Adaptive channel equalisation with the multirate optimising order statistics equaliser (MOOSE)

In this paper a number of filters are applied to the problem of equalisation in impulsive noise. The multirate optimising order statistics equaliser (MOOSE) is shown to be far superior to FIR and L1-filters in comparisons of bit error rate performance over a range of additive noise types and signal to noise ratios. MOOSE is also shown to be significantly faster, and much smaller in its number of parameters and robust with respect to sudden variations in the channel statistics.