COGNITIVE EQUALIZATION FOR HF CHANNELS

Abstract

In the High Frequency (HF) band, ranging from 3-30 MHz, long-range communications can be obtained by bouncing signals off the ionosphere without any significant infrastructure. However, the ionosphere changes rapidly, which can cause potentially harmful effects to the transmitted signal. This has motivated research into using adaptive equalization in this band to reverse these effects. However, a disadvantage of this technique is that based on the equalizer model and learning algorithm used, the error propagation may become significantly large, resulting in insufficient equalization to respond to these variations. To counter this, we investigate the usage of cognitive equalization, where an adaptive equalizer is equipped with the ability to change its structure (i.e. number of taps, step size, etc.) based on the current channel conditions and use probability of error to characterize its performance.