Performance of an echo canceller based on pseudo-noise training sequences

Different design issues and performance aspects of a lowcomplexity echo canceller for digital On-Channel Repeaters (OCRs) are described in this paper. Locally generated lowpower PN sequences are injected in the repeated signal to facilitate the estimation process of the coupling channel between the transmitting and the receiving antennas. In particular, we analyse a low-complexity channel estimation technique based on the autocorrelation properties of the training sequence. An expression for the cancelling performance is found analytically, in terms of Mean Rejection Ratio (MRR), and the impact of the filtering stages within the repeater is also addressed.