ABSTRACT
A Framework
for the Estimation of Time-Variant Channels
in OFDM
OFDM combines the advantages
of high achievable rates and relatively
easy implementation. However, for proper
recovery of the input, the OFDM receiver
needs accurate channel information. In this
talk, we propose an algorithm for channel
(and data) recovery in OFDM transmission
over time-variant environments.
The algorithm makes use of
the rich structure of the underlying communication
problem– a structure induced by the data
and channel constraints. These constraints
include pilots, the cyclic prefix, space-time
code, and the finite alphabet constraints
on the data. The constraints also include
sparsity, finite delay spread, and the statistical
((frequency and time correlation) and spatial
correlation of the channel. The algorithm
boils down to a forward-backward (FB) Kalman
filter. We also suggest a suboptimal modification
(essentially, a forward-only Kalman) that
is able to track the channel and recover
the data with no latency
We finally introduce two
recent extensions of our framework. We present
an implementation of the algorithm in the
frequency-domain that helps reduce computational
complexity. We also demonstrate how the
algorithm can be made robust against uncertainties
that result from imperfect channel estimates
and from Inter-carrier interference.