The application of adaptive echo cancellation technique in analog full-duplex speech scrambling systems

The adaptive echo cancellation technique is used to deal with the echo problems associated with analog full-duplex speech scrambling systems. The influence of scrambling process on adaptive algorithms is analyzed through computer simulation. It is shown that the whitening effect caused by the scrambling process has little influence on least-squares (LS) algorithms but is especially beneficial to least-mean-square (LMS) algorithms. The authors thus conclude that it is very effective to use LMS algorithms in analog scrambling systems. Based on this conclusion, a high-performance adaptive echo canceller using a single TMS32010 processor has been implemented, the experimental results coincide with computer simulation.<<ETX>>