Performance enhancement of blind adaptive equalizers using environmental knowledge

In this paper, a multichannel equalizer that utilizes information of the surrounding environment is proposed. The structure of this receiver consists of a bank of prefilters matched to the deterministic part of the channel impulse response, whose outputs are fed to a conventional blind multichannel equalizer. If information on the source-receiver geometry as well as physical properties of the channel are available the unknown impulse responses can be estimated using a numerical propagation model. Assuming a Pekeris model for the physical channel and utilizing a ray propagation, model the receiver structure is tested with measured shallow water data gathered during the ROBLINKS 1999 sea trial. The obtained results demonstrate that incorporating environmental knowledge into the signal processing scheme leads to an improved performance compared to a conventional blind receiver.