Blind interference cancellation in wireless digital communication systems.

摘要

Interference such as multiple access (or interchannel) and/or intersymbol interference resulting from imperfect characteristics of communication channels is a major problem in digital communication systems. However, due to the inherent nature of the signals in a digital communication system, it is possible to develop advanced signal processing techniques that can alleviate this problem.; In this dissertation, new interference suppression criteria and adaptive algorithms are proposed for multichannel decision feedback equalizers/filters. Advantage of these methods is that they do not require training sequences that erode capacity. An analysis on the convergence shows advantages of the proposed algorithms over the conventional adaptive algorithms such as constant modulus and minimum mean-squared error methods with imprecise and partial knowledge of the communication channels.; Despite its superiority over the linear equalizer and its popularity, a major drawback of the decision feedback equalizer is that it suffers from error propagation. An intuitive choice is one where we replace the hard limiter with a ‘softer’ function. In this research, a different approach based on the joint entropy maximization is proposed to introduce this soft decision device. This information theoretical framework also allows one to introduce this soft decision device naturally not only in the feedback path but also in the adaptive algorithms. It is also shown that there is a link between the proposed joint entropy maximization algorithm and classical Bussgang-type algorithms that are currently used in channel equalization.; Finally, these new techniques are applied to cancel interference in a direct sequence code division multiple access system. The new multistage receiver structure is enable to mitigate the time-dispersive effect of multipath channels. The proposed receiver offers improved performance and reduced computational load over the conventional ones. Although application to CDMA multiuser system is considered in this research, algorithms derived in this dissertation also find applications in multiple head digital recording system, cellular radio system, and dually polarized radio system. Indeed, computer simulations show effectiveness of the proposed receiver structures and algorithms in this dissertation.