Pulse or digital communications – Equalizers
Reexamination Certificate
1999-11-30
2003-10-14
Tran, Khai (Department: 2631)
Pulse or digital communications
Equalizers
C375S232000
Reexamination Certificate
active
06633606
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a signal processing apparatus for a digital television, and more particularly, to a blind channel equalizer for determining a blind coefficient of a channel equalizer. This application is based on Korean Application No. 99-10694 which is incorporated herein by reference.
2. Description of the Related Art
In a digital television using ground waves, broadcasting signals are transmitted from a transmitter to a receiver. The transmitted signals travel from the transmitter to the receiver along different routes, and are time-delayed by obstacles along the different routes. Accordingly, even if the same signals are transmitted from the transmitter, the signals arrive at the receiver at different times. For that reason, channel equalization is required. Channel equalization is a method of summing up all received signals and recovering original signals in the receiver.
A digital television employing the same broadcasting mode as the existing National Television System Committee (NTSC) analog television may be interfered with by an NTSC signal. To prevent deterioration of reception performance resulting from such interference, a digital television receiver includes an NTSC rejection filter (NRF).
Generally, a blind error used in a self-recovering channel equalizer can be calculated as follows.
e
blind
(
i
)=&ggr;sgn(
y
)−
y
(
i
) (1)
where a blind error at a particular time i is equal to a value obtained by subtracting an output y(i) of a feed forward filter
100
described below at the particular time i from a value obtained by multiplying a blind decision coefficient y by a sign function sgn(y). At this time, the blind decision coefficient y with respect to a transmission symbol string {x} can be calculated as follows.
γ
=
E
[
x
2
]
E
[
&LeftBracketingBar;
x
&RightBracketingBar;
]
(
2
)
The blind decision coefficient is 5.25 at an 8-vestigial sideband (VSB), 8 at a 15-VSB, and 10.625 at a 16-VSB on the basis of a level value.
For blind channel equalization, the transmission symbol string should have a constant value or have a value in a sub-Gaussian distribution. In the case of the 15-VSB which has passed through a comb filter, outputs of a channel equalizer have a triangular distribution, theoretically limiting the application of blind channel equalization. According to conventional technology, an output symbol string is approximated to a binary sequence of +&ggr; and −&ggr; to calculate an error signal. When symbol values converge on a level value 0 as in the 15-VSB, a blind error is measured using inverse transformation and then the measured blind error is forward transformed to obtain a desired value.
FIG. 1
is a block diagram of a conventional blind channel equalizer. A NRF output signal which has passed through the feed forward filter
100
is delayed by 12 symbols in an inverse transformer
110
before being output. A blind decision coefficient allocator
120
allocates a blind decision coefficient +&ggr; or −&ggr; to a resultant of the inverse transformation. A forward transformer
130
forward-transforms a blind decision coefficient signal. The forward transformer
130
outputs a blind decision coefficient corresponding to a 15-VSB level. A subtracter
140
performs subtraction with respect to the blind decision coefficient from the forward transformer
130
and the signal from the feed forward filter
100
to obtain a blind error value.
In such method, a transmission symbol string is inverse-transformed and its distribution is made uniform. Thereafter, a blind decision coefficient is obtained and passed through a partial response system, thereby obtaining a blind error in the form of a partial response. When this blind error is applied to the 15-VSB passed through the comb filter, the blind decision coefficient becomes either +10.5 or −10.5.
The blind error can be obtained in a usual partial response system. However, in the 15-VSB introduced to remove NTSC interference signals, NTSC components are restored during the inverse transformation and cause errors during the process of measuring the blind error. Specifically, the blind error is determined in accordance with a sign of a receiving symbol. When inverse transformation is performed in the 15-VSB, the NTSC components are restored and cause the sign of the receiving symbol to be disturbed. This phenomenon appears particularly when the level value of a symbol is small.
SUMMARY OF THE INVENTION
It is an object of the present invention to provide a blind channel equalizer for allocating a blind decision coefficient to improve the performance of a self-recovering channel equalizer for a system having a partial response.
To achieve the object of the present invention, there is provided a blind channel equalizer in a vestigial sideband (VSB) receiving system. The blind channel equalizer includes: a comparator for determining a boundary level value according to the distribution of occurrence probabilities of VSB symbol levels and comparing the determined boundary level value with a received symbol; a blind decision coefficient allocator for allocating one of three blind coefficients according to the output result of the comparator; and a subtracter for performing subtraction with respect to the received symbol and an output of the blind decision coefficient allocator to provide a blind error.
REFERENCES:
patent: 5537439 (1996-07-01), Choi
patent: 6418164 (2002-07-01), Endres et al.
Samsung Electronics Co,. Ltd.
Sughrue & Mion, PLLC
LandOfFree
Blind channel equalizer does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Blind channel equalizer, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Blind channel equalizer will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3161520