PAPR Reduction Technique for Mobile Communications System Based on Pattern Learning

This work presents a new scheme for reducing PAPR in OFDM systems based on pattern learning. In the proposed scheme, a set of subcarriers is reserved for controlling the PAPR. The pilot values associated with each subcarrier within this set are defined, aiming to reduce the PAPR to a given target. The sets of pilot subcarriers that lead to the target PAPR are saved and associated with a parameter of the input sequence. This same set is used whenever an input sequence presents the same parameter, reducing the overall system complexity. The proposed technique demonstrates superior performance in terms of PAPR reduction when compared with standard techniques, such as PTS, providing PAPR values up to 5 dB below the benchmark. One main advantage of the proposed scheme is that it allows the maximum tolerable PAPR to be defined at the cost of higher complexity. Typically, the complexity grows with the reduction of the target PAPR. Hence, the system can be configured according to the trade-off between PAPR reduction and complexity.