The Research On The Papr Reduction For Large-scale Multi-user MIMO-OFDM Systems

Large scale Multiple-input Multiple-output(MIMO) technology has been widely studied over the past two decades, since it can significantly increase the capacity and reliability of the system. Although large scale MIMO technology can deal with the multipath fading, its resistance to the frequency selective fading is not good enough. It is known that Orthogonal Frequency Division Multiplexing(OFDM) technology has high robustness to multi-path fading. Thus, the OFDM technology can be combined with the large scale MIMO technology, and generates large scale MIMO-OFDM communication system. Moreover, the performance of multi-user MIMO(MU-MIMO) systems is generally less sensitive to the propagation environment than in the point-to-point MIMO case, thus, focus has shifted to large scale MU-MIMO-OFDM. However, the inherent Peak-to-Average Power Ratio(PAPR) problem in OFDM systems is also introduce into the large scale MU-MIMO-OFDM systems. Since single-antenna OFDM systems is the basis of the large scale MU-MIMO-OFDM systems, we firstly investigate the single-antenna OFDM systems and proposes a novel PAPR reduction method. Then, a low complexity method is proposed for the large scale MU-MIMO-OFDM systemsFirstly, we propose a novel method named Constellation Amplitude Modification to reduce the PAPR of OFDM signal. The CAM method formulates the PAPR reduction as a convex optimization by two constraints, power constraint and amplitude constraint. Furthermore, we also propose a customized interior-point method to solve the problem, and a detailed derivation is presented. Simulation results show that the CAM method converges faster and offers a better PAPR reducton performance than the ACE method.Secondly, we propose a TR-based method to reduce the PAPR of large-scale multi-user MIMO-OFDM system. The method is based on the fact that only a small set of signal exceed the linear region of HPA, thus, we only process the signal with maximum PAPR using TR method in each circulation. Although finding the signal with maximum PAPR needs computation, the proposed method can reduce more PAPR with lower computational complexity compared with applying the TR method on every antenna.