Research On Adaptive Narrowband Suppression Techniques In Spread Spectrum Communication Systems

One of the important properties of the spread spectrum is that it inherently provides some protection against interference. As is well known, this inherent coding gain and processing gain determine the interference rejection capability. In fact, any level of interference protection can be obtained by designing the signal with sufficient processing gain. The price for greater protection, however, is an increase in the bandwidth of the transmitted signal for a given data bandwidth. Practical considerations such as transmitter/receiver complexity and available frequency spectrum can serve to limit the reasonably attainable processing gain. Hence, the system will not function properly when the level of interference surpasses the jamming margin.Narrowband interference arises in a number of types of practical spread spectrum systems. A classical, and still important, instance of this is narrowband jamming in tactical spread spectrum communications because that NBI is easy to produce, its power spectrum density (PSD) can reach to very high degree, and several NBI's can overlay a given bandwidth. A further situation in which NBI can be a significant factor in spread spectrum systems is in systems which share bandwidth with narrowband communication systems. Thus, the issue of active NBI suppression in spread spectrum is one that is of increasing importance.It has been known for decades that active methods of NBI suppression can significantly improve the performance of such systems. Existing active NBI suppression techniques have focused on three basic types of techniques: predictive techniques, transform-domain techniques, and code-aided techniques. However, these techniques do not aim at improving the bit error rate (BER) of the spread spectrum in the most.This dissertation studies the optimum interference suppressor in the minimum bit error rate (MBER) sense. The NBI suppression performance of this suppressor is investigated, and then compared with performance of the other suppression techniques.The efficient implementation of the linear optimum interference suppressor is then arisen. By examining the relationship between the MBER, maximum signal-to-interference ratio (MSIR), minimum mean square error (MMSE) and the constrained minimum mean output energy (MMOE), we resolve the above problem into the implementation of the optimum suppressor in the MMSE and the constrained MMOE criterion. As a consequent, the adaptive and blind adaptive algorithms are developed.This dissertation addresses the least-mean-square (LMS), transform domain least-mean-square (TrDLMS), and recursive least-squares (RLS) adaptive versions of the MMSE suppressor. The QR decomposition-based RLS (QR-RLS) algorithm can overcome two major problems associated with the RLS, that is, numerical instability and difficult for parallel implementation. This dissertation discusses the systolic array processing for parallel implementation of the QR-RLS algorithm. By establishing the state equation based on the weight vector of theoptimum interference suppressor and the measurement equation based on the multiple linear regression model of the desire signal, the (Caiman filtering algorithm is then presented.In the above-mentioned adaptive algorithms, the desired signal is assumed to be known to the receiver. The linear constrained MMOE suppressor admits a blind adaptive implementation. After a procedure for transforming constrained MMOE problem to unconstrained MMSE problem, the blind LMS (BLMS), blind recursive least squares (BRLS), and blind Kalman filtering adaptive versions of the constrained MMOE criterion for interference suppressor is developed and analyzed.The blind algorithm without known data sequences has tracking capabilities in a time-varying environment. However, the NBI suppression performance of the blind algorithms is shown to be significantly inferior to the corresponding adaptive algorithms. Therefore, it is desirable to cooperate with a decision-directed adaptation mode as soon as the blind adaptation convergences. Whenever a sudden change in the en...