Interference suppression in spread spectrum code division multiple access communications

摘要

This dissertation covers three different research topics, all falling under the umbrella of code division multiple access (CDMA) communications systems, specifically direct-sequence spread spectrum. The first two topics concern narrowband interference suppression for radio frequency (RF) systems employing spread spectrum techniques to overlay mobile communications signals on pre-existing narrowband systems. The third topic deals with the effect of phase noise on the performance of a CDMA system, an issue of concern in coherent optical systems. The RF work applies to mobile radio and personal communications services. The optical results are of interest in fiber optic as well as indoor wireless communications, e.g., wireless local area networks.;The first section of this thesis extends and improves recent advances in the use of nonlinear filtering methods to suppress narrowband interference. Interpolating filter structures are derived and previous results extended to the case of multiple spread spectrum users. Simulations demonstrate how nonlinear filtering can dramatically outperform linear filtering. The stochastic nonlinear difference equation that governs the dynamics of the proposed filtering is analyzed.;The second research area introduces a novel approach to the interference suppression problem, exploiting the structure of an interferer that is a digital communications signal. For such an interferer, significant performance improvement over previous suppression techniques is achieved by employing multiuser detection theory. The functions of detection and interference suppression are combined and an asymptotically optimal linear receiver is derived and analyzed.;The third section of this dissertation investigates the effects of phase noise on coherent optical CDMA. In designing RF receivers for CDMA, it is reasonable to assume that the phases of all users will be known or tracked. In optical systems, however, the phase can drift significantly over the integration times of the receiver. We investigate the effects of this phase drift on: multiuser receivers that ignore the drift, noncoherent receivers that ignore any coherence, and partially coherent receivers that try to find a middle ground between these extremes. The analysis involves the proof of a central limit theorem for dependent sequences, specifically a sampled phase noise modulated by independent equiprobable, binary random variables.