Narrow-band interference excision in spread-spectrum systems usingself-orthogonalizing transform-domain adaptive filters

摘要

Among many transform-domain interference excision techniques,ntransform-domain adaptive filtering has many advantages. It is based onna true optimization of some particular performance parameters such asnthe bit-error rate (BER). Moreover, it is insensitive to jammernfrequency. However, transform-domain adaptive filtering also has thendrawback of being incapable of tracking a rapidly changing interferencenbecause most adaptive algorithms require time to converge to the optimalnsolution. In this paper, a self-orthogonalizing transform-domain leastnmean square (SO-TRLMS) algorithm is used to speed up the convergence.nCompared to a traditional transform-domain least mean square (TRLMS)nalgorithm, the SO-TRLMS algorithm can significantly improve thenconvergence rate of the LMS algorithm, thus making the transform-domainnadaptive filtering technique more suitable for real-time processing. Innorder to show how the system performance is affected by various factorsnsuch as interference power and the transform used, this paper presentsnan analytical result for the BER performance that is applicable fornarbitrary orthogonal linear transforms. Simulation results are alsonpresented to demonstrate the validity of the analysis