 |
||||
Blind and semi-blind channel estimation and equalization
What is Blind Channel Estimation and Equalization?
In a communication system the transmitted information signal is distorted
by the propagation medium (channel) and corrupted by the additive noise.
Therefore, the receiver has to compensate for these effects in order to
ensure a correct reception. An equalizer is designed to get rid of the
received signal distortions, by trying 'to invert' the channel effect.
The equalizer has to know the transmitted sequence, or at least a part
of it, in order to estimate the channel impulse response and compensate
its effect in the transmitted data signal.
A new class of equalizers have been the subject of research in last
decades. They estimate the channel by using only the received data and
some characteristics of the transmitted signal (statistics, finite alphabet,
constant envelope etc. ) These techniques constitute so-called blind equalization
or self recovering equalization.
Since the second order statistics of the received symbol rate sampled
signal do not have any information about the phase of the channel impulse
response they cannot be used in blind identification. Therefore higher
than second order statistics have to be used. The blind equalization algorithms
based on higher order statistics (HOS) are impractical in many applications
such wireless communications due to the long convergence time.
If the received signal is sampled at a rate higher than the symbol
rate the resultant sequence is cyclostationary and the second order cyclostationary
(SOCS) statistics are sufficient for blind channel identification. The
blind equalization algorithms based on SOCS have a higher faster rate but
they have to fulfill some conditions (ex. no common zeros among the fractionally
spaced subchannels). These conditions became unpractical in applications
where the channel is time-variant.
What is Semi-Blind Channel Estimation and Equalization?
If training sequence is available at the receiver the quality of the
reception can be increase by using hybrid semi-blind equalization
algorithms which take advantage of the both blind and data aided
equalization schemes.
In future wireless communications the higher data rates require a very
good channel estimation and tracking. Usually blind techniques are not
capable to track fast variations of the wireless channels therefore more
powerful channel estimation algorithms are required (Kalman filtering,
RLS etc). These algorithms may be used in a decision directed (DD) scheme
where they use the decisions of an equalizer (DFE, Viterbi etc.) for channel
estimation and tracking.
Helsinki University of Technology