Research On WFRFT-based Interference-suppression Methods Under Doubly-selective Channels

With the diversification of current communication application scenarios, the phys-ical wireless channel that signals pass through is becoming more complex. Such situa-tion can be seen in wireless and underwater acoustic communication systems, when the communication terminals are moving rapidly against each other, and suffering from mul-tipath propagation. The consequent Doppler shift and multipath propagation result in signal dispersions appearing in both the time and the frequency domains. The so-called doubly selective/dispersive channel causes performance degradations of single carrier fre-quency domain equalization (SC-FDE) and orthogonal multi-carrier modulation (MCM) systems. First, the relative motion between transceivers results in a time variant channel response, which requires more complicated channel equalization methods. Besides, the signal dispersion in the time domain causes inter-sample interference (ISI), and the signal dispersion in the frequency domain causes inter-carrier interference. Such ISI/ICI will significantly influence the performance of detection and transmission efficiency.For above reasons, robust ISI/ICI suppressing methods are developed to meet tech-nical applications over doubly selective channels. For the past few years, relevant articles have developed time domain windows, channel equalization methods and novel modula-tion schemes to repress ISI/ICI in single carrier modulation (SCM) and MCM systems. This thesis launches the research on ISI/ICI suppression methods under a novel mod-ulation scheme, that is, the hybrid carrier modulation (HCM) based on weighted-type fractional Fourier transform (WFRFT), in order to improve performances of systems suf-fering from doubly selective channels. The reason that we choose WFRFT is:WFRFT can be regarded as an expansion of classical Fourier transform, which gives an established basis for signal conversions in the HCM system. Additionally, HCM is the expansion of the SCM and MCM systems, which allows a compatible carrier scheme evolution. Con-sidering the significant influence of Doppler shifts, this thesis aims at interference sup-pression methods for doubly selective channels with different Doppler shifts. Besides, for an uplink multi-user application, this thesis develops an FDMA system based on HCM architecture and a multi-user joint detection method.Firstly, this thesis focuses on wireless doubly selective channels with low Doppler spread, and proposes an HCM multi-tap linear equalization method. In addition, we ana-lyze performances of this combination of HCM and channel equalization under different channel conditions, and further explain the superiority of HCM over conventional carrier schemes:the amplitude of WFRFT domain residual interference after equalization shows a lower peak probability. Three linear equalization methods are employed in this the-sis, including block linear equalization and two serially linear equalization methods. The performance of HCM system with linear equalization depends on the HCM order, power delay profile and Doppler shift of the doubly selective channel. With increase of Doppler spread, HCM scheme begins to exhibit a superiority over SCM and MCM in terms of bit error ratio (BER).In order to reduce the influence of doubly selective channels with high Doppler shifts, such as underwater acoustic channels, this thesis improves the classical time do-main iterative equalization method, and proposes a time domain iterative MMSE equal-ization (TD-IMSE) based on the HCM architecture, which is implemented by updating and feeding back WFRFT domain prior information in each iteration. The accuracy of lin-ear MMSE estimation in each iteration depends on the convergence of prior information. Leveraging the combination of HCM and iterative method, the prior information in the WFRFT domain exhibits a better convergence, thus, the HCM-based TD-IMSE achieves a desirable BER performance under doubly selective channels with high Doppler shifts, and exhibits obvious superiority over SCM-and MCM-based TD-IMSE. However, with in-crease of multipath delay spread, the complexity of TD-IMSE significantly increases. An-other disadvantage of TD-IMSE is its inapplicability in FDMA systems. With these two factors in mind, this thesis further improves classical frequency domain iterative equal-ization methods, and proposes an HCM-based frequency domain iterative MMSE equal-ization (FD-IMSE) method. Leveraging the combination of HCM and this FD-IMSE, the system achieves better trade-off between performance and complexity. Due to the better convergence of WFRFT prior information, compared with SCM-based and MCM-based FD-IMSE, the HCM-based FD-IMSE exhibits obvious BER superiority under channels with high Doppler shifts.Finally, focusing on a scenario for multi-user uplink communication under doubly selective channels, this thesis leverages WFRFT to expand classical FDMA architectures, and proposes an hybrid carrier FDMA (HC-FDMA) architecture, which proves the com- patibility of HCM to the two conventional carrier schemes. This thesis further defines a multi-user joint frequency domain channel matrix for the uplink doubly selective channel, and proposes iterative multi-user joint detection methods based on this joint frequency do-main channel matrix. The PAPR of transmitted signals in HC-FDMA systems becomes higher than that in single carrier FDMA (SC-FDMA) systems with increase of HCM or-ders. Nevertheless, desirable BER performance can be achieved by modulating signals of different users in the HC-FDMA system with different channel-dependent HCM or-ders. When localized FDMA is adopted, HC-FDMA exhibits a BER superiority over SC-FDMA, and this superiority becomes more obvious with increase of Doppler shifts.