Conventional Minimum-Variance (MV) channel estimation is affected by two sources of error, namely the finite number of samples used to estimate the covariance matrix and the asymptotic bias due to interference and additive noise. On the other hand, widely linear (WL) filtering has been shown to improve the estimation of improper complex signals. Researchers have recently demonstrated that the application of WL processing principles can significantly improve the performance of subspace-based channel estimation algorithms. However, in contrast to MV estimation algorithms, subspace-based algorithms assume knowledge of the total number of users in the system, and must be coupled with sophisticated user enumeration algorithm at the expense of increased complexity. In this work, in an effort to combine the practical advantages of MV channel estimation algorithms with the performance of WL filters we propose a widely linear version of the MV channel estimator in the context of multicarrier(MC) CDMA systems employing real modulation. We use numerical simulations to demonstrate that the widely linear minimum-variance algorithm yields more accurate channel estimates compared to the conventional MV algorithm. By considering two simplified transmission/reception models, we also show analytically that the widely linear estimator on average reduces both types of error.

Modern communications has become far more than point-to-point calling and wireless communications is part of every-day life. Driven by ever growing demand for high data rate communication, multiple-access techniques are of interest for allowing multiple users to share limited resources, such as frequency, time and space. Commercially introduced in 1995, Code-Division Multiple-Access (CDMA) quickly became one of the world's fastest-growing wireless technologies. However, CDMA is subject to some limiting factors, such as multiple-access interference (MAI), which dramatically affects the capacity of the wireless system and degrades performance. Fortunately, these effects can be alleviated by applying advanced signal processing techniques such as multiuser detection (MUD), which potentially provides a large increase in system capacity, enhances spectral efficiency, and relaxes requirements for power control. / Further improvements of MUD can be obtained through joint multiuser detection/decoding. However this is a very complex approach. Inspired by Turbo codes and iterative decoding, Turbo-MUD and iterative multiuser decoding have been proposed. The main objective of this research is to analyse the existing iterative techniques applied to Turbo multiuser decoding for coded CDMA systems and propose new decoder structures to improve the system performance. / In this thesis, we observe that many of the iterative multiuser decoding algorithms in the literature are focused on exchanging information obtained within the most current iteration. However, if correlations over iterations are low, then in principle the bit error rate (BER) performance can be improved by combining signal estimates over iterations. Inspired by this idea, iterative maximal ratio combining (MRC) is proposed in this thesis for application to iterative decoding structures. With this approach all previous estimates are recursively weighted and combined to refine the current signal estimates. The derivation of the corresponding weighting factors is based on the statistics of the decoder outputs over iterations, which leads to maximizing the resultant signal-to-noise ratio (SNR) for each current signal estimate. It is shown that the recursive MRC scheme can be widely applied to many existing iterative structures and provide significantly improved system performance with acceptable computational complexity. In addition, the analytic and numerical results illustrate that the resulting performance gain from the application of MRC is inversely proportional to the correlation of the decoder estimates across iterations. The more correlated the signal estimates over consecutive iterations are, the slower system convergence will be, if MRC is employed over all iterations. MRC over only a few initial iterations where correlation across those iterations is low provides faster convergence. A truncated MRC is suggested, which provides better performance while maintaining low computational complexity. Simulation results based on monte carlo averaging demonstrate that the system performance for the proposed techniques is better than many existing algorithms in the literature. / Thesis (MA(Telecommunications))--University of South Australia, 2005.

Thesis (Ph.D.)--University of South Australia, 1996.

The introduction of cellular wireless systems in the 1980s has resulted in a huge demand for personal communication services. This demand has made larger capacity systems necessary. This has been partially satisfied by the introduction of second generation digital systems. New third generation systems are now under going standardisation and will require even more efficient utilisation of the spectrum if the high bandwidth features and larger capacity are to become a reality. Motivated by these growing requirements we discuss methods of achieving large improvements in spectral efficiency and performance. Multiple-user communications over a channel can only be achieved with some form of diversity. In this work we point out that the efficient utilisation of the dimensions of space, time, and frequency will ultimately maximise the system capacity of a multiple-user system. We apply our receiver techniques solely to the base-station design where capacity limitations are currently present. We note however, that some of these techniques could also be applied at the mobile terminal receiver. We primarily focus our attention on the direct-sequence code-division multiple-access (DS/CDMA) channel, since this channel is inherently interference limited by other users in the cell of interest. We exploit a new powerful channel coding technique named " turbo coding" for its iterative decoding approach. We show how we can substitute the inner convolutional code of a turbo code encoder with the CDMA channel. By " iterative detection/decoding" or " turbo equalisation" at the receiver we achieve performance results which show the interference from other users to approach complete removal. We develop and analyse a new, low complexity, iterative interference canceller/decoder. This receiver has complexity per user linear with the memory of the channel and independent of the number of users in the system. We extend this receiver to more realistic channels that are asynchronous and include multi-path, and include spatial diversity by using an antenna array at the receiver. The CDMA channel we study exclusively uses randomly generated spreading codes. With this channel model we still achieve single user performance (no interference from other users) with a 10logL gain from L antenna elements and a gain of up to 10logP from P multi-path components. With any new receiver design, sensitivity to channel parameter errors is of paramount interest. We find that the sensitivity of our receiver is low with respect to the parameter errors induced. This is as we desire for a realisable receiver design. Finally we investigate the application of this new iterative interference canceller/decoder receiver to a number of other interference channels. These include the intersymbol interference (ISI) channel, partial response signalling (PRS), and continuous phase modulation (CPM). For these channels excellent performance improvement is generally achieved by the utilisation of the iterative interference canceller/decoder solution. / Thesis (PhD)--University of South Australia, 1999

Thesis (Ph. D.)--University of Hawaii at Manoa, 2005. / Includes bibliographical references (leaves 84-89).

Thesis (M.S.)--Virginia Polytechnic Institute and State University, 1994. / Vita. Abstract. Includes bibliographical references (leaves 87-91). Also available via the Internet.

Project (M.Eng.) - Simon Fraser University, 2004. / Theses (School of Engineering Science) / Simon Fraser University.

Thesis (M. Phil.)--University of Hong Kong, 2000. / Includes bibliographical references (leaves 116-121).

Thesis (Ph. D.)--University of Hong Kong, 2000. / Includes bibliographical references (leaves 108-114).

Thesis (M. Phil.)--University of Hong Kong, 1998. / Includes bibliographical references (leaf 92-96).