Preprocessor is arranged to apply window function to the output of described finite filter.Described window function has the duration of the duration that equals data block in fact.

According to a further aspect in the invention, a kind of equalizer comprises preprocessor, finite filter and preprocessor.Described preprocessor applies coefficient b to the ghost image of main signal that is received and the main signal that is received, so that main signal and described ghost image that modulation is received.Described finite filter applies coefficient a to modulated main signal that receives and ghost image, so that eliminate described ghost image in fact.Described preprocessor applies coefficient c as window function to the main signal that is received in the output of described finite filter, forces at modulation on the main signal that is received so that remove by described coefficient b.

According to another aspect of the invention, a kind of method of eliminating the ghost image of the main signal that comprises data block that is received in fact, may further comprise the steps: a) main signal and the ghost image that is received applied coefficient a, so that eliminate described ghost image in fact, thereby produce the signal that does not have ghost image in fact, wherein said coefficient a has the duration than the longer duration of data block; And b) signal that does not have ghost image is in fact applied coefficient c, wherein said coefficient c forms the window function of the duration with the duration that equals data block in fact.

The accompanying drawing summary

By the detailed consideration of following connection accompanying drawing, these and other characteristics and advantage of the present invention will become clearer, in the accompanying drawing:

Fig. 1 shows the interference pattern that can produce when two signals in the substantially identical time reception same frequency band of receiver;

Fig. 2 has illustrated the response of the FIR filter that the time-domain equalizer in the Chang Zuowei receiver uses in order to eliminate ghost image;

Fig. 3 has illustrated the frequency-domain equalizer that is used for receiver in order to eliminate ghost image;

Fig. 4 has illustrated in order to eliminate ghost image by the employed one group of exemplary coefficient A of the frequency-domain equalizer of Fig. 3 i

Fig. 5 has illustrated the protection interval that can use between the emission vector in the system that uses equalizer;

Fig. 6 has illustrated the equalizer that comprises vector field transfer pair (being vector field conversion and the conversion of reverse vector territory);

Fig. 7 has illustrated the 1st embodiment according to equalizer of the present invention;

Fig. 8 has illustrated the 2nd embodiment according to equalizer of the present invention;

Fig. 9 has illustrated the 1st embodiment of response of the preprocessor of the equalizer shown in Fig. 7 and 8;

Figure 10 has illustrated the response of the acoustic convolver of the equalizer shown in Fig. 7;

Figure 11 has illustrated the real part of response of the multiplier of the equalizer shown in Fig. 8;

Figure 12 has illustrated the imaginary part of response of the multiplier of the equalizer shown in Fig. 8;

Figure 13 has illustrated the 1st embodiment of response of the preprocessor of the equalizer shown in Fig. 7 and 8;

Figure 14 has illustrated the 2nd embodiment of response of the preprocessor of the equalizer shown in Fig. 7 and 8;

Figure 15 has illustrated the 2nd embodiment of response of the preprocessor of the equalizer shown in Fig. 7 and 8;

Figure 16 is the main signal that received and the time domain explanation of ghost image thereof;

Figure 17 has illustrated the output of preprocessor response in time domain;

Figure 18 has illustrated the 3rd embodiment according to equalizer of the present invention;

Figure 19 has illustrated the 4th embodiment according to equalizer of the present invention;

Figure 20 has illustrated the real part of response of the multiplier of the equalizer shown in Figure 19;

Figure 21 has illustrated the imaginary part of response of the multiplier of the equalizer shown in Figure 19; And

Figure 22 has illustrated the response of the preprocessor of the equalizer shown in Figure 19.

Describe in detail

Illustrated among Fig. 7 according to equalizer 100 of the present invention, it comprises preprocessor 102, finite filter 104 and preprocessor 106.The preprocessor 102 usefulness coefficient b of equalizer 100 multiply by the signal that comes self-channel that is received.In Fig. 7, the signal that comes self-channel that is received is appointed as the data input.Preprocessor 102 is the received main signal of modulation and the modulation operations of ghost image thereof, makes ghost image less than received main signal.Therefore, described ghost image no longer is 100% ghost image.

As shown in Figure 7, finite filter 104 is acoustic convolvers 108.Therefore, in acoustic convolver 108, use the multiplication result of coefficient a convolution preprocessor 102.From the multiplication result of preprocessor 102, eliminated ghost image by the convolution that acoustic convolver 108 carries out.

Preprocessor 106 usefulness coefficient c multiply by the convolution results from acoustic convolver 108, make that the output of preprocessor 106 is the data that are transmitted in the channel.In Fig. 7, the data of output place of preprocessor 106 are appointed as data output.Preprocessor 106 reverses the effect of the modulation that preprocessor 102 is forced, and the output of acoustic convolver 108 is applied window function.This window function has the duration that equates with the duration of data input block in fact.

Because the output of 106 pairs of acoustic convolvers 108 of preprocessor applies window function, make the data IOB temporarily mate corresponding data input block, for example, can be embodied as the FIR filter to acoustic convolver such as above-mentioned described like that about Fig. 2.That is to say that owing to the window function that is applied by preprocessor 106, the quantity of the tap of FIR filter needs not be unlimited, and can be limited to rational quantity.For example, these taps can have the duration of the duration that doubles the data input block.

The main signal that provides controller 109 to measure separately to be received and the time interval d of ghost image thereof.As described below, can be used for form factor b, a and c to interval d.Controller 109 provides coefficient b to preprocessor 102, provides coefficient a to acoustic convolver 108, and provides coefficient c to preprocessor.Controller 109 also is synchronized to every blocks of data by described equalizer 100 to preprocessor 102, acoustic convolver 108 and preprocessor 106.Available protection separates per two blocks of data at interval.

Fig. 8 has illustrated and the equalizer 110 of equalizer 100 equivalences shown in Fig. 7 that it comprises preprocessor 112, finite filter 114 and preprocessor 116.Finite filter 114 comprises fast Fourier transform 118, multiplier 120 and inverse FFT 122.Thereby finite filter 104 is operated in time domain.On the contrary, because multiplier 120 is exported the frequency domain that complex coefficient A (following) is applied to fast Fourier transform 122, finite filter 114 is operated in frequency domain in fact.

Therefore, the preprocessor 112 usefulness coefficient b of equalizer 110 multiply by the signal that comes self-channel that is received.Once more, preprocessor 112 is actually the main signal that modulation receives and the modulation operations of ghost image thereof, makes that described ghost image and the main signal that is received are unequal.Therefore, ghost image no longer is 100% ghost image.By fast Fourier transform 118 multiplication result of preprocessor 112 is transformed into frequency domain, multiplier 120 usefulness complex coefficient A multiply by the frequency domain multiplication result from fast Fourier transform 118, so that eliminate ghost image from the multiplication result of preprocessor 112, and inverse FFT 122 will not have ghost image, frequency domain, modulated reception main signal and be transformed into time domain.Preprocessor 116 usefulness coefficient c multiply by the output from finite filter 114, so that the effect of the modulation that preprocessor 112 is forced is reversed as mentioned above, and the output of inverse FFT 122 are applied window function.

Controller 124 is measured d at interval, provides coefficient b, A and c to preprocessor 112, multiplier 120 and preprocessor 116 respectively, and preprocessor 112, finite filter 114 and preprocessor 116 are synchronized to every blocks of data by described equalizer 110.

As shown as an example among Fig. 9, the coefficient b that is applied by preprocessor 102 and 112 can be a discrete step.Each of these steps has and equals the at interval width along time shaft of d, and described interval d is the main signal that separately received and the time interval of ghost image thereof.Similarly, the ratio of the amplitude of the amplitude of arbitrary step and the last step of adjacency is α, and wherein α is a constant and preferably less than 1.In the example shown in Fig. 9, α is 0.8.And, coefficient b is applied to each data input block as piece, therefore, at the t that begins to locate of the piece of coefficient b 0With t in the end of the piece of coefficient b B+dBetween the difference temporal length that is equivalent to the data input block add d, wherein d separates the received main signal and the time interval of ghost image thereof as mentioned above.For example, as shown in Figure 9, if each data input block has the duration in 256 sampling times, and d has the duration in 32 sampling times, t so 0And t bBetween difference be 288 sampling times.In addition, each side at the piece of coefficient b should have suitable protection at interval.

Notice main signal and ghost image thereof that coefficient b modulation is received, make the amplitude of the ghost image after the applying of coefficient b preferably less than the amplitude of the main signal that is received.Thereby if by by along the interval d of time shaft and the pulse main signal that receives 130 shown in Figure 16 and the ghost image 132 thereof that separate, then after the applying of coefficient b, described signal 130 and ghost image 132 thereof can have surface shown in Figure 17.Be also noted that described coefficient b carries out window function having eliminated on the meaning that interval outside the data input block adds the energy that is received among the d at interval.

In Figure 10, show the coefficient a that applies by finite filter 104 as an example.As can be as seen from Figure 10, by under the situation of FIR filter, applying coefficient a.Each of these coefficients is contiguous to all being separated by interval d.Similarly, the ratio of the amplitude of the last coefficient of the amplitude of arbitrary coefficient and adjacency is α.Because α is less than 1, the amplitude of described coefficient reduces towards 0 progressively.Coefficient a preferably occupies the time interval of the length that doubles the data input block.For example, if the data input block has the duration in 256 sampling times, coefficient a preferably has the duration in 512 sampling times so.Apply the result of coefficient a as finite filter 104, from the output of preprocessor 102, eliminated ghost image.

In Figure 10 and 11, show the coefficient A that applies by multiplier 120 as an example.Because the output of fast Fourier transform 122 is multiple, coefficient A also must be multiple.Therefore, coefficient A has real part shown in Figure 11 and the described imaginary part of Figure 12.As finding out from Figure 11 and 12, coefficient A is based on and postpones d and ratio cc.Similarly, the duration of each real part of coefficient A and the imaginary part twice of the length of the duration of data input block preferably.Apply the result of coefficient A as multiplier 120, from the output of preprocessor 112, eliminated ghost image.

Showing the coefficient c that is applied by preprocessor 106 and 116 among Figure 13 as an example can be discrete step.Each of these steps has the width d along time shaft.Similarly, under the situation of coefficient c, the ratio cc of the amplitude of the step subsequently of the amplitude of arbitrary step and adjacency is preferably less than 1.In example shown in Figure 13, α is 0.8.Therefore and, coefficient c is applied to the output of finite filter 104 and the output of inverse FFT 122 by piece, and at the section start t of coefficient c piece 0The t of end with coefficient c piece cBetween the difference and the time span of data input block suitable.Do not require described t 0And t cBetween difference comprise d because eliminated ghost image already, d separates the received main signal and the time span of ghost image thereof as mentioned above.For example, if the data input block has the duration in 256 sampling times, t so 0And t cBetween difference also be 256 sampling times.In addition, on each side of coefficient c piece, should there be suitable protection at interval.Coefficient c reverses the modulation on the main signal of applying of the coefficient b reception of forcing at.Coefficient c also provides window function, makes data IOB in output place of finite filter 104 and 114 have in fact the duration that the duration with the data input block is complementary.Therefore, in order to eliminate 100% ghost image, the quantity of the pulse in finite filter 104 and 114 the response needs not be unlimited, and can be the quantity that gears to actual circumstances.

The above-mentioned priori that generally requires d about Fig. 9 and 13 described coefficient b and c.Do not require the priori of d below about Figure 14 and 15 described coefficient b and c.Curve as the curve of the coefficient b that illustrates as an example among Figure 14 is, promptly the ratio along the amplitude of x1 place, any point curve of time shaft and an amplitude of the curve at x2 place is a constant alpha, wherein separate x1 and x2 with d, wherein d can have any value, and x2 along time shaft than the first appearance of x1.Constant alpha is preferably less than 1.In example shown in Figure 14, α is 0.8.And, as before, coefficient b is applied to the data input block by piece, therefore, at the section start t of curve 0The t of end with curve B+dBetween difference and the time span of data input block add that d is suitable, wherein d separates the received main signal and the time span of ghost image thereof as mentioned above.In addition, on each side of coefficient b piece, should there be suitable protection at interval.

Following equation has provided the curve of coefficient b as shown in figure 14:

b = k 0 α - x k 1 . . . ( 1 )

Wherein x is along time shaft t 0And t B+dBetween poor, α as mentioned above, k 0Be that such constant makes at a t 0The b of place has desirable value, and k 1Relevant with d.

Curve as the curve of the coefficient c that illustrates as an example among Figure 15 is, promptly the ratio along the amplitude of x1 place, any point curve of time shaft and an amplitude of the curve at x2 place is α, wherein separate x1 and x2 with d, wherein d can have any value, and x2 occurs after than x1 along time shaft.As shown in Figure 15, α is 0.8.Therefore as before, coefficient c is applied to the output of finite filter 104 and the output of inverse FFT 122 by piece, and at the section start t of coefficient c piece 0The t of end with coefficient c piece cBetween the difference and the time span of data input block suitable.Do not require described t 0And t cBetween difference comprise d because eliminated ghost image already.In addition, on each side of coefficient c piece, should there be suitable protection at interval.Coefficient c reverses the modulation on the main signal of applying of the coefficient b reception of forcing at.Similarly, as described above, coefficient c provide window function, makes data IOB in output place of finite filter 104 and 114 have in fact the duration that the duration with the data input block is complementary.

Following equation has provided the curve of coefficient c as shown in figure 15:

c = k 0 α x k 1 . . . ( 2 )

Wherein x is along time shaft t 0And t cBetween poor, α as mentioned above, k 0Be that such constant makes at a t 0The c of place has desirable value, and k 1Relevant with d.

Notice that the quantity of the calculating of being undertaken by conversion shown in Figure 6 is by n 2Increase with n increases, and wherein n is the quantity of the data element in the data block.Notice that further the number of computations of being undertaken by the acoustic convolver such as the acoustic convolver 108 of Fig. 7 is also by n 2Increase with n increases.Yet the quantity of the calculating of being undertaken by the finite filter 114 of Fig. 8 increases with the increase of n by nlogn.Thereby the calculating of being undertaken by equalizer 110 is less than the calculating of being undertaken by the conversion of Fig. 6 significantly.

About Fig. 7 and 8 two embodiment according to equalizer of the present invention have been described above.Yet, be possible according to other embodiment of equalizer of the present invention.For example as shown in figure 18, preprocessor 150 usefulness coefficient b multiply by received main signal and ghost image thereof.To frequency domain, and shown in Figure 11 and 12, multiplier 154 usefulness complex coefficient A multiply by the frequency domain output of fast Fourier transform 152 the output transform of preprocessor 150 in fast Fourier transform 152.The output of acoustic convolver 156 usefulness coefficient c convolution multipliers 154 is so that the data that recovery sends by channel.In this case, coefficient c must be multiple.Same, the inverse FFT as the inverse transformation of fast Fourier transform 152 is arranged in the transmitter, with wherein conversion of signals to time domain to transmit by channel.

Figure 19 has illustrated the equalizer 160 that comprises finite filter 162 and preprocessor 164.Finite filter 162 comprises fast Fourier transform 166, multiplier 168, inverse FFT 170.Fast Fourier transform 166 is arrived frequency domain to the conversion of signals of coming self-channel that is received, multiplier 168 usefulness complex coefficient A multiply by the frequency-region signal from fast Fourier transform 166, so that eliminate ghost image from the signal that is received, and 170 conversion of signals no ghost image, frequency domain of inverse FFT are to time domain.Preprocessor 172 usefulness coefficient c multiply by the output from finite filter 162, so that the output to inverse FFT 170 applies window function, make each data IOB in output place of finite filter 162 have the suitable duration of duration of corresponding with it the in fact data input block of handling by equalizer 160.

Controller 172 is measured d at interval, so that determine coefficient A, coefficient A and c is applied to multiplier 168 and preprocessor 164 respectively, and finite filter 162 and preprocessor 164 are synchronized to every blocks of data by described equalizer 160.

Show the coefficient A that applies by multiplier 168 among Figure 20 and 21 as an example.Because the output of fast Fourier transform 166 is multiple, coefficient A also must be multiple.Therefore, coefficient A has real part shown in Figure 20 and imaginary part shown in Figure 21.As finding out from Figure 20 and 21, coefficient A is based on d and ratio cc at interval.Similarly, the duration of each real part of coefficient A and the imaginary part twice of the length of the duration of data input block preferably.Apply the result of coefficient A as multiplier 168, from the output of preprocessor 102, eliminated ghost image.

Show the coefficient c that applies by preprocessor 164 among Figure 22 as an example.Owing to do not modulate the received main signal and the preprocessor of ghost image thereof in the equalizer 160, do not require that coefficient c removes the effect of any modulation.Therefore, from t 0To t cWindow in, coefficient can have non-0 constant value.By piece the coefficient c shown in Figure 22 is applied to the output of finite filter 162, and therefore at the t of the section start of coefficient c 0T with the end of coefficient c cBetween difference and duration of each data input block suitable.The same with the front is if the data input block has the duration in 256 sampling times, t so 0And t cBetween difference also be 256 sampling times.In addition, on each side of coefficient c piece, should there be suitable protection at interval.Coefficient c provide window function, makes the data IOB of output place of finite filter 162 is restricted in fact the duration that is complementary with its duration of corresponding data input block.Therefore, in order to eliminate 100% ghost image, the quantity of the pulse in the response of finite filter 162 needs not be unlimited, and can be the quantity that gears to actual circumstances.

Some correction of the present invention described above and selecting fully.For the personnel that are practiced in the field of the invention, other correction will occur and select fully.For example, because the present invention can operate the most satisfactorily, comprise ghost image and/or other linear distortion about term ghost image used in the present invention here in the situation that has ghost image and other linear distortion.

And, show coefficient b by non-complex coefficient above.Yet coefficient b can be multiple, such as at the signal that is received be in the situation of QAM signal like that.

Therefore, it only is illustrative that description of the invention is interpreted as, and is to implement optimal mode of the present invention for those of ordinary skill in the art is lectured.Can change details in fact, and not deviate from main idea of the present invention, and keep the private right of all corrections within the scope of accessory claim.

Claims (18)

1. the method for an equalization channel is characterized in that comprising:
Reception has the signal of master signal component and ghost image component;
Multiply by the signal that is received with coefficient, to produce treated signal, wherein said master signal component and described ghost image component have unequal amplitude; And
Utilize finite filter from described treated signal, to remove described ghost image component in fact, do not have described ghost image component in fact so that export described master signal component.
2. the method for claim 1 is characterized in that removing in fact described ghost image component and comprises with the time domain finite filter described treated signal is carried out filtering from described treated signal.
3. method as claimed in claim 2 is characterized in that described time domain filtering comprises the FIR filter.
4. the method for claim 1 is characterized in that removing in fact described ghost image component and comprises with the frequency domain finite filter described treated signal is carried out filtering from described treated signal.
5. method as claimed in claim 4, it is characterized in that described frequency domain finite filter comprise be arranged to the main signal that is received and described ghost image be transformed into frequency domain fast Fourier transform, be arranged to multiply by the main signal that received and described ghost image so that the inverse FFT of eliminating the multiplier of ghost image in fact and being arranged to the output of described multiplier is changed back time domain with coefficient A.
6. the method for claim 1, it is characterized in that describedly multiply by the signal that is received with coefficient and comprising with coefficient b and multiply by described master signal component and described ghost image component, wherein said coefficient b comprises the step of different amplitudes, each of wherein said step has and equals the main signal and duration in the time interval between the described ghost image of being received in fact, and the amplitude of one of described step is not equal to 1 with ratio in abutting connection with the amplitude of step.
7. the method for claim 1 is characterized in that describedly multiply by the signal that is received with coefficient and comprising with coefficient b and multiply by described master signal component and described ghost image component that wherein said coefficient b constitutes exponential curve.
8. the method for claim 1, it is characterized in that describedly multiply by the signal that is received with coefficient and comprising with first coefficient and multiply by the signal that is received, and wherein said method comprises that further the master signal component that multiply by output with second coefficient will be putting on effect counter-rotating on the described master signal component owing to multiply by the signal that is received with first coefficient.
9. method as claimed in claim 8, it is characterized in that the master signal component that multiply by described output with second coefficient comprises the master signal component that multiply by described output with coefficient c, wherein said coefficient c comprises the step with different amplitudes, wherein each step has the duration that equals the duration of data block on the time in fact, and the amplitude of one of described step is not equal to 1 with ratio in abutting connection with the amplitude of step.
10. method as claimed in claim 9, it is characterized in that described usefulness first coefficient multiply by the signal that is received and comprises with coefficient b and multiply by described master signal component and described ghost image component, wherein said coefficient b comprises the step of different amplitudes, each of wherein said step has and equals the main signal and duration in the time interval between the described ghost image of being received in fact, the amplitude of one of described step is not equal to 1 with ratio in abutting connection with the amplitude of step, the step of one of wherein said coefficient b and c is that amplitude reduces gradually, increases gradually and the step of other coefficient b and c is an amplitude.
11. method as claimed in claim 8 is characterized in that the master signal component that described usefulness second coefficient multiply by described output comprises the master signal component that multiply by described output with coefficient c, wherein said coefficient c constitutes exponential curve.
12. method as claimed in claim 11, it is characterized in that described usefulness first coefficient multiply by the signal that is received and comprises with coefficient b and multiply by described master signal component and described ghost image component, wherein said coefficient b constitutes exponential curve, the exponential curve amplitude of one of wherein said coefficient b and c reduces gradually, and the amplitude of the exponential curve of other coefficient b and c increases gradually.
13. the method for claim 1 is characterized in that further comprising that the master signal component to described output applies window function.
14. method as claimed in claim 13 is characterized in that the described window function that applies comprises the master signal component that multiply by described output with coefficient.
15. method as claimed in claim 14, it is characterized in that the described master signal component that multiply by described output with coefficient comprises the master signal component that multiply by described output with coefficient c, wherein said coefficient c comprises the step with different amplitudes, wherein each step has the duration of the duration that equals data block in fact, and the amplitude of one of described step is not equal to 1 with ratio in abutting connection with the amplitude of step.
16. method as claimed in claim 15, it is characterized in that describedly multiply by the signal that is received with coefficient and comprising with coefficient b and multiply by described master signal component and described ghost image component, wherein said coefficient b comprises the step of different amplitudes, each of wherein said step has and equals the main signal and duration in the time interval between the described ghost image of being received in fact, the amplitude of one of described step is not equal to 1 with ratio in abutting connection with the amplitude of step, the step of one of wherein said coefficient b and c is that amplitude reduces gradually, increases gradually and the step of other coefficient b and c is an amplitude.
17. method as claimed in claim 14 is characterized in that the described master signal component that multiply by described output with coefficient comprises the master signal component that multiply by described output with coefficient c, wherein said coefficient c constitutes exponential curve.
18. method as claimed in claim 17, it is characterized in that describedly multiply by the signal that is received with coefficient and comprising with coefficient b and multiply by described master signal component and described ghost image component, wherein said coefficient b constitutes exponential curve, the exponential curve amplitude of one of wherein said coefficient b and c reduces gradually, and the amplitude of the exponential curve of other coefficient b and c increases gradually.
CNB008195420A 2000-05-17 2000-05-17 Ghost eliminating equalizer CN1193501C (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/US2000/013620 WO2001089087A1 (en) 2000-05-17 2000-05-17 Ghost eliminating equalizer

Publications (2)

Publication Number Publication Date
CN1452809A CN1452809A (en) 2003-10-29
CN1193501C true CN1193501C (en) 2005-03-16

Family

ID=21741392

Family Applications (1)

Application Number Title Priority Date Filing Date
CNB008195420A CN1193501C (en) 2000-05-17 2000-05-17 Ghost eliminating equalizer

Country Status (7)

Country Link
CN (1) CN1193501C (en)
AU (1) AU5140600A (en)
BR (1) BR0015868A (en)
CA (1) CA2409518A1 (en)
HK (1) HK1059154A1 (en)
MX (1) MXPA02011140A (en)
WO (1) WO2001089087A1 (en)

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5331416A (en) * 1992-12-02 1994-07-19 Samsung Electronics Co., Ltd. Methods for operating ghost-cancelation circuitry for TV receiver or video recorder
US5285280A (en) * 1993-03-19 1994-02-08 Industrial Technology Research Institute Division method and system for ghost cancellation
IL127134A (en) * 1996-03-04 2002-11-10 Oren Semiconductor Ltd Dsp apparatus

Also Published As

Publication number Publication date
AU5140600A (en) 2001-11-26
MXPA02011140A (en) 2003-04-25
CN1452809A (en) 2003-10-29
BR0015868A (en) 2003-08-19
WO2001089087A1 (en) 2001-11-22
CA2409518A1 (en) 2001-11-22
HK1059154A1 (en) 2005-08-19

Similar Documents

Publication Publication Date Title
US20170250772A1 (en) Orthogonal frequency division multiplexing-code division multiple access system
US9083595B2 (en) Signal modulation method resistant to echo reflections and frequency offsets
JP6001760B2 (en) Filter bank multicarrier signal transmission and channel estimation method and apparatus
US8542772B2 (en) Multi-user detection using equalization and successive interference cancellation
US20140247898A1 (en) Ofdm communications methods and apparatus
US7206349B2 (en) Multicarrier receiver with channel estimator
EP1488589B1 (en) Apparatus and method for estimating a plurality of channels
KR100925909B1 (en) Staggered pilot transmission for channel estimation and time tracking
CN1233137C (en) Apparatus and method for keeping frame synchronization in multi-carrier transmission system
US7042929B2 (en) Scaling using gain factors for use in data detection for wireless code division multiple access communication systems
KR100679554B1 (en) Method and apparatus for error reduction in an orthogonal modulation system
CN1078410C (en) Diversity receiver for signals with multipath time dispersion
JP3366794B2 (en) Spread spectrum code pulse position modulation receiver with delay spread compensation
US7848438B2 (en) Method and apparatus for pilot signal transmission
US7002904B2 (en) Method and apparatus for reducing peak power in partial transmit sequence OFDM
KR100835147B1 (en) Time filtering for excess delay mitigation in ofdm systems
CA2226723C (en) Pulse shaping for multicarrier modulation
CN1148922C (en) Optimization of cannel equalizer
EP1382146B1 (en) Data communication apparatus and method based on orthogonal frequency division multiple access
EP2162985B1 (en) Method and apparatus for communicating with root-nyquist, self-transform pulse shapes
EP1181797B1 (en) A windowing function for maintaining orthogonality of channels in the reception of ofdm symbols
CN102088436B (en) Steering diversity for the multiple antenna communication based on OFDM
US7634023B2 (en) Orthogonal frequency digital multiplexing correlation canceller
CN1211944C (en) Radio frequency noise canceller
JP2015520954A (en) Signal modulation method resistant to echo reflection and frequency offset

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
REG Reference to a national code

Ref country code: HK

Ref legal event code: DE

Ref document number: 1059154

Country of ref document: HK

C14 Grant of patent or utility model
GR01 Patent grant
C17 Cessation of patent right
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20050316