On signal transmission and detection over fading channels.

摘要

There has been a great deal of interest in the last years in developing new schemes that efficiently utilize the resources of a communication system in order to increase its spectral efficiency, minimize its power consumption, improve link quality, and reduce overall complexity. These goals become more formidable in a wireless system mainly due to medium impairments such as fading, shadowing, and path losses. Coupling these goals with reality (practical implementation and medium nature), we pursue the following research thrusts. The first thrust deals with low power applications and proposes a variable-rate and variable-power (VRVP) non-coherent M-ary frequency shift keying (NC-MFSK) scheme that offers considerable power saving when compared to the traditional adaptive M-ary quadrature amplitude modulation (M-QAM). An adaptive switching-thresholds module can also be added to render the system more robust against outdated channel knowledge. For sake of merging the power saving and spectral efficiency merits of FSK and QAM schemes, respectively, the second thrust proposes a hybrid NC-MFSK/M-QAM scheme that extends the range of coverage and channel utilization. In the third thrust, we study a wideband communication system that capitalizes on the orthogonal frequency division multiplexing (OFDM) technique. Optimum algorithms for bit and power allocation are provided so as to minimize the average bit error rate (BER) while respecting total power and rate constraints. Multiple-antenna implementation at the receiver side is addressed in the fourth thrust wherein we propose an adaptive discrete power loading scheme that minimizes the average BER. Since all the proposed schemes hinge upon the channel knowledge at transmitter, the fifth thrust aims at investigating the effect of imperfect channel knowledge in a generalized setup of multiple-input multiple-output (MIMO) antenna configuration. Namely, we jointly optimize both pilot and data sequences in order to boost the ergodic system capacity under realistic assumptions of imperfect receiver estimates and transmit correlation. The final thrust addresses one source of complete lack of knowledge, namely unknown-signal transmission. We consider the simplest version of this problem by designing energy and generalized maximum likelihood detectors that can just tell whether a signal exists or not.