CN1798116B - Weighting type equalizer in adaptivity, and related method - Google Patents

Abstract

The self-adapting equalizer consists of a reference signals generator used to produce a reference signal in term of first signals source, an equalizing unit used to make calculation for a received signal in term of multi equalizing coefficient, a weighted signal generator used to produce a weighted signal in term of the second signal source, and a equalizing coefficient self-adaptor used to adjust the equalizing coefficient in term of the reference signal, the equalizing signal, the weighted signal and the received signal.

Description

Weighting type equalizer in adaptivity and correlation technique thereof

Technical field

The invention provides a kind of equalizer in adaptivity and its method of operation, refer to a kind of equalizer in adaptivity and method of utilizing a weighted signal to control the weight of adjusting equalizing coefficient especially.

Background technology

In electrical communication system now, the proportion of digital data transfer has the trend that increases day by day, for more high-quality delivery means is provided, must manage to overcome the unfavorable characteristic of passage.Therefore, in general communication system,, often need add equalizer (equalizer) at receiver front end in order to reduce of the influence of the unfavorable characteristic of passage to system, influence with reduction channel noise or other interference is earlier deciphered again, to reach the purpose of high-quality transmission.For instance: equalizer in adaptivity comes a balanced received signal according to a plurality of equalizing coefficients, and above-mentioned equalizing coefficient is adjusted according to a received signal and a reference signal, this reference signal may by a training sequence (training sequence) or carrying out the bit combination of being exported after the decoding computing by cutting decoder (slicer decoder) or Vitebi detector (Viterbi detector) be produced, then according to the balanced above-mentioned received signal of adjusted equalizing coefficient.In addition, above-mentioned cutter decoder directly compares the signal after the equilibrium and a cleavage standard, to judge that the data in the received signal is " 0 " or " 1 "; Then when the judgement data was " 0 " or " 1 ", the relevance before and after will consider between the signal made that the correctness of decoding is higher to Vitebi detector.

See also Fig. 1, Fig. 1 is the function block schematic diagram of existing equalizer in adaptivity 10.Equalizer in adaptivity 10 includes a balanced unit 12, a reference signal generator 14 and an equalizing coefficient from suitable device (coefficient adapting circuit) 15, and wherein equalizing coefficient is provided with an error arithmetic element 16 and a coefficient arithmetic element 18 in suitable device 15.At first, balanced unit 12 is used for according to a plurality of equalizing coefficient C ₀(0), C ₁(0) ..., C _N(0) a received signal y is carried out computing to produce an equalizing signal y _Eq, its detailed operation mode will be described further hereinafter.Then, reference signal generator 14 can utilize a training sequence (training sequence) or cut the bit combination d that is exported after decoder, the Vitebi detector execution decoding computing and produce an expection signal (desired signal)

And then will expect signal by error arithmetic element 16

With equalizing signal y _EqSubtract each other to produce an error signal e.At last, equalizer in adaptivity 10 is convenient carries out lowest mean square with coefficient arithmetic element 18 according to received signal y and error signal e (Least Mean Square, LMS) computing is to produce preferable a plurality of equalizing coefficient C ₀(1), C ₁(1) ..., C _N(1), carries out so repeatedly.Following equation can be used to illustrate the calculating process of existing coefficient arithmetic element 18:

C _j(k)=C _j(k-1)-τ e (k) y (k-j) equation ()

In the equation (), τ represents the class interval of coefficient adjustment, and it can be a default value or change according to system requirements, for example, when external environment changes when violent, τ can be made as bigger numerical value, and this moment, coefficient can be adjusted to stable status more quickly; Relative, if τ is made as less numerical value,, relatively is not easy the situation that causes equalizing coefficient not restrain, thereby can avoids causing grave error though balanced unit 12 can reach stable status the long time of cost.Computing and upgrade equalizing coefficient C so repeatedly ₀, C ₁... C _NThe result can make error signal e can level off to zero gradually, this moment equalizing coefficient C ₀, C ₁... C _NAlso can stablize by convergence, certainly, when external environment again during change, error signal e can be amplified thereupon, and this equalizing coefficient also can produce tangible change, so equalizer in adaptivity 10 can dynamically be adjusted the equalizing coefficient that balanced unit 12 is adopted according to aforesaid operations again.

See also Fig. 2, Fig. 2 is the enlarged diagram of balanced unit 12 shown in Figure 1.Balanced unit 12 includes a plurality of delay cells 22,24,26, the multiplier 32,34,36,38 of a plurality of adjustable integral coefficients, and a plurality of adder 42,44,46.22,24,26 time of delays that provided of delay cell are corresponding to the sample time of received signal y, therefore convenient received signal y (k), the y (k-1) that produces a plurality of different time points of taking a sample with received signal y respectively ... y (k-N), that is inhibit signal.Multiplier 32,34,36,38 is used for according to this adjustable equalizing coefficient C ₀, C ₁... C _NBe multiplied by its corresponding inhibit signal y (k), y (k-1) respectively ... y (k-N), multiplied result added up promptly produce equalizing signal y at last _EqSo the function mode of balanced unit 12 can be illustrated by following equation:

y _Eq(k)=C ₀Y (k)+C ₁Y (k-1)+... C _NY (k-N) equation (two)

Yet, can find that in practical application probability that some fragment in the signal makes a mistake is than higher, especially near received signal generation transition (transition), because existing equalizer in adaptivity also can't be done special processing at the above-mentioned fragment that is easier to make a mistake, therefore, if can reduce the error rate of these fragments then can significantly improve whole average error rate.

Summary of the invention

Therefore, one of main purpose of the present invention is to provide a kind of weighting type equalizer in adaptivity and correlation technique, to address the above problem.

The present invention discloses a kind of equalizer in adaptivity (equalizer), and this equalizer in adaptivity includes: a reference signal generator is used for producing a reference signal according to one first derived reference signal; One balanced unit (equalization unit) is used for coming a received signal is carried out computing according to a plurality of equalizing coefficients, to produce an equalizing signal; One weighted signal generator is used for producing a weighted signal according to one second derived reference signal; And the certainly suitable device (coefficient adapting circuit) of an equalizing coefficient, be used for according to this reference signal, this equalizing signal, this weighted signal, and this received signal to be to adjust this a plurality of equalizing coefficients, and wherein, this equalizing coefficient includes from fitting device: an error arithmetic element, be electrically connected to this balanced unit and this reference signal generator, be used for adjusting the difference of this reference signal and this equalizing signal to export an error signal; And a coefficient arithmetic element, be used for according to this received signal and these a plurality of equalizing coefficients of this error signal adjustment.

In addition, the present invention discloses a kind of equalizer in adaptivity in addition, and it includes: a reference signal generator is used for producing a reference signal according to one first derived reference signal; One balanced unit is used for coming a received signal is carried out computing according to a plurality of equalizing coefficients, to produce an equalizing signal; And the certainly suitable device of an equalizing coefficient, in order to vectorial according to a reflection, this reference signal, this equalizing signal, and this received signal is adjusted this a plurality of equalizing coefficients; Wherein, this equalizing coefficient includes from suitable device: an error arithmetic element, be electrically connected to this balanced unit and this reference signal generator, and be used for exporting the error signal of this equalizing signal and this reference signal; One unit map is electrically connected to this error generator, is used for producing a reflection received signal according to a reflection vector and this received signal, and produces a reflection error signal according to this reflection vector and this error signal; And a coefficient arithmetic element, be electrically connected to this unit map, be used for adjusting this a plurality of equalizing coefficients according to this reflection error signal and this reflection received signal.

The present invention discloses a kind of self adaptive equalization methods in addition.This self adaptive equalization methods includes: produce a reference signal according to one first derived reference signal; Come a received signal is carried out computing according to a plurality of equalizing coefficients, to produce an equalizing signal; Produce a weighted signal according to one second derived reference signal; And according to this weighted signal, this reference signal, this equalizing signal and this received signal are adjusted this a plurality of equalizing coefficients, and wherein, the step of adjusting these a plurality of equalizing coefficients includes: the difference of adjusting this reference signal and this equalizing signal is to produce an error signal; According to this received signal and these a plurality of equalizing coefficients of this error signal adjustment.

In addition, the present invention also discloses a kind of self adaptive equalization methods.This self adaptive equalization methods includes: produce a reference signal according to one first derived reference signal; Come a received signal is carried out computing according to a plurality of equalizing coefficients, to produce an equalizing signal; And according to a reflection vector, this reference signal, this equalizing signal and this received signal are adjusted this a plurality of equalizing coefficients; Wherein, the step of adjusting these a plurality of equalizing coefficients includes: produce an error signal according to this equalizing signal and this reference signal; Produce a reflection received signal according to a reflection vector and this received signal, and produce a reflection error signal according to this reflection vector and this error signal; And adjust this a plurality of equalizing coefficients according to this reflection error signal and this reflection received signal.

Therefore, the present invention utilizes the weighted signal generator at the fragment that is easier in the signal make a mistake, weight when strengthening adjusting equalizing coefficient, and utilize a unit map to capture the noise contribution that decoding error more easily takes place according to above-mentioned reflection vector, calculate preferable equalizing coefficient at this noise contribution then and use, to reach the purpose of significantly improving the ensemble average error rate for balanced unit.

Description of drawings

Fig. 1 is the function block schematic diagram of existing equalizer in adaptivity.

Fig. 2 is the function block schematic diagram of balanced unit shown in Figure 1.

Fig. 3 is the function block schematic diagram of first embodiment of first kind of equalizer in adaptivity of the present invention.

Fig. 4 is the function block schematic diagram of second embodiment of first kind of equalizer in adaptivity of the present invention.

Fig. 5 is the function block schematic diagram of first embodiment of second kind of equalizer in adaptivity of the present invention.

Fig. 6 is the function block schematic diagram of second embodiment of second kind of equalizer in adaptivity of the present invention.

Fig. 7 is the function block schematic diagram of the 3rd embodiment of second kind of equalizer in adaptivity of the present invention.

Fig. 8 is the function block schematic diagram of the embodiment of the third equalizer in adaptivity of the present invention.

Component symbol

10,50,70,110 equalizer in adaptivity, 12,52,72,112 balanced units

14,54,74,114 reference signal generators, 15,58,78,118 equalizing coefficients are from suitable device

16,62,82,122 error arithmetic elements, 18,64,86,126 coefficient arithmetic elements

22,24,26 delay cells, 32,34,36,38 multipliers

42,44,46 adders, 51,71,111 Vitebi detectors

56,76 weighted signal generators 84,84a, 84b, 124 unit map

Embodiment

See also Fig. 3, Fig. 3 is the function block schematic diagram of first embodiment of first kind of equalizer in adaptivity 50 of the present invention.As shown in FIG., equalizer in adaptivity 50 is electrically connected on the special ratio sensor 51 of one dimension, and Vitebi detector 51 is used for handling the equalizing signal y that equalizer in adaptivity 50 is exported _Eq, to export a bit combination d.In addition, equalizer in adaptivity 50 includes a balanced unit 52, a reference signal generator 54, a weighted signal generator 56 and an equalizing coefficient from suitable device (coefficient adaptingcircuit) 58, wherein, equalizing coefficient is provided with an error arithmetic element 62 and a coefficient arithmetic element 64 in suitable device 58.Balanced unit 52 is according to a plurality of equalizing coefficient C ₀, C ₁..., C _NCarry out balancing operational to produce an equalizing signal y with a received signal y _Eq, for instance, received signal y can be a fundamental frequency signal in the wireless telecommunication system or a CD readback signal (disc read-back signal).

Reference signal generator 54 produces an expection signal according to one first derived reference signal In the present embodiment, first derived reference signal can be the signal that other used in the bit combination d that exported of follow-up Vitebi detector 51 or the system, and for example: received signal y, received signal y are through the output of an existing cutting decoder (slicer decoder) or the equalizing signal y of corresponding received signal y _Eq, note that employed reference signal generator 54 among the present invention running (that is the expection signal Production method) do not exceed with present embodiment because this one produces expection signal Technology be widely used in the existing equalizer in adaptivity, so belong to the known technology of industry, so the function mode of reference signal generator 54 is no longer in hereinafter giving unnecessary details, therefore, any being applied to can produce the expection signal in the existing equalizer in adaptivity Method and corresponding device thereof, all can be applicable to reference signal generator 54 of the present invention.For instance, present embodiment adopts bit combination d as this first derived reference signal, because the bit combination d that Vitebi detector 51 is exported in decode procedure is regarded as a correct decoding data, so the expection signal that is produced according to this decoding data

Can be considered received signal y should be corresponding desired value.For example: when channel pattern is (1,2,2,1), can produce the expection signal according to following equation

$\hat{y}\left(k\right)=1\·d\left(k\right)+2\·d(k-1)+2\·d(k-2)+1\·d(k-3)$ Equation (three)

Note that this difference can be eliminated by suitable delay element owing to the difference that may have between reference signal and received signal on the sequential.

Weighted signal generator 56 produces a weighted signal w according to one second derived reference signal, and drives equalizing coefficient from suitable device 58 according to weighted signal w.In the present embodiment, this second derived reference signal can be that (for example: received signal y, received signal y are through output, the expection signal of an existing cutting decoder (slicer) for the signal that other used in the bit combination d that exported of above-mentioned Vitebi detector 51 or the system

Or equalizing signal y _Eq).For convenience of description, hereinafter will set forth respectively and utilize the signal that is used in bit combination d and other system to produce the judgment mechanism of weighted signal w.In the present embodiment, at first when the bit combination d that utilizes above-mentioned Vitebi detector 51 to be exported is used as this second derived reference signal, if this bit combination d meets specific combination, then weighted signal generator 56 produces the weighted signal w of special value, the accurate position of tangible logic for example takes place change (transition), then the weighted signal w of weighted signal generator 56 generations is " 1 "; Under other situation, the weighted signal w of weighted signal generator 56 outputs is " 0 "; Therefore, equalizing coefficient can carry out the computing of optimum equalizer coefficient according to weighted signal w from suitable device 58.Following equation can be used to illustrate that weighted signal generator 56 is used for producing the mechanism of weighted signal w;

$w\left(k\right)=\left\{\begin{array}{c}1,\mathrm{if}\left[\begin{array}{ccccc}{d}_{k-2}& d_{k-1}& d& d_{k+1}& d_{k+2}\end{array}\right]=\left[00X11\right]\mathrm{or}\left[11X00\right]\\ 0,\mathrm{else}\end{array}\right\}$ Equation (four)

In the equation (four), X represents any logical value, that is X can be " 0 " or " 1 ", in other words, when weighted signal generator 56 detects bit stream (bit stream) among the bit combination d for " 00011 ", " 00111 ", " 11100 ", " 11000 ", weighted signal generator 56 can judge that just this bit combination d produces the accurate position of logic and changes, and is the weighted signal w of " 1 " so weighted signal generator 56 is an output valve.Note that the bit length of being judged in the equation (four) (bit length) can be a system default value or arbitrary value, does not exceed with present embodiment.Please note, weighted signal w can correspond to plural any number (that is not being subject to 0 and 1), yet weighted signal w only uses 0 and during 1 two kind of numerical value, this special case can be considered as a kind of method of switch in the present embodiment, controls equalizing coefficient and activates from suitable device 58 and adjust equalizing coefficient C ₀, C ₁..., C _N

In addition, if the signal that utilizes in the system to be used is used as this second derived reference signal, then with equalizing signal y _EqBe example, as equalizing signal y _EqIn the different preset range in position, then weighted signal w can be set to different special values.For instance, if equalizing signal y _EqThere is following multiple numerical value in pairing ideal value: 6,4,2,0,-2,-4,-6, its production method replaces position " 0 " for a bit combination with four positions is replaced position " 1 " with+1 with-1, do inner product operation with vector (1 22 1) then, for instance, the numerical value that bit combination (1111) is produced is " 6 ", the numerical value that bit combination (1110) is produced is " 4 ", the numerical value that bit combination (0110) is produced is " 2 ", the numerical value that bit combination (1000) is produced is " 4 ", hence one can see that, and the absolute value of its corresponding ideal value of bit combination that the accurate position of normal more occurrence logic changes is more little.Because it is frequent more that the accurate position of the logic of bit combination changes, its probability that makes a mistake is also high more, so have only in the present embodiment as equalizing signal y _EqNumerical value be 2,0 or-2 o'clock, weighted signal generator 56 is promptly exported the weighted signal w that equals " 1 "; On the other hand, if equalizing signal y _EqCorresponding numerical value 6,4 ,-4 ,-6 o'clock, then weighted signal generator 56 is just exported the weighted signal w that equals " 0 "; And weighted signal w is passed to equalizing coefficient from suitable device 58, make it carry out the computing of optimum equalizer coefficient according to weighted signal w.Therefore, the producing method of weighted signal w can be illustrated with following equation:

$w\left(k\right)=\left\{\begin{array}{c}1,\mathrm{if}-3<y_{\mathrm{eq}}<3\\ 0,\mathrm{else}\end{array}\right\}$ Equation (five)

Note that and only use two preset range in the equation (five), that is (3＜y _EqAnd y _Eq＜3) with (y _Eq＞3 or y _Eq＜-3) two kinds, but the present invention can also make weighted signal w can correspond to plural any number according to the environment for use difference so that a plurality of preset range to be provided according to system design.As mentioned above, equalizing coefficient is from fitting device 58 according to weighted signal w, received signal y, expection signal

And equalizing signal y _EqCarry out computing, to produce preferable a plurality of equalizing coefficient C ₀, C ₁..., C _N, and upgrade a plurality of equalizing coefficients that use in the equalizer in adaptivity 50 further, so that obtain the higher equalizing signal y of reliability _EqIn the present embodiment, error arithmetic element 62 is used for calculating the expection signal And equalizing signal y _EqDifference to produce error signal e, then, coefficient arithmetic element 64 is utilized weighted signal w, error signal e, is carried out known lowest mean square computing to adjust equalizing coefficient C with received signal y ₀, C ₁..., C _NFollowing equation can be used to illustrate the calculating process of coefficient arithmetic element 64:

C _j(k)=C _j(k-1)-τ e (k) w (k) y (k-j) equation (six)

In the equation (six), τ represents the class interval of coefficient adjustment, can be a default value or change according to system requirements, in addition, when system judges that this moment, received signal had higher error rate, the weighted signal w that weighted signal generator 56 is produced corresponds to a bigger numerical value, so coefficient arithmetic element 64 can be adjusted equalizing coefficient C with bigger amplitude ₀, C ₁..., C _NOtherwise when system judged that this moment, received signal had lower error rate, the weighted signal w that weighted signal generator 56 is produced also can be less, so coefficient arithmetic element 64 also can be adjusted equalizing coefficient C with less amplitude ₀, C ₁..., C _NAt last, utilize the preferable equalizing coefficient C that calculates again ₀(k), C ₁(k) ... C _N(k) upgrade equalizer in adaptivity 50 employed equalizing coefficient C originally ₀(k-1), C ₁(k-1) ... C _N(k-1).Please pay special attention to, weighted signal w is used for control coefrficient arithmetic element 64 in the present embodiment, but find by being not difficult in the equation (six), if weighted signal w also can be used for departure arithmetic element 62, even also weighted signal w is multiplied by error signal e earlier and then input coefficient arithmetic element 64 is carried out computing, its result comes the operation result of control coefrficient arithmetic element 64 identical with utilizing weighted signal w, therefore, one of present embodiment change type comes departure arithmetic element 62 for utilizing weighted signal w in advance.

See also Fig. 4, Fig. 4 is the function block schematic diagram of second embodiment of first kind of equalizer in adaptivity 50 of the present invention.As shown in FIG., present embodiment and equalizer in adaptivity shown in Figure 3 50 unique differences are that weighted signal generator 56 is electrically connected to error generation unit 62, make error generation unit 62 adjust the expection signal according to weighted signal w And equalizing signal y _EqResidual quantity, with produce after the weighting error signal e to follow-up coefficient arithmetic element 64 to carry out the lowest mean square computing, produce preferable equalizing coefficient C ₀, C ₁... C _NAnd upgraded.

See also Fig. 5, Fig. 5 is the function block schematic diagram of first embodiment of second kind of equalizer in adaptivity 70 of the present invention.In the present embodiment, equalizer in adaptivity 70 is electrically connected on the special ratio sensor 71 of one dimension, and wherein, Vitebi detector 71 is used for handling the equalizing signal y that equalizer in adaptivity 70 is exported _Eq, to export a bit combination d.As shown in FIG., equalizer in adaptivity 70 includes a balanced unit 72, a reference signal generator 74, a weighted signal generator 76 and an equalizing coefficient from suitable device 78, and wherein equalizing coefficient is provided with an error arithmetic element 82, a unit map (projector) 84 and a coefficient arithmetic element 86 in suitable device 78.Because all the element of the same name with Fig. 3 is identical with function mode for the framework of balanced unit 72, reference signal generator 74, weighted signal generator 76 and error arithmetic element 82, so do not describe in detail one by one at this.

The equalizer in adaptivity 70 that present embodiment proposed is to have increased newly unit map 84 with the main difference of equalizer in adaptivity 50 shown in Figure 3.The function of unit map 84 is to filter out the interference noise of certain particular form, and this is because the interference noise of some form is easier to cause decoding error, and institute is so that adjust equalizing coefficient C at this interference noise especially ₀, C ₁..., C _N, that is wish and can carry out reducing to the influence of interference noise minimum before the follow-up Vitebi detector 71.For instance: " 6,4,0 ,-4 " and " 4; 0 ,-4 ,-6 " all are issuable normal symbol metasequences, supposes the data that " 4; 0 ,-4 ,-6 " are sent for tendency to develop, when through a communication channel and an interference noise synthesize, and this interference noise just is " 2,4,4,2 ", so 71 of Vitebi detectors that are positioned at receiving terminal can be judged the data that is received by accident and are " 6,4; 0 ,-4 ", follow-up thus decoding is moving to be done the bit combination of being exported and also can make a mistake.Therefore, present embodiment utilizes a reflection vector Reflect above-mentioned specific interference noise, make that unit map 84 can be according to the reflection vector

Y and error signal e are done inner product operation to produce a reflection received signal y to received signal respectively _vWith a reflection error signal e _vFollow procedure can be used to explanation reflection received signal y _vWith the reflection error signal e _vProducing method:

e _v(k)=[e (k-1) e (k) e (k+1) e (k+2)] [v ₁v ₂v ₃v ₄] equation (seven)

y _v(k)=[y (k-1) y (k) y (k+1) y (k+2)] [v ₁v ₂v ₃v ₄] equation (eight)

In equation (seven) and equation (eight), because reflection vector [v ₁v ₂v ₃v ₄] be interference noise with respect to above-mentioned particular form, so when channel pattern is (1,2,2,1), can adopt the reflection vector At last, again via coefficient arithmetic element 86 according to weighted signal w, reflection received signal y _v, with the reflection error signal e _vCarry out the lowest mean square computing and produce preferable equalizing coefficient C ₀(k), C ₁(k) ... C _N(k), to upgrade balanced unit 72 present employed equalizing coefficient C further ₀(k-1), C ₁(k-1) ..., C _N(k-1).Following equation can be used to illustrate the calculating process of coefficient arithmetic element 86:

C _j(k)=C _j(k-1)-τ w (k) e _v(k) y _v(k-j) equation (nine)

Note that according to equation (nine) as can be known weighted signal w also can be used for controlling unit map 84.

See also Fig. 6, Fig. 6 is the function block schematic diagram of second embodiment of second kind of equalizer in adaptivity 70 of the present invention.As shown in FIG., present embodiment and equalizer in adaptivity shown in Figure 5 70 unique differences are that weighted signal generator 76 is electrically connected to unit map 84, make unit map 84 select one according to weighted signal w and adjust reflection received signal y _vWith the reflection error signal e _v, received signal y then will video _vWith the reflection error signal e _vBe sent to coefficient arithmetic element 86 and carry out the lowest mean square computing, to produce preferable equalizing coefficient C ₀(k), C ₁(k) ... C _N(k) and upgrade existing equalizing coefficient C ₀(k-1), C ₁(k-1) ..., C _N(k-1).

See also Fig. 7, Fig. 7 is the function block schematic diagram of the 3rd embodiment of second kind of equalizer in adaptivity 70 of the present invention.As shown in FIG., the main difference point of present embodiment and equalizer in adaptivity 70 shown in Figure 6 is that equalizer in adaptivity shown in Figure 6 70 only includes a unit map 84, that equalizer in adaptivity 70 shown in Figure 8 then includes is a plurality of (for example: M) individual unit map 84a ... 84b, all the unit map 84 with shown in Figure 6 is identical with framework in its running, and this is because equalizer in adaptivity 70 can be at the different reflection vector of array in practical application Do reflection,, for example when channel pattern is (1,2,2,2,1), can adopt the reflection vector to handle the interference noise that the M kind is easier to cause decoding error respectively

Follow procedure can be used to illustrate its corresponding reflection received signal y _vWith the reflection error signal e _vProducing method:

$e_{v,i}\left(k\right)=\left[\begin{array}{ccccc}e(k-1)& e\left(k\right)& e(k+1)& ...& e(k-2+L_i)\end{array}\right]\&CenterDot;{\stackrel{\&RightArrow;}{v}}_i$ Equation (ten)

$y_{v,i}\left(k\right)=\left[\begin{array}{ccccc}y(k-1)& y\left(k\right)& y(k+1)& ...& y(k-2+L_i)\end{array}\right]\&CenterDot;{\stackrel{\&RightArrow;}{v}}_i$ Equation (11)

Wherein, L _iBe to correspond to the reflection vector

Dimension, unit map 84a is according to reflection vector

Produce the reflection error signal e _{V, 1}With reflection received signal y _{V, 1}The rest may be inferred, and unit map 84b is according to the reflection vector Produce the reflection error signal e _{V, n}With reflection received signal y _Vn, therefore, 86 of coefficient arithmetic elements are according to n reflection error signal e _{V, 1}..., e _{V, n}With n reflection received signal y _{V, 1}..., y _{V, n}Adjust equalizing coefficient C ₀, C ₁..., C _N, following equation can be used to illustrate equalizing coefficient C ₀, C ₁..., C _NCalculating process:

$C_j\left(k\right)=C_j(k-1)-\mathrm{\&tau;}\&CenterDot;({\mathrm{\&Sigma;}}_{i=1}^n{w}_i\left(k\right)\&CenterDot;e_{v,i}\left(k\right)\&CenterDot;y_{v,i}(k-j))$ Equation (12)

W wherein _i(k) expression corresponds to the reflection vector

Weighted signal, wherein correspond to each reflection vector

Weighted signal w _iCan be according to different reflection vectors

And different, one of method is according to reflection received signal y _{V, 1}..., y _{V, n}Produce weighted signal w _iThe another one method makes and corresponds to each reflection vector Weighted signal w _iAll identical, that is the M of weighted signal generator output signal w _iAll identical, maybe can be simplified to and only use an output signal w.Note that the weighted signal w in the equation (12) _iCan also be by coefficient arithmetic element 86 according to each reflection vector Importance adjust weighted signal w _iSize, for example: reflection vector

Importance big more (that is big more) to the correctness influence of follow-up decoding, the weighted signal w that it is corresponding _iAlso big more.

See also Fig. 8, Fig. 8 is the function block schematic diagram of the embodiment of the third equalizer in adaptivity 110 of the present invention.Equalizer in adaptivity 110 is electrically connected on Vitebi detector 111 in the present embodiment, and wherein, Vitebi detector 111 is used for handling the equalizing signal y that equalizer in adaptivity 110 is exported _Eq, to export a bit combination d.As shown in FIG., equalizer in adaptivity 110 includes a balanced unit 112, a reference signal generator 114 and an equalizing coefficient from suitable device 118, and wherein equalizing coefficient is provided with an error arithmetic element 122, a unit map 124 and a coefficient arithmetic element 126 in suitable device 118.Because all the element of the same name with Fig. 5 is identical with function mode for the framework of balanced unit 112, reference signal generator 114, error arithmetic element 122, unit map 124 and coefficient arithmetic element, so do not describe in detail one by one at this.Please consult the equalizer in adaptivity 70,110 shown in Fig. 5 and Fig. 8 simultaneously, its main difference point is that equalizer in adaptivity shown in Figure 8 110 does not use the weighted signal generator, to simplify equalizer in adaptivity 70 calculating processes, reduce system cost simultaneously, therefore, the coefficient arithmetic element 126 among Fig. 8 is only according to the reflection error signal e _vWith reflection received signal y _vAdjust equalizing coefficient C ₀, C ₁..., C _N, following equation can be used to illustrate equalizing coefficient C ₀, C ₁..., C _NCalculating process:

C _j(k)=C _j(k-1)-τ e _v(k) y _v(k-j) equation (13)

Embodiment among Fig. 8 only uses a unit map, and this embodiment also can expand into a plurality of unit map of use according to the mode of Fig. 7.

Compared to prior art, equalizer in adaptivity of the present invention is by the weighted signal generator, when some wrong incidence higher signal fragment produces, weight when coming departure arithmetic element, unit map and coefficient arithmetic element to carry out computing according to the weighted signal that is produced, and capture the noise contribution that decoding error more easily takes place by unit map, calculate preferable equalizing coefficient at this noise contribution then and use for balanced unit.Thus, by assisting of weighted signal generator and unit map, equalizer in adaptivity of the present invention just can significantly reduce the system mistake rate.

The above only is preferred embodiment of the present invention, and all equalizations of being done according to the present patent application claim change and modify, and all should belong to covering scope of the present invention.

Claims (50)

1. an equalizer in adaptivity is characterized in that, includes:

One reference signal generator is used for producing a reference signal according to one first derived reference signal;

One balanced unit is used for coming a received signal is carried out computing according to a plurality of equalizing coefficients, to produce an equalizing signal;

One weighted signal generator is used for producing a weighted signal according to one second derived reference signal; And

One equalizing coefficient is used for according to this reference signal from suitable device, this equalizing signal, and this weighted signal, and this received signal to be adjusting these a plurality of equalizing coefficients,

Wherein, this equalizing coefficient includes from suitable device:

One error arithmetic element is electrically connected to this balanced unit and this reference signal generator, is used for adjusting the difference of this reference signal and this equalizing signal to export an error signal; And

One coefficient arithmetic element is used for according to this received signal and these a plurality of equalizing coefficients of this error signal adjustment.

2. equalizer in adaptivity as claimed in claim 1 is characterized in that, this balanced unit is electrically connected on the special ratio sensor of one dimension, and this Vitebi detector produces an output signal according to this equalizing signal and is used as this second derived reference signal.

3. equalizer in adaptivity as claimed in claim 2 is characterized in that, this weighted signal generator is compared this second derived reference signal and a plurality of bit string, and produces this weighted signal according to the result of comparison.

4. equalizer in adaptivity as claimed in claim 1, it is characterized in that, this weighted signal generator is detected this second derived reference signal and is positioned at a preset range, then this weighted signal generator is according to producing this weighted signal with the corresponding weighting proportion of this preset range, to control this equalizing coefficient from fitting device according to these a plurality of equalizing coefficients of this weighted signal adjustment.

5. equalizer in adaptivity as claimed in claim 1 is characterized in that, this weighted signal generator uses this equalizing signal to be used as this second derived reference signal.

6. equalizer in adaptivity as claimed in claim 5, it is characterized in that, this weighted signal generator is detected this equalizing signal and is positioned at a preset range, and then this weighted signal generator produces this weighted signal, to control this equalizing coefficient from fitting device according to these a plurality of equalizing coefficients of this weighted signal adjustment.

7. equalizer in adaptivity as claimed in claim 1 is characterized in that, this received signal is used as this second derived reference signal.

8. equalizer in adaptivity as claimed in claim 7, it is characterized in that, this weighted signal generator is detected this received signal and is positioned at a preset range, then produces this weighted signal to control this equalizing coefficient from fitting device according to these a plurality of equalizing coefficients of this weighted signal adjustment.

9. equalizer in adaptivity as claimed in claim 1 is characterized in that, the reference signal of being exported by this reference signal generator is this second derived reference signal.

10. equalizer in adaptivity as claimed in claim 9 is characterized in that, this weighted signal generator is detected this reference signal and is positioned at a preset range, then produces this weighted signal to control this equalizing coefficient from suitable these a plurality of equalizing coefficients of device adjustment.

11. equalizer in adaptivity as claimed in claim 1 is characterized in that, this balanced unit is electrically connected on the special ratio sensor of one dimension, and this Vitebi detector produces an output signal according to this equalizing signal and is used as this first derived reference signal.

12. equalizer in adaptivity as claimed in claim 1 is characterized in that, this reference signal generator is used as this first derived reference signal with this equalizing signal.

13. equalizer in adaptivity as claimed in claim 1 is characterized in that, this reference signal generator is to be used as this first derived reference signal with this received signal.

14. equalizer in adaptivity as claimed in claim 1 is characterized in that, produces this first derived reference signal according to this received signal.

15. equalizer in adaptivity as claimed in claim 1 is characterized in that, this received signal is a CD readback signal.

16. equalizer in adaptivity as claimed in claim 1 is characterized in that, this error arithmetic element is electrically connected to this weighted signal generator, and the difference that is used for adjusting this reference signal and this equalizing signal according to this weighted signal to be exporting this error signal, and

This coefficient arithmetic element is used for adjusting this a plurality of equalizing coefficients according to this error signal and this received signal.

17. equalizer in adaptivity as claimed in claim 1 is characterized in that, this coefficient arithmetic element is electrically connected to this weighted signal generator, is used for according to this weighted signal this error signal and these a plurality of equalizing coefficients of this received signal adjustment.

18. equalizer in adaptivity as claimed in claim 1 is characterized in that, this equalizing coefficient includes from suitable device:

One unit map is electrically connected to this error arithmetic element, is used for producing a reflection received signal according to a reflection vector and this received signal, and produces a reflection error signal according to this reflection vector and this error signal; Wherein

This coefficient arithmetic element is electrically connected to this unit map and this weighted signal generator, is used for according to this weighted signal this reflection error signal and these a plurality of equalizing coefficients of received signal adjustment of should videoing.

19. equalizer in adaptivity as claimed in claim 1 is characterized in that, this equalizing coefficient includes from suitable device:

One unit map, be electrically connected to this error generator and this weighted signal generator, be used for producing a reflection received signal according to a reflection vector and this received signal, and produce a reflection error signal, and adjust this reflection received signal and one of them person of this received signal according to this weighted signal according to a reflection vector and this error signal; Wherein

This coefficient arithmetic element is electrically connected to this unit map, is used for adjusting this a plurality of equalizing coefficients according to this reflection error signal and this reflection received signal.

20. an equalizer in adaptivity is characterized in that, includes:

One reference signal generator is used for producing a reference signal according to one first derived reference signal;

One balanced unit is used for coming a received signal is carried out computing according to a plurality of equalizing coefficients, to produce an equalizing signal; And

One equalizing coefficient is from suitable device, and is vectorial in order to video according to one, this reference signal, and this equalizing signal, and this received signal is adjusted this a plurality of equalizing coefficients;

Wherein, this equalizing coefficient includes from suitable device:

One error arithmetic element is electrically connected to this balanced unit and this reference signal generator, is used for exporting the error signal of this equalizing signal and this reference signal;

One unit map is electrically connected to this error generator, is used for producing a reflection received signal according to a reflection vector and this received signal, and produces a reflection error signal according to this reflection vector and this error signal; And

One coefficient arithmetic element is electrically connected to this unit map, is used for adjusting this a plurality of equalizing coefficients according to this reflection error signal and this reflection received signal.

21. equalizer in adaptivity as claimed in claim 20 is characterized in that, this received signal is a CD readback signal.

22. equalizer in adaptivity as claimed in claim 20 is characterized in that, this balanced unit is electrically connected on the special ratio sensor of one dimension, and this Vitebi detector produces an output signal according to this equalizing signal and is used as this first derived reference signal.

23. equalizer in adaptivity as claimed in claim 20 is characterized in that, this reference signal generator is used as this first derived reference signal with this equalizing signal.

24. equalizer in adaptivity as claimed in claim 20 is characterized in that, this reference signal generator is used as this first derived reference signal with this received signal.

25. equalizer in adaptivity as claimed in claim 20 is characterized in that, produces this first derived reference signal according to this received signal.

26. a self adaptive equalization methods is characterized in that, includes:

Produce a reference signal according to one first derived reference signal;

Come a received signal is carried out computing to produce an equalizing signal according to a plurality of equalizing coefficients;

Produce a weighted signal according to one second derived reference signal; And

According to this weighted signal, this reference signal, this equalizing signal, and this received signal adjusts this a plurality of equalizing coefficients,

Wherein, the step of adjusting these a plurality of equalizing coefficients includes:

The difference of adjusting this reference signal and this equalizing signal is to produce an error signal;

According to this received signal and these a plurality of equalizing coefficients of this error signal adjustment.

27. self adaptive equalization methods as claimed in claim 26 is characterized in that, the output signal that is to use a Viterbi detecting unit is as this second derived reference signal, and wherein this Viterbi detection unit produces this output signal according to this equalizing signal.

28. self adaptive equalization methods as claimed in claim 27 is characterized in that, the step that produces this weighted signal is this second derived reference signal of comparison and a plurality of bit string, and produces this weighted signal according to the result of comparison.

29. self adaptive equalization methods as claimed in claim 27, it is characterized in that, the step that produces this weighted signal is that this second derived reference signal of detecting is positioned at a preset range, and produce this weighted signal according to the result of detecting, with control this equalizing coefficient from suitable device according to these a plurality of equalizing coefficients of this weighted signal adjustment.

30. self adaptive equalization methods as claimed in claim 26 is characterized in that, this equalizing signal is intended for this second derived reference signal.

31. self adaptive equalization methods as claimed in claim 30, it is characterized in that, the step that produces this weighted signal is whether this second derived reference signal of detecting is positioned at a preset range, and produce this weighted signal according to the result of detecting, with control this equalizing coefficient from suitable device according to these a plurality of equalizing coefficients of this weighted signal adjustment.

32. self adaptive equalization methods as claimed in claim 26 is characterized in that, this received signal is intended for this second derived reference signal.

33. self adaptive equalization methods as claimed in claim 32, it is characterized in that, whether the step that produces this weighted signal is positioned at a preset range for this received signal of detecting, and produce this weighted signal according to the result of detecting, with control this equalizing coefficient from suitable device according to these a plurality of equalizing coefficients of this weighted signal adjustment.

34. self adaptive equalization methods as claimed in claim 26 is characterized in that, this reference signal is used as this second derived reference signal.

35. self adaptive equalization methods as claimed in claim 34, it is characterized in that, whether the step that produces this weighted signal is positioned at a preset range for this reference signal of detecting, and produce this weighted signal according to the result of detecting, with control this equalizing coefficient from suitable device according to these a plurality of equalizing coefficients of this weighted signal adjustment.

36. self adaptive equalization methods as claimed in claim 26 is characterized in that, the output signal of using a Viterbi detecting unit is as this first derived reference signal, and wherein this Viterbi detection unit produces this output signal according to this equalizing signal.

37. self adaptive equalization methods as claimed in claim 26 is characterized in that, the step that produces this reference signal utilizes this equalizing signal to be used as this first derived reference signal.

38. self adaptive equalization methods as claimed in claim 26 is characterized in that, produces the step of this reference signal, is to be used as this first derived reference signal with this received signal.

39. self adaptive equalization methods as claimed in claim 26 is characterized in that, produces this first derived reference signal according to this received signal.

40. self adaptive equalization methods as claimed in claim 26 is characterized in that, this received signal is a CD readback signal.

41. self adaptive equalization methods as claimed in claim 26 is characterized in that, the difference of adjusting this reference signal and this equalizing signal according to this weighted signal to be producing this error signal, and

Utilize this error signal and this received signal to adjust this a plurality of equalizing coefficients.

42. self adaptive equalization methods as claimed in claim 26 is characterized in that,

According to this weighted signal, this error signal and this received signal are adjusted this a plurality of equalizing coefficients.

43. self adaptive equalization methods as claimed in claim 26 is characterized in that, the step of adjusting these a plurality of equalizing coefficients includes in addition:

Produce a reflection received signal according to a reflection vector and this received signal, and produce a reflection error signal, and adjust this reflection error signal and one of them person of this reflection received signal according to this weighted signal according to this reflection vector and this error signal; Wherein

Adjust this a plurality of equalizing coefficients according to this reflection error signal and this reflection received signal.

44. self adaptive equalization methods as claimed in claim 26 is characterized in that, the step of adjusting these a plurality of equalizing coefficients includes in addition:

Utilize a reflection vector and this received signal to produce a reflection received signal, and utilize this reflection vector and this error signal to produce a reflection error signal; Wherein

Control the weight of this reflection error signal and this reflection received signal according to this weighted signal, and according to this reflection error signal and this these a plurality of equalizing coefficients of received signal adjustment of videoing.

45. a self adaptive equalization methods is characterized in that, includes:

Produce a reference signal according to one first derived reference signal;

Come a received signal is carried out computing according to a plurality of equalizing coefficients, to produce an equalizing signal;

According to a reflection vector, this reference signal, this equalizing signal, and this received signal is adjusted this a plurality of equalizing coefficients;

Wherein, the step of adjusting these a plurality of equalizing coefficients includes:

Produce an error signal according to this equalizing signal and this reference signal;

Produce a reflection received signal according to a reflection vector and this received signal, and produce a reflection error signal according to this reflection vector and this error signal; And

Adjust this a plurality of equalizing coefficients according to this reflection error signal and this reflection received signal.

46. self adaptive equalization methods as claimed in claim 45 is characterized in that, this received signal is a CD readback signal.

47. self adaptive equalization methods as claimed in claim 45 is characterized in that, the output signal of using a Viterbi detecting unit is as this first derived reference signal, and wherein this Viterbi detection unit produces this output signal according to this equalizing signal.

48. self adaptive equalization methods as claimed in claim 45 is characterized in that, the step that produces this reference signal is to use this equalizing signal to be used as this first derived reference signal.

49. self adaptive equalization methods as claimed in claim 45 is characterized in that, produces the step of this reference signal, is used as this first derived reference signal with this received signal.

50. self adaptive equalization methods as claimed in claim 45 is characterized in that, produces this first derived reference signal according to this received signal.

Images