Coefficient tap leakage for fractionally-spaced equalizers

Claim 2

An arrangement for processing a received data signal which was transmitted over a transmission channel at a rate of 1/T symbols per second comprising sampling means (11, 12, 16) for forming a succession of samples of said signal at a predetermined rate means for forming a decision as to the value of each one of said symbols and for forming an error signal associated with each symbol, the last-mentioned means comprising means (25, 26) for multiplying each successive one of an ordered plurality of coefficients with a successive respective one of a plurality of said samples associated with said decision, means (27, 34, 35, 51) responsive to the resulting products for generating an equalized signal, and means (41, 42) for quantizing said equalized signal to form said decision, the value of said error signal being a function of the value of said equalized signal and the value of said one of said symbols, and updating means (121, 122) for determining at least a first value for each one of said coefficients by additively combining first and second terms with a previous value of that coefficient, said first term being a function of the error signal associated with a predetermined one of said symbols, said updating means including means (180) for subtracting the magnitude of said second term from said previous value if that value has a positive sign and for adding the magnitude of said second term to said previous value if that value has a negative sign, said updating means characterized in that the magnitude of said second term is independent of the magnitudes of any of said coefficients.

Claim 3

The invention of claim 1 wherein the magnitude of said second term is always non-zero.

Claim 4

The invention of claim 2 wherein the magnitude of said second term is a selected constant.

Claim 5

The invention of claim 3 wherein said predetermined rate exceeds 1/T samples per second.

Claim 6

The invention of claims 1, 3 or 4 wherein said equalized signal generating means includes means (127, 128) for summing said products.

Claim 7

The invention of claim 3 wherein said data signal is a data signal modulated at a predetermined carrier frequency and wherein said equalized signal generating means includes means (127, 128) for summing said products to form a modulated equalized signal and means (27, 51) for demodulating said modulated equalized signal to form said equalized signal.

Claim 8

The invention of claim 6 wherein said error signal is equal to the difference, modulated at said carrier frequency, between the pre-quantized value of said equalized signal and said decision and wherein said updating means further includes means (170) for forming said first term as a predetermined fraction of the product of said modulated difference with the sample with which said each one of said coefficients was multiplied in the formation of said decision.

Claim 9

The invention of claims 1, 3, 4 or 7 wherein said predetermined rate is at least (1+α)/T samples per second, where α is the fractional excess bandwidth.

Claim 10

An arrangement including sampling means (11, 12, 16) for forming samples of a received data signal which was transmitted over a transmission channel at a predetermined symbol rate, said sampling means operating at greater than said symbol rate, means (25, 26) operative at said symbol rate for multiplying ones of said samples by predetermined coefficients, decision circuit means (41, 42) responsive to the sum of the resulting products for forming a decision as to the value of each transmitted data symbol and for forming a corresponding error signal, and means (121, 122) for updating each of said coefficients by additively combining first and second updating terms therewith, said first term being a function of an error signal previously formed by said decision circuit means, said updating means characterized in that the magnitude of said second term is a constant, said magnitude being added to each negative-valued coefficient and subtracted from each positive-valued coefficient.

Claim 11

An arrangement including sampling means (11, 12, 16) for forming samples of a received data signal which was transmitted over a transmission channel at a predetermined symbol rate, said sampling means operating at greater than said symbol rate, means (25, 26) operative at said symbol rate for multiplying ones of said samples by predetermined coefficients, decision circuit means (41, 42) responsive to the sum of the resulting products for forming a decision as to the value of each transmitted data symbol and for forming a corresponding error signal, and means (121, 122) for updating each of said coefficients by additively combining first and second updating terms therewith, said first term being a function of an error signal previously formed by said decision circuit means, said updating means characterized in that the magnitude of said second term is a constant, said magnitude being added to each negative-valued coefficient and subtracted from each positive-valued coefficient.

Claim 12

In a receiver to which a carrier signal, modulated by a succession of data symbols occurring at T second intervals, is transmitted over a band-limited channel, apparatus comprising sampling means (11, 12, 16) for forming a succession of samples of said signal at a predetermined rate and equalizer means operative during each one of a succession of T second intervals for forming a decision as to the value of an individual one of said symbols, said equalizer means comprising means (25, 26) for multiplying at least individual ones of a predetermined number of the most recently formed ones of said samples with individual ones of a plurality of coefficients, each coefficient having a predetermined value associated with said one of said intervals, means (27, 51) for demodulating the sum of the resulting products to form a demodulated equalized signal, means (37, 41, 42) responsive to said demodulated equalized signal for forming said decision and for forming a passband error signal, said passband error signal being equal to the difference, multiplied by said carrier signal, between said demodulated equalized signal and a signal representing said symbol value, and means (121, 122) for determining the value of said each coefficient associated with a subsequent one of said intervals by additively combining first and second updating terms therewith, the magnitudes of said first term being equal to a predetermined fraction of the product of a passband error signal formed during a predetermined one of said intervals with the sample with which said each coefficient was multiplied during said predetermined one of said intervals, the magnitude of said second term being a constant, and the sign of said second term being such as to drive the value of said each coefficient toward zero.

Claim 13

In a receiver to which a carrier signal, modulated by a succession of data symbols occurring at T second intervals, is transmitted over a band-limited channel, apparatus comprising sampling means (11, 12, 16) for forming a succession of samples of said signal at a predetermined rate and equalizer means operative during each one of a succession of T second intervals for forming a decision as to the value of an individual one of said symbols, said equalizer means comprising means (25, 26) for multiplying at least individual ones of a predetermined number of the most recently formed ones of said samples with individual ones of a plurality of coefficients, each coefficient having a predetermined value associated with said one of said intervals, means (27, 51) for demodulating the sum of the resulting products to form a demodulated equalized signal, means (37, 41, 42) responsive to said demodulated equalized signal for forming said decision and for forming a passband error signal, said passband error signal being equal to the difference, multiplied by said carrier signal, between said demodulated equalized signal and a signal representing said symbol value, and means (121, 122) for determining the value of said each coefficient associated with a subsequent one of said intervals by additively combining first and second updating terms therewith, the magnitudes of said first term being equal to a predetermined fraction of the product of a passband error signal formed during a predetermined one of said intervals with the sample with which said each coefficient was multiplied during said predetermined one of said intervals, the magnitude of said second term being a constant, and the sign of said second term being such as to drive the value of said each coefficient toward zero.

Claim 14

The invention of claim 10 wherein said predetermined rate is at least equal to (1+α)/T samples per second, where α=(2Tf co -1), f co being the highest spectral component about said carrier signal having at least a predetermined minimum energy, and wherein said signal representing said symbol value is derived from said decision.

Claim 15

The invention of claim 11 wherein the magnitude of said second term is subtracted from said each coefficient if the latter has a positive sign and is added to said each coefficient if the latter has a negative sign.

Claim 16

A data receiver for a received modulated data signal transmitted to said receiver over a bandlimited channel, said signal including a succession of data symbols transmitted at a symbol rate of 1/T symbols per second, said receiver comprising first means (11, 12, 16) for generating first and second time-spaced passband samples R j and R j ' of said received signal during the j th one of a succession of T second intervals, second means (25, 26, 34, 35) for forming the summation ##EQU6## where N is a selected integer and the C i (j)'s and C i '(j)'s are coefficients having at least first respective values, third means (27, 41, 42, 51) responsive to said summation Z j for forming a decision A j * as to the value of an individual one of said symbols, and fourth means (121, 122) for determining the values of the coefficients C i (j+1) and C i '(j+1) in accordance with C.sub.i (j+1)=C.sub.i (j)-αF(j)-αμsgn[C.sub.i (j)] C.sub.i '(j+1)=C.sub.i '(j)-αF'(j)-αμsgn[C.sub.i '(j)] where F(j) and F'(j) are predetermined correction terms, α and μ are predetermined constants, and the function sgn [x] has unity magnitude and has the sign of x.

Claim 17

A data receiver for a received modulated data signal transmitted to said receiver over a bandlimited channel, said signal including a succession of data symbols transmitted at a symbol rate of 1/T symbols per second, said receiver comprising first means (11, 12, 16) for generating first and second time-spaced passband samples R j and R j ' of said received signal during the j th one of a succession of T second intervals, second means (25, 26, 34, 35) for forming the summation ##EQU6## where N is a selected integer and the C i (j)'s and C i '(j)'s are coefficients having at least first respective values, third means (27, 41, 42, 51) responsive to said summation Z j for forming a decision A j * as to the value of an individual one of said symbols, and fourth means (121, 122) for determining the values of the coefficients C i (j+1) and C i '(j+1) in accordance with C.sub.i (j+1)=C.sub.i (j)-αF(j)-αμsgn[C.sub.i (j)] C.sub.i '(j+1)=C.sub.i '(j)-αF'(j)-αμsgn[C.sub.i '(j)] where F(j) and F'(j) are predetermined correction terms, α and μ are predetermined constants, and the function sgn [x] has unity magnitude and has the sign of x.

Claim 18

The invention of claim 13 wherein said third means includes means for forming a passband error signal E j corresponding to said decision A j * and wherein in said fourth means, F(j)=E j-1 R j-i-1 and F'(j)=E j-1 R j ' -i-1 .

Claim 19

The invention of claim 13 wherein said third means includes means (27, 51) for forming a signal A j which is a demodulated version of said summation Z j , and means (41, 42) for quantizing said signal A j to form said decision A j *.

Claim 20

The invention of claims 14 or 15 wherein said passband error signal forming means includes means (37, 51) for generating a signal equal to a remodulated version of the quantity (A j -A j *) to form said passband error signal E j .

Claim 21

The invention of claim 16 wherein modulated data signal is a complex data signal, wherein said sample generating means includes means (11) for forming first and second replicas of said received signal, said replicas being a Hilbert Transform pair, and means (12) for individually sampling said replicas to form first and second components of each of said samples R j and R j ', and wherein said signal A j is equal to Z j e -j (θ.sbsp.j.spsp.*) and said passband prior signal E j is equal to (A j -A j *)e +j (θ.sbsp.j.spsp.*), θ j * being an estimate of the current phase of the carrier of said data signal.

Claim 24

The invention of claim 18 wherein the magnitude of said second term is always non-zero.

Claim 25

The invention of claim 19 wherein the magnitude of said second term is a selected constant.

Claim 26

The invention of claim 20 wherein said predetermined rate exceeds 1/T samples per second.

Claim 27

The invention of claims 18, 20 or 21 wherein said equalized signal generating step includes the step of summing said products.

Claim 28

The invention of claim 20 wherein said data signal is a data signal modulated at a predetermined carrier frequency and wherein said equalized signal generating step includes the steps of summing said products to form a modulated equalized signal and demodulating said modulated equalized signal to form said equalized signal.

Claim 29

The invention of claim 23 wherein said error signal is equal to the difference, modulated at said carrier frequency, between the pre-quantized value of said equalized signal and said decision and wherein said determining step includes the step of forming said first term as a predetermined fraction of the product of said modulated difference with the sample with which said each one of said coefficients was multiplied in the formation of said decision.

Claim 30

The invention of claims 18, 20, 21 or 24 wherein said predetermined rate is at least (1+α)/T samples per second, where α is the fractional excess bandwidth.

Claim 35

The invention of claim 27 wherein said predetermined rate is at least equal to (1+α)/T samples per second, where α=(2Tf co -1), f co being the highest spectral component about said carrier signal having at least a predetermined minimum energy, and wherein said signal representing said symbol value is derived from said decision.

Claim 36

The invention of claim 28 wherein the magnitude of said second term is subtracted from said each coefficient if the latter has a positive sign and is added to said each coefficient if the latter has a negative sign.

Claim 39

The invention of claim 30 wherein said A j * forming step includes the step of forming a passband error signal E j corresponding to said decision A j * and wherein in said determining step F(j)=E j-1 R j-i-1 and F'(j)=E j-1 R j ' -i-1 .

Claim 40

The invention of claim 30 wherein said A j * forming step includes the steps of forming a signal A j which is a demodulated version of said summation Z j , and quantizing said signal A j to form said decision A j *.

Claim 41

The invention of claims 31 or 32 wherein said passband error signal forming step includes the step of generating a signal equal to a remodulated version of the quantity (A j -A j *) to form said passband error signal E j .

Claim 42

The invention of claim 33 wherein modulated data signal is a complex data signal, wherein said sample generating step includes the steps of forming first and second replicas of said received signal, said replicas being a Hilbert Transform pair, and individually sampling said replicas to form first and second components of each of said samples R j and R j ', and wherein said signal A j is equal to Z j e -j (θ.sbsp.j.spsp.*) and said passband prior signal E j is equal to (A j -A j *)e +j (θ.sbsp.j.spsp.*), θ j * being an estimate of the current phase of the carrier of said data signal.