My research interest is in signal processing for digital communications.
Starting from blind equalization and sources separation to mitigate
multipath propagation in hostile environments, I enlarged the transmission
problems I considered at the same time I learned about communication systems.
I also considered multi-user systems based on spread-spectrum
such as HS-DPA or GPS in multipath propagation, multicarrier systems such as
UWB based on multi-band OFDM, and OFDMA for LTE.
Thanks to Alain Glavieux, I got interested in turbo-equalization,
i.e. joint equalization and decoding using soft information passing in an
iterative processing.
We proposed an original analysis of the turbo-equalizer and proposed joint channel estimation
inside the turbo-equalizer.
The application to Hybrid-ARQ (retransmission strategy) has also been developped.
This first joint approach led me to consider the transmission more globally
in a cross-layer fashion.
Therefore, I started to deal with resources allocation (of subcarriers, bands, power, modulation order, code rate...)
that can be performed from the study of the system performance.
The originality of the approach I considered is to consider a very limited
channel state information based on the mean channel gain over some band
in order to have a low feedback throughput.
Considering the entire transmission chain, it also important to deal with the
effect of the RF frontal which is never totally ideal.
Non-linearities, frequency selectivity, mis-synchronisation (IQ-imbalance or
carrier frequency offset) impact heavily on the system performance.
We analyze these "dirty-RF" impairments and propose numerical processing
to mitigate their effects.