PAPR Reduction for OFDM System with a Class of HNN

The PAPR issue in OFDM system can be formulated to a combinational optimization problem. This paper proposes a class of approaches to reduce the PAPR value of multi-carrier/OFDM system by solving this combinational optimization problem with some improved Hopfield neural networks (HNN) . Furthermore, the general theoretical framework of PAPR reduction based on all kinds of HNN is presented as well. By adopting new neural output function and random state disturbance, many kinds of HNN networks are implemented and used to solve this problem. By analyzing the PAPR reduction performance, we give some significative suggestion on PAPR issues based on HNN and verify the effectivity of HNN approaches. Simulation results show PAPR is improved greatly compared to the traditional PAPR reduction methods. The general improvement is about 3 dB. So it is a class of effective and practical algorithms for PAPR reduction in OFDM system