Robust Decision Feedback Equalizers
Mikael Sternad ,
Anders Ahlén and
IEEE International Conference on Acoustics,
Speech and Signal Processing,
Minneapolis, MN, vol III, pp 555-558, April 27-30, 1993.
© 1993 IEEE.
Paper In Pdf (with figures) 291K.
In Postscript, without figures :
compressed(gz) 32K ,
When digital data are transmitted over dispersive radio channels,
a Decision Feedback Equalizer (DFE) can be utilized to
estimate transmitted symbols.
The paper presents a model-bases methodology for
robust design, aimed at reducing the effect of model uncertainties and
of decision errors in the feedback loop.
Design equations are presented for robust and realizable
decision feedback equalizers, based on uncertain models of
IIR channels with coloured noise. Given a probabilistic measure
of model uncertainty, the MSE, averaged over a whole
class of possible channel models and noise models, is minimized.
A second type of robustification,
which reduces the error propagation due to the feedback,
is also introduced: the design is optimized under the assumption
that previous decisions are affected by white noise.
When the variance of this noise is increased, the gain of
the DFE feedback filter is reduced.
In the limit, we obtain a recursive linear equalizer
which minimizes the model-averaged MSE.
By treating the decisioned data as uncertain,
the length of error bursts due to feedback is reduced.
Simpler coding schemes, which use less interleaving, can then be used.
For some channels, the probability of decision errors can
also be reduced significantly by selecting a nonzero noise variance.
The resulting design equations define a large
class of equalizers, with DFE's and linear equalizers
based on nominal models being special cases.
- Related publications:
PhD Thesis by Erik Lindskog, which explains "self-robustification"
by Erik Lindskog, May 1995, which includes the paper.
Robustification against time-variations
within a GSM TDMA frame.
Robust filter design
and uncertainty modelling, as described in Automatica 1993.
by Kenth Öhrn May 1996, describing robust filter
design in detail.
Personal use of this material is permitted.
However, permission to reprint/republish this material for
advertising or promotional purposes or for creating new
collective works for resale or redistribution to servers or lists,
or to reuse any copyrighted component of this work in other works
must be obtained from the IEEE.