29 April 2021 in IEEE Communications Letters
Deep learning (DL) has dramatically improved the peak-to-average power ratio (PAPR) performance. However, the high computational complexity and excessive training data constitute a significant hurdle. In this letter, a model-driven deep learning algorithm is proposed for PAPR reduction in orthogonal frequency division multiplexing (OFDM) system. Precisely, an iterative peak-canceling signal generation scheme is unfolded as a layer structure of the DL network. The scheme falls into the category of tone reservation technique. A set of trainable parameters, which optimizes the clipping threshold and weights time-domain kernel function, has been designed and introduced into the iterative scheme. Compared with the existing approaches, the simulation results demonstrate that the proposed algorithm achieves comparable PAPR performance with low complexity and training costs.
27 August 2020 in IEEE Wireless Communications Letters
One of the limitations of the Neural Network-based peak-to-average power ratio (PAPR) reduction methods lies in the long-term training process through a large amount of training data. To address this issue, an extreme learning machine-based (ELM) tone reservation scheme is proposed, which could achieve comparable PAPR performance with low complexity compared to other Neural Network-based algorithms. An ELM-based mapper, trained on the clipping control tone reservation (CC-TR) signals, is designed to reduce the PAPR of orthogonal frequency division multiplexing (OFDM) system. Simulation results indicate that the proposed algorithm owns the advantages of fast learning capability and short training length.