基于改进贝叶斯压缩感知的正交频分复用系统信道估计

摘要

Orthogonal frequency division multiplexing (OFDM)systems use traditional compressed sensing algorithm to make channel estimation,this needs the information such as the known channel sparsity,etc.,and the algorithm has high complexity and long reconstruction time as well.To solve these problems,we proposed to improve Bayesian compressive sensing algorithm for OFDM channel estimation.The proposed algorithm converts the OFDM’s channel estimation to Bayesian compressive sensing reconstruction problem.Under the condition of not knowing the channel sparsity information in advance,it gets the channel estimation values and error range by optimising the basis function selection in reconstruction process and gradually adding the basis function from one to all rather than deleting,this makes the algorithm have higher speed in convergence.Simulation results indicated that compared with traditional channel estimation algorithm,the proposed algorithm did not need sparsity information of channel and had higher reconstruction accuracy,better estimation performance in lower signal-to-noise ratio condition.It improved the operation speed and reduced the complexity.%针对正交频分复用（OFDM）系统利用传统压缩感知算法进行信道估计需要已知信道稀疏度等消息，且算法复杂度高，重构时间长的问题，提出改进贝叶斯压缩感知算法进行 OFDM信道估计。该算法将正交频分复用系统的信道估计转化为贝叶斯压缩感知重构问题，在不需要预先知道信道稀疏度信息的情况下，通过优化重构过程中的基函数选择方法，将基函数从1个开始逐渐增加，而不是删除，进而得到信道估计值以及误差范围，使该算法具有更快的收敛速度。仿真结果表明，与传统信道估计算法相比，该算法不需要信道的稀疏度信息，并且重构精度更高，在低信噪比的情况下估计效果更好，提高了运算速度，降低了复杂度。