Complexity-constrained LS estimation for sparse systems

In an increasing number of applications the complexity of so-called fast RLS algorithms is prohibitive for real-time RLS estimation. One notable example is the adaptive equalization problem for intersymbol interference channels, where the number of equalizer taps grows linearly with the data rate. Due to the substantial memory in some ISI channels (for example the shallow-water acoustic telemetry channel) the uncoded data rate of real-time telemetry is limited not by the channel distortion or background noise, but by the complexity of the tap update algorithms