Research On Suppressing Narrowband Interferences In DSSS Communication Systems

The fast increase of wireless communication services requires support of more and higher rate wireless links, which results in severe shortage of frequency resource. Recently, people have been making every effort to solve this problem. Except developing new frequency band and raising frequency band efficiency by using new modulation/coding and space-division multiplexing methods, it is also a very important way to make frequency resource shared by both broadband and narrow-band communication systems.Spread spectrum is one of the most common signaling schemes in current and emerging wireless services, which would find more and more applications. Spread spectrum communication systems often suffer strong narrowband interferences resulted from various factors. If the intensity of the interferences exceeds the interference margin of the system, its bit error rate will increase rapidly, even the system does not work. Fortunately, the interference margin of the system can be remarkably raised if the narrow interferences are excised or reduced from the received signal by preprocessing before dispreading, which proves to be very useful for increasing its robustness and communication capabilityThis thesis researches on the narrowband interference suppression in direct sequence spread spectrum communication systems. The author's main research contents and contributions are summarized as follows:After analyzing the performance and drawback of the narrow-band interference suppression based linear prediction and summarizing the existing algorithms of nonlinear processing, an effective non-linear interference suppressing algorithm is proposed, i.e. a non-linear QR-LMS fast algorithm for narrow-band interference suppression based on QR decomposition. The proposed algorithm has a complexity remarkably lower than the conventional ones while its performance is equivalent to the least square algorithm based on QR decomposition.For the asynchronous CDMA system over the frequency selective fading channel, the thesis proposes a blind algorithm of interference suppression based on MMSE criterion, moreover, has deduced a close expression of the filter weights in which both the ISI and MAI are takes into consideration. Furthermore, an effective narrow interference excising method is formed by approaching the expression with subspace decomposition combined with a robust subspace tracking based on project approaching.For the case of the multiple independent narrowband interferences appear in the spread spectrum system, the independent component analysis model is introduced and a corresponding method of narrowband interferences suppression is proposed. The proposed method takes the canonical correlation in the reproduced Hilbert space as the contrast function, and introduces the kernel method into the canonical correlation. The sources are separated with the interferences by maximizing the contrast function, so that the interference suppression is realized.For the case of the multiple narrowband interferences with unknown models exist in the system, a blind suppression algorithm based on the generalized side-lobe canceller is proposed. Through adding a linear constraint such as the signature of the desired user on the interference suppression filter, the filter weights update only in the subspace orthogonal to the constraint matrix. The filter can effectively suppress the narrow-band interferences while keep the desired signal unchanged. Besides, another novel algorithm to suppress narrow interferences in the frequency domain is proposed, which is realized by reconstructing the signal with the hard limited magnitude spectrum of the received signal combined with its original phase spectrum. The algorithm can not only improve the signal/interference ratio remarkably but also be applied to the case that the bandwidth of the interference is wider than that of the conventional.