OFDM系统中基于改进阈值的信道估计算法的研究

发布时间：2018-04-21 02:24

本文选题：正交频分复用 + 信道估计　；参考：《东北大学》2014年硕士论文

【摘要】：正交频分复用(OFDM)是一种多载波调制和复用技术,具有较高的频谱利用率和较强的抗干扰能力,目前已经广泛应用于各种高速数据传输领域。信道估计对OFDM系统的性能有着直接的影响,是OFDM系统的研究热点之一,也是进行相干检测、解调和均衡的基础。首先,论文介绍了无线信道的传输特性,重点研究了多径传播和多普勒效应的原理。详细阐述了OFDM系统的基本原理,包括OFDM的调制与解调、保护间隔和循环前缀的作用、OFDM参数的选择。在此基础上,重点研究了基于导频的信道估计典型算法,分析了各种典型算法的性能和复杂度。其次,论文深入研究了基于阈值的DFT信道估计。在分析这些原有阈值的缺点和不足的基础上,提出了一种新的将噪声样本点能量的中位数作为改进阈值的DFT信道估计。仿真实验表明,所提算法在大脉冲噪声情况下,可以很好地解决信道有效样本点被滤除的问题,且能使循环前缀内的噪声得到有效抑制,误码率和均方误差得到进一步降低,提高了信道估计的准确性。最后,论文深入研究了基于小波去噪的信道估计。研究发现由最小二乘信道估计得到的信道频率响应具有某些不连续点或快速变化点,所以当使用小波进行去噪时,会出现伪吉布斯现象,因此本文将平移不变量小波去噪应用于信道估计,并提出了一种处理小波高频系数的改进阈值函数。仿真实验表明,本文提出的改进算法能够有效地抑制伪吉布斯现象,更好的滤除噪声,使误码率和均方误差进一步降低,得到了更加准确的信道频率响应。
[Abstract]:Orthogonal Frequency Division Multiplexing (OFDM) is a kind of multicarrier modulation and multiplexing technology, which has high spectrum efficiency and strong anti-jamming ability. It has been widely used in various high-speed data transmission fields. Channel estimation has a direct impact on the performance of OFDM systems. It is one of the research hotspots in OFDM systems and is also the basis for coherent detection demodulation and equalization. Firstly, the transmission characteristics of wireless channel are introduced, and the principle of multipath propagation and Doppler effect is studied. The basic principle of OFDM system is described in detail, including modulation and demodulation of OFDM, protection interval and function of cyclic prefix. On this basis, the typical channel estimation algorithms based on pilot frequency are studied, and the performance and complexity of the typical algorithms are analyzed. Secondly, DFT channel estimation based on threshold is studied in detail. Based on the analysis of the shortcomings and shortcomings of these original thresholds, a new DFT channel estimation method using the median energy of noise sample points as the improved threshold is proposed. The simulation results show that the proposed algorithm can solve the problem that the channel effective sample points are filtered out under the condition of large impulse noise, and can effectively suppress the noise in the cyclic prefix and further reduce the bit error rate and mean square error. The accuracy of channel estimation is improved. Finally, the channel estimation based on wavelet denoising is studied in detail. It is found that the channel frequency response estimated by the least square channel has some discontinuous points or fast changing points, so the pseudo-Gibbs phenomenon will occur when the wavelet is used for denoising. In this paper, the translation invariant wavelet denoising is applied to channel estimation, and an improved threshold function is proposed to deal with the high frequency coefficients of wavelet. Simulation results show that the proposed improved algorithm can effectively suppress the pseudo-Gibbs phenomenon, better filter noise, further reduce the bit error rate and mean square error, and obtain more accurate channel frequency response.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.53

【相似文献】