Pulse or digital communications – Equalizers – Automatic
Reexamination Certificate
1998-11-03
2001-10-09
Pham, Chi (Department: 2731)
Pulse or digital communications
Equalizers
Automatic
C375S321000, C348S404100, C708S322000
Reexamination Certificate
active
06301298
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates generally to adaptive equalizers and more particularly to an adaptive equalizer and method of operation that regenerates ideal real and imaginary component values from a complex valued digital signal for producing a complex valued error signal.
Data-bearing signals are often distorted by the transmission path, or channel, which they traverse between the transmitter and receiver. The distortion may be caused by the physical channel medium, such as free space, cable and the like, or nonlinear behavior of the transmitter and receiver. When such distortion consists of frequency-dependent magnitude and phase deviations of the signal, it can be corrected at the receiver by filtering the received signal with a linear filter which has inverse magnitude and phase response with respect to that of the channel. The inverse filter is called an equalizing filter, and the process of determining the filter coefficients and performing the filtering is called equalization.
The Stochastic Gradient (SG) algorithm is an efficient method to derive the value of the equalizing filter coefficients. The SG algorithm requires that a reference waveform of ideal (undistorted) data symbols be available at the receiver. When the transmitted signal contains random data unknown to the receiver, the ideal waveform can be estimated by regenerating it from the received and possibly distorted symbols. This operation can be implemented by “slicing” the received symbols to the closest ideal symbol points or levels. As long as relatively few error are made in the slicing process, the SG algorithm will converge toward the correct equalization filter.
FIG. 1
show a representative adaptive equalizer
10
for digital signals that is known in the art and generally includes a multi-tap digital filter
12
through which an unequalized digital signal passes. A slicer
14
continuously processes the equalized digital signal from the digital filter and generates ideal amplitude values of the digital levels in the signal. The ideal amplitude values are processed with the corresponding digital levels of the equalized digital signal to generate an error signal. A summation circuit
16
providing a difference output is an example of a circuit for producing the error signal. A correlator
18
receives the error signal and the unequalized digital signal. The correlator
18
performs a series of multiplications of the error signal with the unequalized digital signal, to produce a scaled output. The scaled correlator
18
output is combined with the old tap values
22
of the digital filter
12
in adder
20
to develop new tap values for the filter
12
. For linear impairments in the digital signal, the circuit nullifies the effects of such impairments in the output signal.
Adaptive equalizers are employed in various configurations depending to the type of digital transmission system. In a quadrature amplitude modulation (QAM) system, for example, both the I (real) and Q (imaginary) channels are independent and data bearing. Slicers are provided for both channels for generating the ideal amplitude values for the real and imaginary data. In the advanced television (ATV) system, implemented using an 8-VSB digital transmission system, complex digital symbols (I+jQ) are generated with the in-phase or data bearing component of the signal contained in the real part of the complex signal. The imaginary part of the complex signal is used for suppressing a portion of the signal spectrum for more efficient bandwidth use. Ideal “I” or real symbols can be obtained by slicing the received “I” symbols to ideal reference levels. However, in 8-VSB, in contrast to signals such as QAM, the Q-channel signal is a function of the I-channel data symbols, and has a nearly continuous range of levels at symbol instances. Therefore, it is not possible to simply slice the received Q levels to ideal levels.
What is needed is an adaptive equalizer for equalizing complex digital signal samples where the Q-channel or imaginary component of the complex signal is a function of the I-channel or real component of the complex signal. The adaptive equalizer needs to regenerate ideal imaginary component levels from the sliced ideal real component levels of the complex signal to generate a complex error signal for driving the convergence of the filter coefficients of the equalizing filter. The adaptive equalizer needs to use the distortion information available in the Q-channel signal for improving the convergence of the equalizer process and the approximation of the inverse channel equalizing filter.
SUMMARY OF THE INVENTION
Accordingly, an object of the present invention is an adaptive equalizer that regenerates an ideal reference Q-channel signal from an ideal reference I-channel signal.
Another object of the present invention is an adaptive equalizer that generates a complex error signal for driving the convergence of the equalizer process.
A further object of the present invention is generating a complex valued reference signal from the complex valued equalized signal samples.
The adaptive equalizer with Q-channel regeneration of the present invention includes a filter system receiving complex valued unequalized signal samples and generating complex valued equalized signal samples. A slicer receives the real component values of the complex valued equalized signal samples and generates ideal real component values. A complex valued error signal generator receives the ideal real component values and the complex valued equalized signal samples and produces a complex valued error signal. The real component values of the error signal are derived from the real component values of the complex valued equalized signal samples and the ideal real component values. The imaginary component values of the complex valued error signal are derived from the imaginary component values of the complex valued equalized signal samples and ideal imaginary component values regenerated from the ideal real component values. The complex valued error signal is combined with time aligned vectors of complex valued unequalized signal samples in the filter system for updating filter coefficient values of the filter system.
The filter system includes a equalization filter having filter coefficient values updatable from initial values that generates the complex valued equalized signal samples from the received complex valued unequalized signal samples. The complex valued unequalized signal samples are also applied to a delay that generates delayed vectors of complex valued unequalized signal samples time aligned with the complex valued error signal values from the complex valued error signal generator. A multiplier receives the time aligned vectors and the complex valued error signal and generates equalization filter coefficient correction values. The coefficient correction values are applied to an adder that receives the current equalization filter coefficient values for generating updated equalization filter coefficient values.
The complex valued error signal generation has a regeneration filter that receives the ideal real component values from the slicer and generates ideal imaginary component values. In the preferred embodiment of the present invention, the regeneration filter is a FIR filter having filter coefficients that produce imaginary component values corresponding to a raised cosine filter. A first delay receives the ideal real component values and generates ideal real component values that are time aligned with the generated ideal imaginary component values. A second delay receives the complex valued equalized signal samples and generates complex valued equalized signal samples that are time aligned with the ideal real and imaginary component values. A combiner receives the time aligned complex valued equalized signal samples and the ideal real and imaginary component values and generates the complex valued error signal from the difference between the time aligned complex valued equalized signal samples and the ideal real an
Deshpande Nikhil
Kuntz Thomas L.
Bucher William K.
Pham Chi
Tektronix Inc.
Tran Khai
LandOfFree
Adaptive equalizer with complex signal regeneration and... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Adaptive equalizer with complex signal regeneration and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Adaptive equalizer with complex signal regeneration and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2578197