| In recent years,the satellite communication network,terrestrial communication network and underwater communication network connected by radio frequency,visib le light and underwater acoustic communication,provide timely and reliable guarantee for human economic development,weather forecast,disaster warning,resource development,military strategy research and transoceanic communication and cooperations.Orthogonal frequency division multiplexing(OFDM)technology in wireless communication systems can effectively solve the inter-symbol interference and inter-carrier interference caused by the dispersion channel.It plays an important role in improving the quality of wireless communication,increasing the utilization of spectrum and the rate of data transmission.Since OFDM signal is the superposition of signals on multiple subcarriers,high peaks cannot be avoided when the signals are in phase,which makes the design of power amplifier and digital to analog converter face a severe challenge.If the peaks exceeds the linear operating range of the devices,it will cause serious signal distortion and system performance deterioration.Therefore,as one of the key te chnologies in OFDM systems,peak to average power ratio(PAPR)reduction has become an important research topic.The PAPR reduction theory of OFDM signals are studied in depth mainly from the three application aspects,namely the radio frequency system,t he indoor visible light system and the underwater visible light system,with the relationship among the PAPR reduction,the BER performance and the computation complexity as the core chain.The main innovation work of this paper are summarized as follows:1.Based on the thorough performance analysis of the classic PAPR reduction methods,a novel PAPR reduction method based on the Back Propagation(BP)algorithm of multi-hidden layer feedforward neural networks is proposed to solve the problem of huge comp utation complexity and difficult trade-off between the PAPR reduction and the bit error rate(BER)performances in the convex optimization problem.During the first training stage,the hidden layer parameters are updated with the gradient descent method by constructing the connection weight and the offset value,which include the reserved subcarrier position index to reduce the signal PAPR.On this basis,the loss function which includes the information of BER is constructed during the second training stage,so that the PAPR can be reduced without sacrificing the bit error rate.2.A PAPR reduction algorithm based on the tone reservation scheme is proposed.The proposed algorithm is designed from the Signal-to-Clipping noise Ratio algorithm,and introduces the least square approximation method in the step length design of the iteration process.In each iteration,only the maximum peak signal is selected for processing,thereby reducing the complexity of the algorithm.The theoretical distribution of the PAPR,the computation complexity,and the performance of the BER performance over the Gauss white noise channel and the Rician fading channel,as well as the out of band interference performance of the proposed algorithm are studied.The implementation steps are given,and the performance of the proposed algorithm and a variety of classic PAPR algorithms are compared and analyzed through simulation experiments.The proposed algorithm makes a good trade-off between the factors such as PAPR reduction,bit error rate,out of band interference,computation complexity and so on.The algorithm obviously has better PAPR reduction performance than the traditional algorithms with fixed iteration step lengths.3.An indoor visible DC-biased optical OFDM(DCO-OFDM)system model with PAPR reduction is constructed based on the tone reservation scheme.The wireless light of sight(LOS)link model of indoor visible light is built through experiments.The parameters of the model are determined,and the transmission channel characteristic curves of the indoor visible light LOS link are analyzed.The method of Hermitian symmetry transformation,the method of extracting effective data subcarriers at the receiver and the method of adding DC-bias at the transmitter in the indoor visible DCO-OFDM system are studied.The block diagrams of the clipping method,the least square approximation method and the proposed method are illustrated,and the performances of the PAPR reduction and the bit error rate of the three methods in the indoor visible light DCO-OFDM system are compared and analyzed by simulations.Theoretical analysis and simulations show that the communication distance is the main factor that affecting the communication quality of indoor visible light DCO-OFDM system.The proposed algorithm has better performances compared with the other two methods.4.A PAPR reduction algorithm based on time-frequency interleaving is proposed.The underwater visible light DCO-OFDM system model is constructed.The related parameters of underwater visible light channel model are determined,and the relationship between the transmitting and the receiving light power over underwater visible LOS distance link is derived.The interleaving method and adjustable DC-bias adding method are studied.The photoelectric conversion mode of blue laser diode is given.A scheme of adaptively adjusting the amplifier at the receiver is proposed based on the least squares estimation algorithm.The computation complexity of the proposed scheme is studied.Through simulation experiments,the PAPR and the BER of the time-frequency interleaving algorithm and the time domain interleaving algorithm in the underwater visible light DCO-OFDM system are compared and analyzed.Simulation results show that by interleaving the t ime domain and frequency domain signals,the degree of autocorrelation of the original signals can be greatly reduced,and the PAPR of OFDM signals can be significantly reduced.The computational complexity of the proposed algorithm can be reduced by 50% compared with the time domain interleaving method.Within the range of best communication distance,the bit error rate can be kept at a low level. |
PDF Full Text Request