Research On Peak-to-average Power Ratio Suppression Technology For OFDM Systems

OFDM has drawn much attention due to its high spectral efficiency, immunity to multi-path fading and the flexibility of implementation, but there are some key issues to be solved urgently. In which, the high peak-to-average power ratio (PAPR) is a main drawback. The background of this problem is discussed firstly, the OFDM technique is briefly reviewed and the research status of its PAPR problem are presented, which leads to the necessity of our research. Then different techniques proposed to solve this problem are introduced, including the distortion techniques, e.g., clipping and filtering, companding, and the distortionless techniques, e.g., coding, selected mapping (SLM), partial transmitted sequence (PTS) and tone reservation (TR).The TR method is deeply studied in this paper, and a high performance and low complexity TR scheme called TKM-TR is proposed. The TKM-TR suppresses multiple peaks in a single iteration by simultaneously applying multiple time-domain kernels to generate the time-domain kernel matrix and optimizing the scaling vector, thus accelerating the convergence speed. And there is no need to iteratively perform IFFT/FFT in the iterations due to the implementation of time-domain kernel, so the complexity is greatly reduced. The simulation results in MATLAB indicate that the performance of the algorithm is improved, compared to some other TR schemes.To further improve the performance of the TKM-TR, the peak regrowth and peak residual problem in the TKM-TR are studied. By applying distinct sign vectors to generate a set of different scaling vectors, alternative OFDM signals are obtained, and the one with the lowest PAPR is transmitted. The simulations of the proposed enhanced-TKM-TR show that the peak regrowth and peak residual are suppressed, and the PAPR performance is improved fairly, compared to the original TKM-TR.FPGA design and implementation of the proposed TKM-TR are accomplished, and the implementability and validity of the algorithm are verified by the comparison of the hardware data and the MATLAB simulation results.