Research On PAPR Reduction And Channel Estimation Of Filter Bank Multicarrier System

The filter bank multicarrier system has received considerable attention for its advantages such as low leakage,strong immunity to interference,spectral efficiency and supporting the asynchronous communication.However,there are also some serious shortcomings in this system.First of all,because of FBMC is a member of multicarrier systems,the peak to average ratio in FBMC is high.More seriously,most traditional PAPR reduction schemes are not effective for FBMC signals due to its overlapping structure.Besides,the subcarrier in system only keeps orthogonally in real value filed.Therefore,channel estimation becomes a difficult problem.In this paper,we mainly study the suppression of peak to average ratio and channel estimation of FBMC system.Clipping method and partial transmission sequence technique are studied in this paper.When using PTS method to suppress the PAPR of FBMC system,considering the overlapping characteristics multiple symbols have to be optimized at the same time.There are two strategies to reduce the complexity of phase factor search,dynamic programming and greedy algorithm respectively.However,the dynamic programming algorithm MBJO-PTS-DP have the drawbacks of high complexity and huge resource cost,while the sub-optimal algorithm MBJO-PTS-Suc performed poor when the number of sub-blocks is large.Two searching schemes,iterative search based way and simulated annealing based way,are proposed according to different strategies.Simulation results show that both two methods can significantly improve the performance of the search with appropriate complexity.In addition,compared with the original algorithms,the simulated annealing based algorithm is more flexible and can better adapt to various occasions.On the channel estimation of FBMC system,both frequency domain and time domain estimation method are studied in this paper.An extension was made to the time domain estimation model to maximize the efficiency of the pilot and a set of optimal pilot is given correspondingly.The simulation results show that,in full case the estimation error was reduced by 2dB after optimization.The estimation performance also obtained a certain degree of improvement when the preamble structure is sparse.In addition,considering the correlation between FBMC symbols,a joint estimation method was proposed to further decreasing the estimation error.In the hardware design,basic modules of FBMC transceiver were designed similar with the method of OFDM system.Besides,the method of reducing the complexity of estimation by using QR decomposition is derived,and a corresponding device is designed based on the method.