Pulse or digital communications – Receivers – Interference or noise reduction
Reexamination Certificate
2000-07-24
2002-07-16
Ghebretinsae, Temesghen (Department: 2631)
Pulse or digital communications
Receivers
Interference or noise reduction
C375S316000, C348S614000
Reexamination Certificate
active
06421403
ABSTRACT:
The invention relates to digital filters used for channel equalization and cancellation of multipath distortion in digital communications radio receivers, such as those employed in digital television receivers.
BACKGROUND OF THE INVENTION
A known configuration of channel equalizer employs a precursor finite-impulse-response (FIR) digital filter followed by a postcursor infinite-impulse-response (IIR) digital filter. The postcursor filter comprises a digital subtractor receiving the IIR precursor filter response as minuend input signal, a quantizer for quantizing the difference output signal from the subtractor, and a feedback FIR digital filter responding to the quantizer output signal for supplying subtrahend input signal to the subtractor. The postcursor filter suppresses post-ghost signals arriving after the principal signal. The precursor filter is commonly called a “feed-forward FIR filter” to distinguish it from the feedback FIR filter in the postcursor filter. The feed-forward FIR filter combines match filtering to reduce intersymbol interference, filtering to suppress pre-ghost signals arriving before the principal signal, and filtering to suppress artifacts otherwise arising in the postcursor filter. Clocking of the digital filters in the channel equalizer must at a rate at least as high as symbol rate in order to satisfy the well-known Nyquist criterion for pulse reproduction without irreparable intersymbol interference (ISI) being introduced.
In a process known as “synchronous equalization” the received signal is subjected to various delays that are multiples of the symbol interval and is summed with the delayed signals in a weighted summation procedure to suppress multipath distortion. Synchronous equalization has been employed in adaptive channel equalizers in which the feed-forward and feedback FIR filters are clocked at symbol rate. In such adaptive channel equalizers the coefficients of the feed-forward and feedback FIR filters are adjusted during operation by a process known as “decision feedback”. Error signal for the decision feedback method is generated by comparing the output signal from the quantizer with its input signal, both signals being clocked at symbol rate.
In a process known as “fractional equalization” the received signal is subjected to various delays that are multiples of a specified fraction of the symbol interval and is summed with the delayed signals in a weighted summation procedure to suppress multipath distortion. Equalization at band edges is known to be much improved in a channel equalizer clocked at twice symbol rate, in which channel equalizer the received signal is subjected to various delays that are multiples of one-half of one symbol epoch. It has been observed that substantially the same degree of improvement can be obtained with a channel equalizer filter with substantially fewer taps, which filter is clocked at three-halves symbol rate. In such channel equalizer the received signal is subjected to various delays that are multiples of two-thirds of one symbol epoch.
In over-the-air digital television, transmission channel characterization is subject to considerable change with time and adaptive coefficient equalization is a practical necessity for a DTV receiver to be commercially acceptable. There is a desire to employ decision feedback techniques for adjusting the coefficients in the feed-forward and feedback FIR filters in order to track changing multipath conditions. Fractional equalization is preferred in the adaptive channel equalizer, so there is less criticality as to the timing of sampling in the component filters. Decision feedback techniques for adjusting the coefficients of a fractional equalizer properly are described by A. L. R. Limberg and C. B. Patel in U.S. patent application ser. No. 09/373,588 filed Aug. 13, 1999 and titled “ADAPTIVE FRACTIONALLY SPACED EQUALIZER FOR RECEIVED RADIO TRANSMISSIONS WITH DIGITAL CONTENT, SUCH AS DTV SIGNALS”, claiming priority from a similarly titled provisional U.S. patent application ser. No. 60/097,614 filed Aug. 24, 1998.
U.S. Pat. No. 5,479,449 titled “DIGITAL VSB DETECTOR WITH BANDPASS PHASE TRACKER, AS FOR INCLUSION IN AN HDTV RECEIVER”, which issued Dec. 26, 1995 to A. L. R. Limberg and C. B. Patel, describes the demodulation of digital television (DTV) signals reposing in an intermediate-frequency (I-F) band offset from zero frequency by no more than a few megahertz. These intermediate-frequency DTV signals are digitized and are then synchrodyned to baseband in the digital regime to recover in-phase and quadrature-phase baseband signals. The in-phase baseband signal contains symbol code that is symbol decoded, error-corrected, and de-randomized in successive signal processing steps. The quadrature-phase baseband signal is lowpass filtered to generate automatic frequency and phase control (AFPC) signal for a local oscillator that is used in the downconversion of the DTV signals to the I-F band offset from zero frequency by no more than a few megahertz.
Passband equalization done on orthogonal components of a digitized I-F signal before digital demodulation to baseband is preferred when the received signal has both upper and lower sideband components. The vestigial sideband (VSB)signals proposed for DTV broadcasting have essentially no lower sideband components, however, so baseband equalization has been used. The customary practice in DTV receiver designs that use baseband equalization has been to equalize just the in-phase baseband signal. The number of digital multipliers that baseband equalization uses for applying weighting coefficients to the kernel taps of the equalization filter is then half the number that would be used in equivalent passband equalization of the DTV signals.
It is here pointed out that the bandpass tracker type of demodulator for DTV signals shares a problem with any other synchronous demodulation scheme in which a local oscillator used in downconversion has automatic frequency and phase control (AFPC) of its local oscillations in which AFPC signal is developed by lowpass filtering the quadrature-phase baseband signal. AFPC seeks to adjust the carrier phasing of the synchrodyne to baseband to minimize the direct component of the quadrature baseband signal. If the in-phase baseband signal is equalized to suppress ghosts, but the quadrature-phase baseband signal is not, the presence of a ghost of appreciable strength will perturb the phase of the AFPC'd local oscillator from the correct phasing for an unghosted quadrature-phase baseband signal. This means that the phase of the AFPC'd local oscillator will not be the correct phasing for an unghosted in-phase baseband signal either. This results in a lower amplitude in-phase baseband signal than would be recovered were the phase of the AFPC'd local oscillator correct. When the multipath conditions are static, the equalizer corrects the amplitude of this lower-amplitude in-phase baseband signal. This correction introduces error in the equalizer response to the principal signal vis-a-vis the response to the static ghosts, which compensates for the phase error in the tracking of the in-phase and quadrature-phase baseband signals. When the multipath conditions change, the adaptation of the equalizer coefficients is generally slow in responding to the change, so the error in the equalizer response to the principal signal vis-a-vis the response to the static ghosts tends to persist. However, the phase error in the tracking of the in-phase and quadrature-phase baseband signals changes immediately as the multipath conditions change. Accordingly, the suppression of static ghosts by the equalizer is affected by dynamic ghosts, which poses a particularly serious problem during data slicing if there is a strong static ghost.
Suppose the in-phase and quadrature-phase baseband signals are each subjected to similar equalization filtering to suppress ghosts. Then, when the multipath conditions are static, the AFPC loop adjusts the phase of the AFPC'd local oscillator to be correct for both the in
Ghebretinsae Temesghen
Samsung Electronics Co,. Ltd.
LandOfFree
Baseband equalization of quadrature-phase as well as... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Baseband equalization of quadrature-phase as well as..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Baseband equalization of quadrature-phase as well as... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2818663