Abstract
The growth of Personal Communications, the keyword of the 90s, has already the signs of a technological revolution. The foundations of this revolution are currently set through the standardization of the Universal Mobile Telecommunication System (UMTS), a communication system with synergistic terrestrial and satellite segments. The main characteristic of the UMTS radio interface, is the provision of ISDN services. Services with higher than voice data rates require more spectrum, thus techniques that utilize spectrum as efficiently as possible are currently at the forefront of the research community interests. Two of the most spectrally efficient multiple access technologies, namely. Code Division Multiple Access (CDMA) and Time Division Multiple Access (TDMA) concentrate the efforts of the European telecommunity.This thesis addresses problems and. proposes solutions for CDMA systems that must comply with the UMTS requirements. Prompted by Viterbi's call for further extending the potential of CDMA through signal processing at the receiving end, we propose new Minimum Mean Square Error receiver architectures. MMSE detection schemes offer significant advantages compared to the conventional correlation based receivers as they are NEar FAr Resistant (NEFAR) over a wide range of interfering power levels. The NEFAR characteristic of these detectors reduces considerably the requirements of the power control loops currently found in commercial CDMA systems. MMSE detectors are also found, to have significant performance gains over other well established interference cancellation techniques like the decorrelating detector, especially in heavily loaded system conditions. The implementation architecture of MMSE receivers can be either Multiple-Input Multiple Output (MIMO) or Single-Input Single-Output. The later offers not only complexity that is comparable to the conventional detector, but also has the inherent advantage of employing adaptive algorithms which can be used to provide both the dispreading and the interference cancellation function, without the knowledge of the codes of interfering users. Furthermore, in multipath fading channels, adaptive MMSE detectors can exploit the multipath diversity acting as RAKE combiners. The later ability is distinctive to MMSE based receivers, and it is achieved in an autonomous fashion, without the knowledge of the multipath intensity profile. The communicator achieves its performance objectives by the synergy of the signal processor and the channel decoder. According to the propositions of this thesis, the form of the signal processor needs to be changed, in order to exploit the horizons of spread spectrum signaling. However, maximum likelihood channel decoding algorithms need not change. It is the way that these algorithms are utilized that needs to be revis ed. In this respect, we identify three major utilization scenarios and an attempt is made to quantify which of the three best matches the requirements of a UMTS oriented CDMA radio interface. Based on our findings, channel coding can be used as a mapping technique from the information bit to a more ''intelligent" chip, matching the ''intelligence" of the signal processor.