Reduced complexity equalizer for multi mode signaling

A transmission channel equalizer comprising:

• a sparse digital filter having N tap positions and M coefficients, where N and M are integers and N is greater than M which is greater than 2;
• means for adaptively selecting M of the N tap positions and for assigning the M coefficients to the selected M tap positions;
• and
• means for adaptively updating the selected M coefficients both in value and in correspondence to tap positions in the sparse digital filter;
• wherein the means for adaptively selecting M of the N tap positions includes:
• means for identifying a plurality of tap positions having corresponding coefficient values that are less than a threshold value, ε, according to an l1-norm measure of the coefficient values;
• means for comparing the coefficients of each of the identified tap positions to a further threshold value δ, according to the l1-norm measure of the coefficients;
• means for deleting, from the M tap positions, all of the tap positions corresponding to the coefficients which have a value less than δ;
• and
• means for selecting one tap position for each deleted tap position to provide M selected tap positions.

A transmission channel equalizer comprising:

• a sparse digital filter having N tap positions and M coefficients, where N and M are integers and N is greater than M which is greater than 2;
• means for adaptively selecting M of the N tap positions and for assigning the M coefficients to the selected M tap positions;
• and
• means for adaptively updating the selected M coefficients both in value and in correspondence to tap positions in the sparse digital filter;
• wherein the means for adaptively selecting M of the N tap positions includes:
• means for identifying one of the M tap positions which has a corresponding coefficient value that is less than any other coefficient value;
• means for deleting the identified tap position from the M tap positions;
• and
• means for selecting one tap position of the N tap positions which is not associated with a coefficient value and for adding the selected tap position to the M tap positions.

A transmission channel equalizer comprising:

• a sparse digital filter having N tap positions and M coefficients, where N and M are integers and N is greater than M which is greater than 2;
• means for adaptively selecting M of the N tap positions and for assigning the M coefficients to the selected M tap positions;
• and
• means for adaptively updating the selected M coefficients both in value and in correspondence to tap positions in the sparse digital filter;
• wherein the means for adaptively selecting M of the N tap positions includes:
• means for identifying a plurality of the M tap positions having corresponding coefficient values that are less than the corresponding coefficient values of other ones of the taps in accordance with a predetermined measure;
• means for deleting the identified plurality of tap positions from the M tap positions in accordance with a predetermined measure;
• and
• means for selecting a plurality of tap positions from the N tap positions, which selected plurality of tap positions are not associated with coefficient values and for adding the selected tap position to the M tap positions.

A transmission channel equalizer comprising:

• a sparse digital filter having N tap positions and M coefficients, where N and M are integers and N is greater than M which is greater than 2;
• means for adaptively selecting M of the N tap positions and for assigning the M coefficients to the selected M tap positions;
• and
• means for adaptively updating the selected M coefficients both in value and in correspondence to tap positions in the sparse digital filter;
• where the sparse digital filter includes:
• a first finite impulse response (FIR) filter which processes modulated channel impaired digital samples to produce filtered modulated channel impaired digital samples;
• and
• an infinite impulse response (IIR) filter which processes the filtered modulated channel impaired digital samples, to produce output equalized digital samples, the IIR filter including:
• summing means which adds further filtered digital samples to the filtered modulated digital samples to produce the output equalized digital samples;
• and
• a second FIR filter which processes the output equalized digital samples to produce the further filtered digital samples.

A transmission channel equalizer comprising:

• a sparse digital filter having N tap positions and M coefficients, where N and M are integers and N is greater than M which is greater than 2;
• means for adaptively selecting M of the N tap positions and for assigning the M coefficients to the selected M tap positions;
• and
• means for adaptively updating the selected M coefficients both in value and in correspondence to tap positions in the sparse digital filter;
• where the sparse digital filter includes:
• a first finite impulse response (FIR) filter which processes modulated channel impaired digital samples to produce filtered modulated channel impaired digital samples;
• a derotator which mixes the filtered modulated channel impaired digital samples with an estimated carrier signal to produce derotated channel impaired digital samples;
• an infinite impulse response (IIR) filter which processes the derotated channel impaired digital samples, to produce output equalized digital samples, the IIR filter including:
• summing means which adds filtered derotated digital samples to the derotated channel impaired digital samples to produce the output equalized digital samples;
• and
• a second FIR filter which processes the output equalized digital samples to produce the filtered derotated digital samples.

A transmission channel equalizer comprising:

• a sparse digital filter having N tap positions and M coefficients, where N and M are integers and N is greater than M which is greater than 2;
• means for adaptively selecting M of the N tap positions and for assigning the M coefficients to the selected M tap positions;
• and
• means for adaptively updating the selected M coefficients both in value and in correspondence to tap positions in the sparse digital filter;
• where the sparse digital filter includes:
• a first finite impulse response (FIR) filter which processes modulated digital samples to produce filtered modulated digital samples;
• and
• an infinite impulse response (IIR) filter which processes the filtered modulated digital samples, the IIR filter including:
• summing means which adds filtered rerotated digital symbols to the filtered modulated digital samples to produce equalized digital samples;
• and
• a derotator which mixes the equalized digital samples with an estimated carrier signal to produce derotated digital samples;
• a slicer which quantizes the derotated digital samples according to a predetermined alphabet to produce digital symbols;
• a rerotator which mixes the digital symbols with the estimated carrier signal to produce rerotated digital symbols;
• and
• a second FIR filter which processes the rerotated digital symbols to produce the filtered rerotated digital symbols.

A transmission channel equalizer comprising:

• a sparse digital filter having N tap positions and M coefficients, where N and M are integers and N is greater than M which is greater than 2;
• means for adaptively selecting M of the N tap positions and for assigning the M coefficients to the selected M tap positions;
• and
• means for adaptively updating the selected M coefficients both in value and in correspondence to tap positions in the sparse digital filter;
• where the sparse digital filter includes:
• a first finite impulse response (FIR) filter which processes modulated digital samples to produce filtered modulated digital samples;
• a derotator which mixes the filtered modulated digital samples with an estimated carrier signal to produce derotated digital samples;
• an infinite impulse response (IIR) filter which processes the derotated digital samples, the IIR filter including:
• summing means which adds filtered symbols to the derotated digital samples to produce equalized digital samples;
• and
• a slicer which quantizes the equalized digital samples according to a predetermined alphabet to produce digital symbols;
• a second FIR filter which processes the digital symbols to produce the filtered digital symbols.

A transmission channel equalizer for equalizing a vestigial sideband modulated (VSB) signal, the equalizer comprising:

• a first finite impulse response (FIR) filter which processes modulated digital samples to produce filtered modulated digital samples;
• means for calculating a single axis constant modulus algorithm (SACMA) error term for the filtered modulated digital samples;
• an infinite impulse response (IIR) filter which processes the filtered modulated digital samples, the IIR filter including:
• summing means which adds further filtered digital samples to the filtered modulated digital symbols to produce output equalized digital samples;
• a second FIR filter which processes the equalized digital samples to produce the further filtered digital samples;
• and
• means for applying the calculated SACMA error term to update coefficient values for both the FIR filter and the IIR filter.

A transmission channel equalizer for equalizing a vestigial sideband modulated (VSB) signal, the equalizer comprising:

• a first finite impulse response (FIR) filter which processes modulated digital samples to produce filtered modulated digital samples;
• a derotator which mixes the filtered modulated digital samples with an estimated carrier signal to produce derotated digital samples;
• means for calculating a baseband single axis constant modulus algorithm (SACMA) error term for the derotated digital samples;
• a rerotator which mixes the SACMA error term with the estimated carrier signal to produce a passband SACMA error term;
• an infinite impulse response (IIR) filter which processes the derotated digital samples, the IIR filter including:
• summing means which adds filtered derotated digital samples to the derotated digital samples to produce output equalized digital samples;
• and
• a second FIR filter which processes the output equalized digital samples to produce the filtered derotated digital samples;
• means for calculating a baseband single axis constant modulus algorithm (SACMA) error term for the output equalized digital samples;
• a rerotator which mixes the SACMA error term with the estimated carrier signal to produce a passband SACMA error term;
• and
• means for applying the baseband SACMA error term to update coefficient values for the IIR filter and for applying the passband SACMA error term to update coefficient values for the FIR filter.

A transmission channel equalizer which equalizes vestigial sideband (VSB) or quadrature amplitude modulated (QAM) signals comprising:

• a first finite impulse response (FIR) filter, having a plurality of coefficients, which processes modulated digital samples to produce filtered digital samples;
• an infinite impulse response (IIR) filter which processes the filtered digital samples, the IIR filter including:
• summing means which adds further filtered digital samples to the filtered digital samples to produce output equalized digital samples;
• and
• a second FIR filter, having a second plurality of coefficients, which processes the output equalized digital samples to produce the further filtered digital samples;
• and
• means for updating the coefficients of the first and second FIR filters comprising:
• acquisition means for calculating a constant modulus algorithm (CMA) error term for QAM signals and a single axis constant modulus algorithm (SACMA) error term for VSB signals;
• tracking means for calculating a decision directed (DD) error term;
• and
• means for applying the CMA error term or the SACMA error term to both the first and second FIR filters and then applying the DD error term to the first and second FIR filters.

A method for adaptively updating a transmission channel equalizer having a sparse digital filter which has N tap positions and M coefficients where N and M are integers and N is greater than M which is greater than 2, the method comprising the steps of:

• identifying a plurality of tap positions having corresponding coefficient values that are less than a threshold value, ε;
• comparing each of the identified plurality of tap positions to a further threshold value δ;
• deleting all of the tap positions having a value less than δ from the M tap positions;
• selecting one tap position for each deleted tap position to provide M selected tap positions;
• and
• adaptively updating the selected M coefficients in value responsive to an objective function to cause the filter to produce an equalized output signal.

A method for adaptively updating a transmission channel equalizer having a sparse digital filter which has N tap positions and M coefficients where N and M are integers and N is greater than M which is greater than 2, the method comprising the steps of:

• identifying one of the tap positions having a corresponding coefficient value which is less than any other coefficient value;
• deleting the identified tap position from the M tap positions;
• selecting one of the tap positions to replace the deleted tap position to create a new set of M selected tap positions;
• and
• adaptively updating the M coefficients corresponding to the M selected tap positions in value responsive to an objective function to cause the filter to produce an equalized output signal.

A method for equalizing received signals which are either vestigial sideband (VSB) or quadrature amplitude modulated (QAM) signals comprising the steps of:

• processing the received signals to produce modulated digital symbols;
• forward filtering the modulated digital symbols according to a first frequency response characteristic to produce filtered modulated digital symbols;
• feedback filtering the filtered modulated digital symbols according to a second frequency response characteristic to produce feedback filtered digital symbols;
• summing the feedback filtered digital symbols and the filtered modulated digital symbols to produce output equalized digital symbols;
• and
• adaptively updating the first and second frequency response characteristics comprising the steps of:
• calculating a constant modulus algorithm (CMA) error term for QAM signals and a single axis constant modulus algorithm (SACMA) error term for VSB signals;
• calculating a decision directed (DD) error term;
• and
• applying one of the CMA error term and the SACMA error term to adapt both the first and second frequency response characteristics and then applying the DD error term to adapt both the first and second frequency response characteristics.