Multi-carrier Communication Systems, Signal Optimization And Simulation

Orthogonal frequency division multiplexing (OFDM) has been popular these years since it is strong in fighting against multipath effect and narrow band interference with high spectrum efficiency. However, there is one major drawback in OFDM systems—the high peak-to-average power radio (PAPR), which requires high dynamic level of power amplifier in transmitter and raises the cost of hardware implementation. Besides, OFDM signals generally have high out-of-band power. In the context of cognitive radio, this will cause the mutual interference between the licened users and rental users, violating the basice principle that rental users must remain invible when they access the spectrum and not cause any strong interterences to the licenced users.This dissertation focuses on the above-mentioned problems and the main work and contributions are summarized as follows:1. Improvement is made on one of the distoritionless PAPR reduction technique– SeLected Mapping (SLM) in OFDM systems and a greedy algorithm in SLM phase sequence selection is proposed. Given that PAPR reduction performance mainly depends on the number of candidate sequences and correlation among them, the proposed algorithm selectes the desired phase sequences with relatively low cross correlation by evaluating and minimizing the variance of correlation of given phase sequence set. For those candidate sequence set which is constructed by distinct rules and used for PAPR reduction, such as Walsh-Hadamard, Shapiro-Rudin and m sequence set, the proposed method will result in better PAPR reduction performance closer to performance bound in theory when compared with the counterpart by randomly choosing from the given sequence set, as confirmed by computer simulation.2. Research on one of the out-of-band suppression methods– N-continuous algorithm:1) A distortionless sidelobes reduction algorithm aiming at making OFDM signals N-continuous is proposed. Compared with traditional N-continuous method, the proposed one makes use of some of the the reserved tones to design N-continuous OFDM signal instead of using all of the tones to transmit data. At the receiver side, it does not need to remove the N-continuous noise by iterative estimation as the the conventional one. On the contrary, it can directly extract data from the data carriers, thus lowering the implementation complexity of receiver and bit-error-rate (BER) of system.2) Extend the N-continuous method from single-input single-output (SISO) scenario to multiple-input multiple-output (MIMO) counterpart. The idea of iterative interference cancellation is taken into account of the design of N-continous receiver in MIMO channels. Simulation shows that the improved N-continuous method can guarantee the the performance of both bit-error-rate and out-of-band power emission of system.3. The joint optimization of peak-to-average power ratio and out-of-band of OFDM signals is carefully investegated. Computer simulation shows that, the combination of serial peak cancellation and N-continuous method, after several iterative processing, can gradually reduce PAPR while keeping the out-of-band power within a satisfying level. Of course, this requires tradeoff between the performances of both the PAPR and out-of-band and the implementation complexity.