[go: up one dir, main page]

CN100562076C - Time Domain Adaptive Equalizer and Its Contained Decision Feedback Filter - Google Patents

Time Domain Adaptive Equalizer and Its Contained Decision Feedback Filter Download PDF

Info

Publication number
CN100562076C
CN100562076C CNB2006100283491A CN200610028349A CN100562076C CN 100562076 C CN100562076 C CN 100562076C CN B2006100283491 A CNB2006100283491 A CN B2006100283491A CN 200610028349 A CN200610028349 A CN 200610028349A CN 100562076 C CN100562076 C CN 100562076C
Authority
CN
China
Prior art keywords
feedback filter
output
filter
input
delayer
Prior art date
Legal status (The legal status 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 status listed.)
Expired - Fee Related
Application number
CNB2006100283491A
Other languages
Chinese (zh)
Other versions
CN101098416A (en
Inventor
居峰
何大治
张文军
管云峰
吴松炎
归琳
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
SHANGHAI HIGH DEFINITION DIGITAL TECHNOLOGY INDUSTRIAL Co Ltd
Original Assignee
SHANGHAI HIGH DEFINITION DIGITAL TECHNOLOGY INDUSTRIAL Co Ltd
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 SHANGHAI HIGH DEFINITION DIGITAL TECHNOLOGY INDUSTRIAL Co Ltd filed Critical SHANGHAI HIGH DEFINITION DIGITAL TECHNOLOGY INDUSTRIAL Co Ltd
Priority to CNB2006100283491A priority Critical patent/CN100562076C/en
Publication of CN101098416A publication Critical patent/CN101098416A/en
Priority to HK08107321.7A priority patent/HK1116970B/en
Application granted granted Critical
Publication of CN100562076C publication Critical patent/CN100562076C/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)

Abstract

本发明涉及一种时域自适应均衡器,其中输入信号流入前向滤波器,前向滤波器的输出流入加法器的一个输入端,加法器的另两个输入为第一反馈滤波器的输出和第二反馈滤波器的输出。加法器对输入信号相加,其输出连至判决器的输入端。误差生成器根据判决器的输入和输出生成误差信号。抽头系数更新器根据误差信号生成抽头系数,来更新所述前向滤波器和所述判决反馈滤波器的抽头系数。第一反馈滤波器使用转置结构,第二反馈滤波器使用非转置结构并且其输入相对于第一反馈滤波器的输入有一定延时。转置结构和非转置结构的级数是根据所需要对付的最长后径的长度决定的。

The invention relates to a time-domain adaptive equalizer, wherein an input signal flows into a forward filter, an output of the forward filter flows into one input of an adder, and the other two inputs of the adder are outputs of a first feedback filter and the output of the second feedback filter. The adder sums the input signals, and its output is connected to the input of the decision unit. The error generator generates an error signal based on the input and output of the decider. The tap coefficient updater generates tap coefficients according to the error signal to update the tap coefficients of the forward filter and the decision feedback filter. The first feedback filter uses a transpose structure, the second feedback filter uses a non-transpose structure and its input has a certain delay relative to the input of the first feedback filter. The series of transposed structures and non-transposed structures is determined according to the length of the longest rear diameter to be dealt with.

Description

Time-domain adaptive equalizer and the decision feedback filter device that comprises thereof
Technical field
The present invention relates to a kind of time-domain adaptive equalizer, relate in particular to a kind of time-domain adaptive equalizer based on the time-delay coupling.
Background technology
In the practical application of many different digital informations such as sound, data and video communication etc., equalizer is a kind of very important element.Equalizer is used as the Echo Canceller (compensator), Digital Television of speakerphone in the full-duplex communication or video anti-ghost device, radio modem and the wireless telephonic signal conditioner etc. in the digital cable transmission.In signals transmission, because the existence of multipath signal can bring intersymbol interference (ISI) in the channel, and intersymbol interference is a major reason that produces error, in most of single carrier digital application, generally all use time-domain adaptive equalizer to revise the ISI mistake.
Fig. 1 is the modular structure block diagram sketch of traditional time-domain adaptive equalizer 100.Usually in digital receiver, time-domain adaptive equalizer 100 comprises 20, one error makers 50 of 40, one feedback filters of 10, one decision devices of 60, one forward-direction filters of a tap coefficient renovator and an adder 30.Forward-direction filter 10 can be realized with finite impulse response (FIR) (FIR) filter.Forward-direction filter 10 is used for receiving inputted signal, eliminates the forward direction multipath, the multipath signal that promptly main relatively transmission signals arrives in advance.The output sum1 of forward-direction filter 10 is connected to an adder 30.Another input of adder 30 receives the output sum2 from feedback filter 20.Adder 30 is with input signal sum1 and sum2 one of adduction output and signal sum_EQ mutually, as the output signal of whole time-domain adaptive equalizer 100.Simultaneously, equalizer output signal sum_EQ flows into decision device 40 again, and decision device 40 is made judgement according to its input information sum_EQ and the relation of constellation point, and the output valve of decision device is from the immediate constellation point numerical value of sum_EQ.Simultaneously, if the output of decision device 40 directly as an input of feedback filter 20, so such feedback filter 20 just belongs to the decision feedback filter device.The output of feedback filter 20 adds and is worth the output signal sum_EQ that sum2 will revise equalizer, to eliminate the back to multipath, promptly is later than the multipath signal that main transmission signals arrives, and can eliminate the residual signal that forward-direction filter 10 produces.Error signal is generated according to the output of decision device 40 and the difference of input signal by error maker 50.Tap coefficient renovator 60 is used for receiving error signal, and generates the filter coefficient of forward-direction filter 10 and feedback filter 20.There are a lot of feasible methods to be applicable to the generation filter coefficient, comprise lowest mean square (LMS) and recursive least square (RLS) algorithm.The form of filter also has multiple, can be real filter, that is, the number average of tap coefficient and register is a real number, and the multiplying of filter also is real arithmetic; Also can be complex filter, that is, the number average of tap coefficient and register be a plural number, and the multiplying of filter is a complex operation; Can also be accurate complex filter, that is, though counting of tap coefficient and register is plural number, the multiplying of filter be real arithmetic.
Fig. 2 shows and is traditional concrete block diagram of time-domain adaptive equalizer modular structure.Compared to Figure 1, Fig. 2 provides the concrete structure of forward-direction filter 10 and feedback filter 20.Specifically, forward-direction filter 10 and feedback filter 20 all adopt the non-transpose configuration shown in Fig. 3 a to realize.Referring to Fig. 2 and Fig. 3 a, be example with the non-transpose configuration of feedback filter 20, wherein comprise the M level delayer of cascade, (the output of decision device of the input signal of feedback filter
Figure C20061002834900071
) through 1 grade, 2 grades ... M level delayer is delayed time respectively, obtain a corresponding M delayer output, be respectively the output of the 1st grade of delayer, the output of the 2nd grade of delayer ..., M level delayer output, the corresponding multiplier of M is exported M delayer respectively and corresponding filter coefficient d 1, d K, d MMultiplying each other obtains a corresponding M multiplier output, and M multiplier output obtains the output of feedback filter 20 at last by an adder addition.Forward-direction filter 10 also has similar non-transpose configuration.
In theory, need only the progression long enough of feedback filter in the equalizer, just can tackle the multipath channel of endless.But actually, the realization of equalizer is subjected to the restriction of 2 factors: hardware resource expense and sequential relationship.For the former, because the development of FPGA technology and asic technology, and consider that the length of multipath signal is always limited, the hardware resource expense no longer has been the bottleneck problem of restriction equalizer realization; For the latter, the sequential anxiety mainly is subject to the arithmetic speed of feedback filter.This is because forward-direction filter (FIR) always can solve with streamline (Pipeline) technology group delay and insensitive.And feedback filter need be finished in a symbol period usually and allly in the filter takes advantage of, adds computing.Especially, when adopting Blind Equalization Technique, the data that enter feedback filter no longer are to declare data firmly, but equalizer adds and soft value, feedback filter will become infinite impulse response (IIR) filter this moment, multiplier in the feedback filter can't replace with XOR gate, and sequential relationship will be nervous more.For this reason, the feedback filter that proposes in the time-domain adaptive equalizer of someone adopts transpose configuration to realize.
Fig. 3 b illustrates the exemplary plot of transpose configuration.Transpose configuration shown in Fig. 3 b comprises M multiplier, a M-1 delayer and M-1 adder, and input signal is earlier without time-delay, but via M multiplier respectively with a corresponding M tap coefficient d M, d M-1..., d 1Multiply each other, obtain M output product value.Input signal and coefficient d MMultiply each other first product value input first order delayer of gained, resulting time-delay output and input signal and coefficient d M-1The multiply each other second product value addition of gained, it adds and is worth and imports second level delayer, resulting time-delay output and input signal and coefficient d M-2Multiply each other the 3rd product value addition of gained, the rest may be inferred, the delayer output of M-1 level and input signal and coefficient d 1The multiply each other M product value addition of gained finally obtains the output of the filter of transpose configuration.The input of the time-delay earlier data of transpose configuration and traditional non-transpose configuration all are consistent with multiply each other the respectively mode output signal of addition then of N level coefficient with filter effect again.Yet the basic structure of add tree is M=2 nIndividual input is carried out obtaining M/2 result behind the first order addition, and partial input is M/2 output of the first order, obtains M/4 output valve after the second level adds, and each grade addition needs the time-delay of a symbol period, and is general, if 2 n<M<=2 N+1Be exactly the progression n+1 level that needs so altogether is always if n+1 the symbol period of delaying time is special M=2 n, the progression that needs so altogether is exactly log 2The M level, log always delays time 2M symbol period, traditional non-transpose configuration will be through log between from the input signal to the output signal 2So the add tree of M level is must be through the log of add tree 2The time-delay of M symbol period just can have output; The shortest between the input and output signal of transpose configuration have only an adder, so time-delay has only the time-delay of one-level addition, just a symbol period.Therefore, because traditional non-transpose configuration generally can use add tree that M product value summed up, can introduce log like this 2The addition time-delay of M symbol period, and the benefit of transpose configuration is reduced to this time-delay 1 symbol period exactly, and do not have any influence for the pipeline processes of handling.If the time-delay of feedback filter is greater than 1 symbol period, be assumed to be d greater than 1, back in the individual symbol period in main back (d-1), footpath not can do with to this feedback filter of multipath so, if traditional non-transpose configuration feedback filter wants can do with all back footpaths, must be all M product value addition output in a symbol period, this just brings corresponding sequence problem, and transpose configuration then can be avoided this sequence problem.But the drawback that realizes with transpose configuration is: because the fixed point above the hardware is handled, whenever do a sub-addition, in order to guarantee not overflow, the bit bit wide of output is imported data relatively will increase by one, along with add and carrying out, the bit accuracy that needs will be more and more higher.The bit bit wide of exporting transpose configuration such as the addition of N level will increase the N-1 position, and traditional non-transpose configuration only can increase log 2N bit bit wide.Than non-transpose configuration, the consumption of the hardware resource of transpose configuration will increase like this.If the progression of feedback filter is a lot, in order to save hardware resource, need the number of bits of restriction input signal, so not only limited the realization of blind equalization, also be a big obstacle for the modulation system that realizes many constellation point.
In sum, the hardware resource expense of non-transpose configuration is few, based on the development of FPGA technology and asic technology, no longer is problem, but the sequential relationship anxiety.And the output result of transpose configuration and non-transpose configuration is identical, alleviated sequential relationship, but the hardware resource expense is very big.
Therefore traditional time-domain adaptive equalizer can't resist the long multipath in the Channel Transmission, and especially long and strong is back to multipath.
Summary of the invention
The present invention is directed to digital terrestrial television transmission system the deficiencies in the prior art and defective, the structure of conventional equalizer is improved, proposed a kind of time-domain adaptive equalizer based on the time-delay coupling.By the delay line of design of feedback filter meticulously, make the feedback filter of longer progression also can realize with traditional transverse filter structure, so not only saved hardware resource, can also effectively improve the precision of equalizer, effectively offset long multipath, and make blind equalization become possibility.
In one embodiment of the invention, provide a kind of decision feedback filter device, having comprised: first feedback filter, it has transpose configuration; Second feedback filter, its input signal has the first predetermined time-delay and has non-transpose configuration with respect to the input signal of first feedback filter.
In another embodiment of the present invention, the decision feedback filter device also comprises a plurality of delayers, described a plurality of delayer is cascade, the previous stage delayer is output as the input of next stage delayer, the input of first delayer is the input signal of described first feedback filter in wherein a plurality of delayers, and the delayer output that the input signal of relative first feedback filter has the described first predetermined time-delay is the input of described second feedback filter.
In another embodiment of the present invention, a kind of time-domain adaptive equalizer is provided, comprising: forward-direction filter, its receiving inputted signal is used to eliminate the forward direction multipath; The decision feedback filter device is used for eliminating the back to multipath; Adder, it receives output from forward-direction filter and decision feedback filter device as input, carries out sum operation and exports an equalizer output; Decision device receives described equalizer output as input signal, and output is offered described decision-feedback wave filter; The error signal maker generates error signal according to the input signal of described decision device and the output of described decision device; The tap coefficient renovator generates tap coefficient to upgrade the tap coefficient of forward-direction filter and decision feedback filter device according to described error signal, and described decision feedback filter device comprises: first feedback filter, and it has transpose configuration; Second feedback filter, its input signal has the first predetermined time-delay and has non-transpose configuration with respect to the input signal of first feedback filter.
In yet another embodiment of the present invention, time-domain adaptive equalizer also comprises a plurality of delayers, described a plurality of delayer is cascade, the previous stage delayer is output as the input of next stage delayer, the input of first delayer is the input signal of described first feedback filter in wherein a plurality of delayers, and the delayer output that the input signal of relative first feedback filter has the described first predetermined time-delay is the input of described second feedback filter.
Description of drawings
Fig. 1 is traditional time-domain adaptive equalizer modular structure block diagram sketch.
Fig. 2 is traditional concrete block diagram of time-domain adaptive equalizer modular structure.
Fig. 3 a is the exemplary plot of non-transpose configuration.
Fig. 3 b is the exemplary plot of transpose configuration.
Fig. 4 is the time-domain adaptive equalizer structured flowchart sketch based on the time-delay coupling.
Fig. 5 is the concrete block diagram of time-domain adaptive equalizer structure based on the time-delay coupling.
Embodiment
Provide following examples in conjunction with content of the present invention, be applied in the digital tv ground broadcasting.
Fig. 4 is the structured flowchart sketch based on the time-domain adaptive equalizer 200 of time-delay coupling.In one embodiment of the invention, the time-domain adaptive equalizer 200 based on the time-delay coupling comprises: forward-direction filter 10, first feedback filter 21, second feedback filter 22, decision device 40, error maker 50, tap coefficient renovator 60 and adder 30.Interconnected relationship between its each parts is: input signal at first enters forward-direction filter 10, and the output sum1 of forward-direction filter 10 is connected to an adder 30.The output that in addition two input sum2 of adder 30 and sum3 are respectively first feedback filter 21 and second feedback filter 22.Adder 30 is with input signal sum1, sum2 and sum3 addition, and output sum_EQ signal, as the output signal of whole time-domain adaptive equalizer 200.Equalizer output sum_EQ flows into decision device 40 again, and decision device 40 is made judgement according to its input information sum_EQ and the relation of constellation point, and the output valve of decision device is from the immediate constellation point numerical value of sum_EQ.Simultaneously, if the output of decision device 40 is directly as an input of first feedback filter 21, the output that the output of first feedback filter 21 adds and be worth the sum2 and second feedback filter 22 adds and is worth the output signal sum_EQ that sum3 will revise equalizer, to eliminate the back to multipath, promptly be later than the multipath signal that main transmission signals arrives, and can eliminate the residual signal that forward-direction filter 10 produces.Error signal is generated according to the output of decision device 40 and the difference of input signal by error maker 50.Tap coefficient renovator 60 is used for receiving error signal, and generates the filter coefficient of forward-direction filter 10, first feedback filter 21 and first feedback filter 22.Wherein, first feedback filter 21 is the B level altogether, second feedback filter 22 is the M level altogether, M=2^ (B-A) wherein, A is a delayer progression, the input of representing second feedback filter 22 is with respect to the input of first feedback filter 21 time-delay through A level delayer, A be more than or equal to zero, smaller or equal to the integer of B.
Referring to Fig. 5, Fig. 5 is the concrete block diagram of structure based on the time-domain adaptive equalizer 100 of time-delay coupling.Wherein, first feedback filter 21 is to be realized by transpose configuration shown in Figure 3, and second feedback filter 22 is to be realized by non-transpose configuration shown in Figure 3.The output SUM_EQ of equalizer can obtain decision value through a decision device 40
Figure C20061002834900111
As the input of first feedback filter 21, and the output of the A level delayer in a plurality of delayers in first feedback filter 21
Figure C20061002834900112
As the input of second feedback filter 22, SUM_EQ and
Figure C20061002834900113
Difference by error maker 50 generated error signals, error signal input tap coefficient renovator 60 is to upgrade the tap of forward-direction filter 10
Figure C20061002834900114
The tap of first feedback filter 21
Figure C20061002834900115
And the tap of second feedback filter
Figure C20061002834900116
The generation of tap coefficient and update mode all are known in those skilled in the art.
Below a plurality of delayers in first feedback filter 21 are illustrated.Described a plurality of delayer is cascade, the previous stage delayer is output as the input of next stage delayer, the input of first delayer is the input signal of described first feedback filter in wherein a plurality of delayers, the output of A level delayer then is the input of described second feedback filter, and the input that it makes second feedback filter has certain time-delay with respect to the input of first feedback filter.
In the above description, the amount of delay of each grade all equates with unit amount of delay in forward-direction filter, transpose configuration and the non-transpose configuration.
Wherein, forward-direction filter 10 is that a length is 400 grades, the complex filter of mark space; First feedback filter 21 is the complex filters that length is 32 grades mark space, the employing transpose configuration is realized, like this can one finish in the symbol beat and add and export, and because progression is shorter, it is too big that the bit width that adds and export is unlikely to, the output that is input as decision device 40 of first feedback filter 21; Second feedback filter 22, totally 2 (32-23)=512 grades.The progression of second feedback filter is index not necessarily, in principle as long as can realize by hardware less than this exponential quantity, but needs only the hardware resource permission, then can select usually to represent near this maximum.Because progression is a lot, adopt transpose configuration to realize relatively difficulty, be difficult to control the bit bit wide of output, be unfavorable for the control hardware resource, so adopt the mode of the time-delay coupling of traditional non-transpose configuration to realize, the tap output of the 23rd grade of delayer that is input as first feedback filter 21 of second feedback filter 22, wherein, the amount of delay of each grade delayer equates.So just allow the structure realization that is used for streamline with the time-delay that (32-23)=9 clock cycle can be arranged that adds of second feedback filter 22, the feedback filter that such structure can satisfy all progression in a symbol beat, finish add and the sequential requirement, again can better controlled the bit bit wide of output, it is 32+2 that whole structure is equivalent to a total progression (32-23)The decision feedback filter device, can guarantee in a symbol period, to finish and add and export, and final equalizer output bit bit wide can be remained in the zone of reasonableness.
This feedback filter arrangement based on the time-delay coupling can be supported the input signal of the output of any one position the position of B+1 altogether after the afterbody time-delay before the first order time-delay of first feedback filter 21 as second feedback filter 22.
For example, in another embodiment of the present invention, first feedback filter 21 still is the complex filter of 32 grades mark space for length, the employing transpose configuration is realized, and the just output before its 1st grade of delayer of first feedback filter 21, the input signal that is equivalent to first feedback filter 21 and second feedback filter 22 is identical, corresponding to the A=0 here, because B still equals 32, the progression of second feedback filter just is 2 so (32-0)So just allow the structure realization that is used for streamline with the time-delay that (32-0)=32 clock cycle can be arranged that adds of second feedback filter 22, the feedback filter that such structure can satisfy all progression in a symbol beat, finish add and the sequential requirement, again can better controlled the bit bit wide of output, it is 32+2 that whole structure is equivalent to a total progression (32-0)DFF 200, also can guarantee in a symbol period, to finish and add and export, and the output bit bit wide of final equalizer 200 can be remained in the zone of reasonableness.
Again for example, in another embodiment of the present invention, first feedback filter 21 still is the complex filter of 32 grades mark space for length, the employing transpose configuration is realized, if and first feedback filter 21 is in its afterbody time-delay back output, so just corresponding to A=B here, owing to B still gets 32, A=32 then, this moment, the progression of second feedback filter 22 just was 2 (32-32)=1, whole feedback filter just is equivalent to progression 32+2 (32-32)=33 feedback filter.
Value about B also can change, and still generally speaking the selection of B and A is to determine according to the required the longest back length directly that tackles.By above description as can be known, second feedback filter in the time-domain adaptive equalizer of the present invention is 2 B-ALevel, A is the integer more than or equal to 0, that is, second feedback filter can be at most 2 BLevel (when getting A=0).That is to say that it is B+2 that feedback filter of the present invention can be equivalent to progression at most BFeedback filter.The longest multipath that tackles if desired needs the feedback filter of X level time-delay, as long as can satisfy X<B+2 so B, just can do with the longest multipath.For example, when the longest multipath that tackles when needs needs 512 grades feedback filter, consider and to satisfy 512<B+2 B, B is again an integer, so the value of B is B 〉=9.Therefore, B gets the longest multipath that any value more than or equal to 9 all can do with needs 512 grades of feedback filters, for example in the present invention, and B=32.After the B value has been determined, again according to the satisfied condition X<B+2 of need B-ADetermine the A value.In this example, because X=512, therefore B=32 draws A more than or equal to 0 and smaller or equal to 23 integer.
The progression one of delayer regularly in first feedback filter 21, the input of using first feedback filter 21 is through the delayer of different progression and the output that obtains is that those of ordinary skill in the art can contemplate easily as the operation of the input of second feedback filter 22, the present invention is not limited to the embodiment of the output at the the 0th, the 23rd, the 32nd grade of delayer place of above-mentioned usefulness as 22 inputs of second feedback filter, and the position of other delayer outputs also is feasible.
In that a further aspect of the present invention, feedback filter also is not limited to the first and second above-mentioned feedback filters, but can comprise three or more feedback filters.For example, if at original B=32, on the basis of A=23, the input of using first feedback filter 21 obtains the input of the 3rd feedback filter through the output of C level delayer, carry out filter operations similarly then, wherein A<C<B.The output of the 3rd feedback filter is input summer 30 also.Decision feedback filter device of Xing Chenging and the adaptive equalizer that comprises described decision feedback filter device also can solve technical problem of the present invention like this.
More than disclosed according to the time-domain adaptive equalizer based on the time-delay coupling of the present invention.
Those skilled in the art will appreciate that information and signal can represent with in multiple different technologies and the technology any.For example, data, instruction, order, information, signal, bit, code element and the chip that may relate in the above-mentioned explanation can be represented with voltage, electric current, electromagnetic wave, magnetic field or its particle, light field or its particle or their combination in any.
Those skilled in the art can further understand, and can be used as electronic hardware, computer software or both combinations in conjunction with the described various illustrative logical blocks of embodiment disclosed herein, module and algorithm steps and realizes.In order to clearly demonstrate the interchangeability between hardware and software, as various illustrative assemblies, block diagram, module, circuit and the step 1 according to its functional elaboration.These are functional realizes specific application program and the design of depending on that whole system adopts as hardware or software actually.The technical staff can recognize the interactivity of hardware and software in these cases, and the described function that how to realize each application-specific best.The technical staff may be realizing described function for the different mode of each application-specific, but this realization decision should not be interpreted as causing and deviates from scope of the present invention.
The realization of various illustrative logical block, module and the algorithm steps of describing in conjunction with embodiment as described herein or carry out and to use: general processor, digital signal processor (DSP), application-specific integrated circuit (ASIC) (ASIC), field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components or for carrying out the combination in any that function described here designs.General processor may be a microprocessor, yet or, processor can be processor, controller, microcontroller or the state machine of any routine.Processor also may realize with the combination of computing equipment, as, the combination of DSP and microprocessor, a plurality of microprocessor, in conjunction with one or more microprocessors of DSP kernel or other this configuration arbitrarily.
In the software module that the method for describing in conjunction with disclosed embodiment here or the step of algorithm may directly be included in the hardware, carried out by processor or in the middle of both.Software module may reside in the storage medium of RAM memory, flash memory, ROM memory, eprom memory, eeprom memory, register, hard disk, removable dish, CD-ROM or any other form as known in the art.The coupling of exemplary storage medium and processor makes that processor can be from read information, or information is write storage medium.Perhaps, storage medium can be integrated with processor.Processor and storage medium may reside among the ASIC.ASIC may reside in the user terminal.Perhaps, processor and storage medium may reside in the user terminal as discrete component.
The description of above preferred embodiment makes those skilled in the art can make or use the present invention.The various modifications of these embodiment are conspicuous for a person skilled in the art, and Ding Yi General Principle can be applied among other embodiment and not deviate from the spirit or scope of the present invention here.Therefore, the embodiment that the present invention is not limited to illustrate here, and will meet and the principle and the novel feature the most wide in range consistent scope that disclose here.

Claims (16)

1.一种判决反馈滤波器,其特征在于包括:1. A decision feedback filter, characterized in that it comprises: 第一反馈滤波器,其具有转置结构;a first feedback filter having a transpose structure; 第二反馈滤波器,其输入信号相对于第一反馈滤波器的输入信号具有第一预定延时并且具有非转置结构。A second feedback filter whose input signal has a first predetermined delay relative to the input signal of the first feedback filter and has a non-transposed structure. 2.如权利要求1所述的判决反馈滤波器,其特征在于,2. decision feedback filter as claimed in claim 1, is characterized in that, 所述转置结构包括:The transpose structure includes: 多个乘法器,所述多个乘法器中的每个乘法器的一个输入端接收第一反馈滤波器输入信号,另一个输入端接收相应的滤波系数,并对它们相乘以得到一相应的乘法器输出;a plurality of multipliers, each multiplier in the plurality of multipliers receives the first feedback filter input signal at one input and receives corresponding filter coefficients at the other input, and multiplies them to obtain a corresponding multiplier output; 多个延时器,所述多个延时器中的第一延时器接收所述多个乘法器中的第一乘法器的输出,对其进行延时以得到经延时的第一乘法器输出以输出到第一加法器,所述多个延时器中除第一延时器之外的其余延时器接收前一级加法器的输出作为输入;以及A plurality of delayers, a first delayer in the plurality of delayers receives the output of the first multiplier in the plurality of multipliers, and delays it to obtain a delayed first multiplication The output of the adder is output to the first adder, and the remaining delayers in the plurality of delayers except the first delayer receive the output of the previous stage adder as input; and 多个加法器,所述多个加法器中的每个加法器的一个输入端接收一个经延时的当前级乘法器的输出,另一个输入端接收一个下一级乘法器输出,并对它们进行相加,所述多个加法器中除最后一个加法器以外的所有加法器的输出均为下一级延时器的输入,所述多个加法器中最后一个加法器的输出为转置结构的输出,A plurality of adders, one input end of each adder in the plurality of adders receives the output of a delayed current stage multiplier, and the other input end receives a next stage multiplier output, and Adding, the output of all adders except the last adder in the plurality of adders is the input of the next stage delayer, and the output of the last adder in the plurality of adders is transposed output of the structure, 所述非转置结构包括:The non-transposed structures include: 多个延时器,所述多个延时器中的第一延时器接收第二反馈滤波器的输入信号,所述多个延时器依次级联;A plurality of delayers, the first delayer in the plurality of delayers receives the input signal of the second feedback filter, and the plurality of delayers are cascaded in sequence; 多个乘法器,所述多个乘法器中的每一个乘法器的一个输入端接收经相应延时器延时的输入信号,另一个输入端接收相应的滤波系数,并对它们相乘以得到多个相应的乘法器输出;以及Multiple multipliers, one input terminal of each multiplier in the multiple multipliers receives the input signal delayed by the corresponding delayer, and the other input terminal receives corresponding filter coefficients, and multiplies them to obtain a plurality of corresponding multiplier outputs; and 一个加法器,所述加法器对所述多个乘法器输出进行相加,输出为非转置结构的输出。An adder, the adder adds the outputs of the multiple multipliers, and the output is an output of a non-transposed structure. 3.如权利要求1或2所述的判决反馈滤波器,其特征在于还包括多个延时器,所述多个延时器是级联的,前一级延时器的输出为下一级延时器的输入,其中多个延时器中第一延时器的输入是所述第一反馈滤波器的输入信号,相对第一反馈滤波器之输入信号具有所述第一预定延时的延时器输出是所述第二反馈滤波器的输入。3. decision feedback filter as claimed in claim 1 or 2, is characterized in that also comprising a plurality of time delay devices, and described multiple time delay devices are cascaded, and the output of previous stage time delay device is next The input of the stage delayer, wherein the input of the first delayer in the plurality of delayers is the input signal of the first feedback filter, and has the first predetermined delay relative to the input signal of the first feedback filter The output of the delayer is the input of the second feedback filter. 4.如权利要求3所述的判决反馈滤波器,其特征在于,所述转置结构为B级,所述非转置结构为M级,M=2B-A,其中A是实现所述第一预定延时的延时器的级数,B和M的值是根据所需要对付的最长后径的长度决定的。4. The decision feedback filter according to claim 3, wherein the transpose structure is B-level, and the non-transpose structure is M-level, M=2 BA , wherein A is to realize the first The number of stages of the delayer of predetermined time-delay, the value of B and M are determined according to the length of the longest rear diameter that needs to deal with. 5.如权利要求3或4所述的判决反馈滤波器,其特征在于,所述第一反馈滤波器和第二反馈滤波器的输出之和为所述判决反馈滤波器的输出。5. The decision feedback filter according to claim 3 or 4, wherein the sum of the outputs of the first feedback filter and the second feedback filter is the output of the decision feedback filter. 6.如权利要求1所述的判决反馈滤波器,其特征在于还包括第三反馈滤波器,其输入信号相对于第一反馈滤波器的输入信号具有第二预定延时,所述第二预定延时大于所述第一预定延时,所述第三反馈滤波器为非转置结构。6. The decision feedback filter as claimed in claim 1, further comprising a third feedback filter whose input signal has a second predetermined time delay with respect to the input signal of the first feedback filter, the second predetermined The delay is greater than the first predetermined delay, and the third feedback filter has a non-transpose structure. 7.如权利要求5或6的所述的判决反馈均衡器,其特征在于,所述第一反馈滤波器、第二反馈滤波器、第三反馈滤波器中的任一个是实数滤波器、复数滤波器和准复数滤波器的其中之一。7. The decision feedback equalizer as claimed in claim 5 or 6, wherein any one of the first feedback filter, the second feedback filter, and the third feedback filter is a real number filter, a complex number One of filter and quasi-complex filter. 8.如权利要求4所述的判决反馈滤波器,其特征在于,A为大于等于0的整数。8. The decision feedback filter according to claim 4, wherein A is an integer greater than or equal to 0. 9.一种时域自适应均衡器,包括:9. A time-domain adaptive equalizer, comprising: 前向滤波器,其接收输入信号,用于消除前向多径;A forward filter, which receives an input signal, is used to eliminate forward multipath; 判决反馈滤波器,用于消除后向多径;decision feedback filter for eliminating backward multipath; 加法器,其接收来自前向滤波器和判决反馈滤波器的输出作为输入,进行求和操作并输出一个均衡器输出;an adder, which receives as input the outputs from the forward filter and the decision feedback filter, performs a summation operation and outputs an equalizer output; 判决器,接收所述均衡器输出作为输入信号,并将输出提供给所述判决反馈波滤器;a decision unit, receiving the equalizer output as an input signal, and providing an output to the decision feedback filter; 误差信号生成器,根据所述判决器的输入信号和所述判决器的输出来生成误差信号;an error signal generator, which generates an error signal according to the input signal of the decision device and the output of the decision device; 抽头系数更新器,根据所述误差信号生成抽头系数以更新前向滤波器和判决反馈滤波器的抽头系数,a tap coefficient updater generating tap coefficients according to the error signal to update the tap coefficients of the forward filter and the decision feedback filter, 其特征在于,所述判决反馈滤波器包括:It is characterized in that the decision feedback filter includes: 第一反馈滤波器,其具有转置结构;a first feedback filter having a transpose structure; 第二反馈滤波器,其输入信号相对于第一反馈滤波器的输入信号具有第一预定延时并且具有非转置结构。A second feedback filter whose input signal has a first predetermined delay relative to the input signal of the first feedback filter and has a non-transposed structure. 10.如权利要求9所述的时域自适应均衡器,其特征在于,10. time-domain adaptive equalizer as claimed in claim 9, is characterized in that, 所述转置结构包括:The transpose structure includes: 多个乘法器,所述多个乘法器中的每个乘法器的一个输入端接收第一反馈滤波器输入信号,另一个输入端接收相应的滤波系数,并对它们相乘以得到一相应的乘法器输出;a plurality of multipliers, each multiplier in the plurality of multipliers receives the first feedback filter input signal at one input and receives corresponding filter coefficients at the other input, and multiplies them to obtain a corresponding multiplier output; 多个延时器,所述多个延时器中的第一延时器接收所述多个乘法器中的第一乘法器的输出,对其进行延时以得到经延时的第一乘法器输出以输出到第一加法器,所述多个延时器中除第一延时器之外的其余延时器接收前一级加法器的输出作为输入;以及A plurality of delayers, a first delayer in the plurality of delayers receives the output of the first multiplier in the plurality of multipliers, and delays it to obtain a delayed first multiplication The output of the adder is output to the first adder, and the remaining delayers in the plurality of delayers except the first delayer receive the output of the previous stage adder as input; and 多个加法器,所述多个加法器中的每个加法器的一个输入端接收一个经延时的当前级乘法器的输出,另一个输入端接收一个下一级乘法器输出,并对它们进行相加,所述多个加法器中除最后一个加法器以外的所有加法器的输出均为下一级延时器的输入,所述多个加法器中最后一个加法器的输出为转置结构的输出,A plurality of adders, one input end of each adder in the plurality of adders receives the output of a delayed current stage multiplier, and the other input end receives a next stage multiplier output, and Adding, the output of all adders except the last adder in the plurality of adders is the input of the next stage delayer, and the output of the last adder in the plurality of adders is transposed output of the structure, 所述非转置结构包括:The non-transposed structures include: 多个延时器,所述多个延时器中的第一延时器接收第二反馈滤波器的输入信号,所述多个延时器依次级联;A plurality of delayers, the first delayer in the plurality of delayers receives the input signal of the second feedback filter, and the plurality of delayers are cascaded in sequence; 多个乘法器,所述多个乘法器中的每一个乘法器的一个输入端接收经相应延时器延时的输入信号,另一个输入端接收相应的滤波系数,并对它们相乘以得到多个相应的乘法器输出;以及Multiple multipliers, one input terminal of each multiplier in the multiple multipliers receives the input signal delayed by the corresponding delayer, and the other input terminal receives corresponding filter coefficients, and multiplies them to obtain a plurality of corresponding multiplier outputs; and 一个加法器,所述加法器对所述多个乘法器输出进行相加,输出为非转置结构的输出。An adder, the adder adds the outputs of the multiple multipliers, and the output is an output of a non-transposed structure. 11.如权利要求9或10所述的时域自适应均衡器,其特征在于还包括多个延时器,所述多个延时器是级联的,前一级延时器的输出为下一级延时器的输入,其中多个延时器中第一延时器的输入是所述第一反馈滤波器的输入信号,相对第一反馈滤波器之输入信号具有所述第一预定延时的延时器输出是所述第二反馈滤波器的输入。11. time-domain adaptive equalizer as claimed in claim 9 or 10, is characterized in that also comprising a plurality of time delay devices, and described multiple time delay devices are cascaded, and the output of preceding stage time delay device is The input of the next-stage delayer, wherein the input of the first delayer in the plurality of delayers is the input signal of the first feedback filter, and has the first predetermined value relative to the input signal of the first feedback filter. The delayed delayer output is the input to the second feedback filter. 12.如权利要求11所述的时域自适应均衡器,其特征在于,所述转置结构为B级,所述非转置结构为M级,M=2B-A,其中A是实现所述第一预定延时的延时器的级数,B和M的值是根据所需要对付的最长后径的长度决定的。12. The time-domain adaptive equalizer as claimed in claim 11, wherein the transpose structure is B-level, and the non-transpose structure is M-level, M=2 BA , wherein A is to realize the The number of stages of the delayer for the first predetermined time delay, the values of B and M are determined according to the length of the longest rear path that needs to be dealt with. 13.如权利要求11或12所述的时域自适应均衡器,其特征在于,所述第一反馈滤波器和第二反馈滤波器的输出之和为所述判决反馈滤波器的输出。13. The time-domain adaptive equalizer according to claim 11 or 12, wherein the sum of the outputs of the first feedback filter and the second feedback filter is the output of the decision feedback filter. 14.如权利要求9所述的时域自适应均衡器,其特征在于还包括第三反馈滤波器,其输入信号相对于第一反馈滤波器的输入信号具有第二预定延时,所述第二预定延时大于所述第一预定延时,所述第三反馈滤波器为非转置结构。14. The time-domain adaptive equalizer as claimed in claim 9, further comprising a third feedback filter whose input signal has a second predetermined time delay with respect to the input signal of the first feedback filter, said first The second predetermined delay is longer than the first predetermined delay, and the third feedback filter has a non-transposed structure. 15.如权利要求13或14所述的时域自适应均衡器,其特征在于,所述第一反馈滤波器、第二反馈滤波器、第三反馈滤波器中的任一个是实数滤波器、复数滤波器和准复数滤波器的其中之一。15. time-domain adaptive equalizer as claimed in claim 13 or 14, is characterized in that, any one in described first feedback filter, the second feedback filter, the 3rd feedback filter is a real number filter, One of a complex filter and a quasi-complex filter. 16.如权利要求12所述的时域自适应均衡器,其特征在于,A为大于等于0的整数。16. The time-domain adaptive equalizer according to claim 12, wherein A is an integer greater than or equal to 0.
CNB2006100283491A 2006-06-29 2006-06-29 Time Domain Adaptive Equalizer and Its Contained Decision Feedback Filter Expired - Fee Related CN100562076C (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CNB2006100283491A CN100562076C (en) 2006-06-29 2006-06-29 Time Domain Adaptive Equalizer and Its Contained Decision Feedback Filter
HK08107321.7A HK1116970B (en) 2008-07-02 Time-domain adaptive equalizer and judging feedback filter included in which

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CNB2006100283491A CN100562076C (en) 2006-06-29 2006-06-29 Time Domain Adaptive Equalizer and Its Contained Decision Feedback Filter

Publications (2)

Publication Number Publication Date
CN101098416A CN101098416A (en) 2008-01-02
CN100562076C true CN100562076C (en) 2009-11-18

Family

ID=39011898

Family Applications (1)

Application Number Title Priority Date Filing Date
CNB2006100283491A Expired - Fee Related CN100562076C (en) 2006-06-29 2006-06-29 Time Domain Adaptive Equalizer and Its Contained Decision Feedback Filter

Country Status (1)

Country Link
CN (1) CN100562076C (en)

Families Citing this family (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101527697B (en) * 2009-03-16 2012-08-29 上海高清数字科技产业有限公司 Time-domain automatic-adaptive equalizer and equalizing method thereof
CN102223328B (en) * 2010-04-14 2014-04-16 华为技术有限公司 Signal decision method and device
CN102437978B (en) * 2010-09-29 2015-03-25 中兴通讯股份有限公司 Method and device for balancing digital microwaves
CN103179060B (en) * 2011-12-26 2016-04-27 富士通株式会社 The coefficient updating device of equalizer and method
EP2613488B1 (en) 2012-01-21 2016-01-20 Huawei Technologies Co., Ltd. Adaptive equalization method and adaptive equalizer
CN103067320B (en) * 2012-12-28 2015-09-16 成都泰格微波技术股份有限公司 Mesh MANET channel adaptive equalization device
CN103516648B (en) * 2013-10-18 2016-09-28 南京信息工程大学 A kind of mould judgement multi-mode blind equalization method of correction
CN104682919B (en) * 2013-11-29 2018-02-27 无锡华润矽科微电子有限公司 A kind of method that the frequency of occurrences is drifted about after solution balanced device fixed point
CN107005307B (en) * 2014-12-09 2019-06-28 华为技术有限公司 A kind of method and balancer that balancer is set
CN107534629B (en) * 2015-04-10 2020-02-14 华为技术有限公司 Decision feedback equalization device and method and optical transmission system
CN105827557A (en) * 2016-05-24 2016-08-03 桂林市思奇通信设备有限公司 Time-domain equalizer based on MIMO (Multiple-Input Multiple-Output)
CN108616265B (en) * 2018-05-04 2022-07-01 重庆邮电大学 Circuit structure of RNS DWT filter bank based on five-modulus residue number basis
CN110224955B (en) * 2019-06-19 2020-07-03 北京邮电大学 Digital equalizer structure and implementation method
CN110957995B (en) * 2019-10-25 2023-06-30 四川汇源光通信有限公司 IIR filtering system of distributed fault monitoring device and design method thereof
CN113595949B (en) * 2021-09-30 2021-12-21 苏州浪潮智能科技有限公司 Adaptive PAM4 Decision Feedback Equalization Circuit
CN113992485B (en) * 2021-10-27 2023-05-30 西安微电子技术研究所 Decision feedback equalization circuit and high-speed signal channel transmission system

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
HDTV中分层结构的均衡器的设计. 宋杭宾,孟利民.浙江工业大学学报,第27卷第3期. 1999 *

Also Published As

Publication number Publication date
CN101098416A (en) 2008-01-02
HK1116970A1 (en) 2009-01-02

Similar Documents

Publication Publication Date Title
CN100562076C (en) Time Domain Adaptive Equalizer and Its Contained Decision Feedback Filter
Meher et al. Critical-path analysis and low-complexity implementation of the LMS adaptive algorithm
JP3357956B2 (en) Decision feedback equalizer
Azadet et al. Low-power equalizer architectures for high speed modems
Vijetha et al. Low power low area VLSI implementation of adaptive FIR filter using DA for decision feed back equalizer
Lawrence et al. Multiprocessor implementation of adaptive digital filters
CN100547988C (en) Adaptive Equalization Circuit in Cable Digital TV
CN102457251B (en) Method and device for realizing universal digital filter
CN107786476B (en) A filter, time domain equalizer and receiver
US5898731A (en) Auto-coefficient renewal digital channel equalizer
Choo et al. Two's complement computation sharing multiplier and its applications to high performance DFE
US6993071B2 (en) Low-cost high-speed multiplier/accumulator unit for decision feedback equalizers
CN101304495B (en) Ghost cancellation equalizer
US20070014345A1 (en) Low complexity Tomlinson-Harashima precoders
TWI260891B (en) Pipelined adaptive decision feedback equalizer
US6944218B2 (en) Adaptive filter having a small circuit scale with a low power consumption and tap-coefficients updating method of adaptive filter
HK1116970B (en) Time-domain adaptive equalizer and judging feedback filter included in which
Lutkemeyer et al. A transversal equalizer with an increased adaptation speed and tracking capability
Perry et al. Pipelined DFE architectures using delayed coefficient adaptation
CN103346822B (en) For the convertible equalizer of demodulator
Wolf et al. Low complexity equalization for cable modems
Lim Pipelined FIR filter architecture and increasing performance of decision-feedback equalizer
Mizuno et al. A high-throughput pipelined architecture for blind adaptive equalization with minimum latency
Nedovic et al. FIR Filter for Adaptive Equalization in PRML Read Channels
Deschamps et al. Pipeline

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: 1116970

Country of ref document: HK

C14 Grant of patent or utility model
GR01 Patent grant
REG Reference to a national code

Ref country code: HK

Ref legal event code: GR

Ref document number: 1116970

Country of ref document: HK

CF01 Termination of patent right due to non-payment of annual fee
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20091118