Comparing RLS and LMS adaptive equalizers for large hydrophone arrays in underwater acoustic communication channels

Abstract

There is an expected growth in underwater acoustic data traffic. To meet the demand of high data throughput, a migration towards the very high frequency band (>100 kHz) will be necessary. The increment in frequency comes with two problems; high propagation losses and rapid channel fluctuations. One solution addressing both of these obstacles is the use of large hydrophone arrays in combination with multi-channel equalizers. We will in this paper examine and compare tracking properties and convergence rate of the RLS and LMS adaptive algorithm for use in large multichannel equalizers. Using data from a very high frequency (250 kHz) channel we will show that in our channel, for large hydrophone arrays, RLS does not offers a significant advantage over LMS when it comes to tracking channel variations nor convergence rate.