【摘要】 针对传统最小二乘(least square,LS)算法估计时变水声OFDM信道误差较大问题,提出一种基于压缩感知的准确估计方法.首先导出了水声OFDM系统接收序列、发送序列及信道传输矩阵之间的关系;在此基础上利用水声信道稀疏特征,探讨了采用正交匹配追踪(orthogonal matching pursuit,OMP)算法估计水声OFDM信道的方法;分析了导频插入方式、导频数及径数等对水声OFDM信道估计均方误差(mean square error,MSE)的影响.研究表明:当导频随机插入时,OMP算法用较少导频即可准确估计水声OFDM信道的传输矩阵.当信噪比大于10dB时,OMP算法的MSE小于-24dB,比传统的LS算法小18dB.更多还原

【Abstract】 A novel and accurate method based on compressed sensing to estimate the time-varying underwater acoustic OFDM channels is introduced for solving the problem of large estimation errors of classical least square(LS)algorithm in the paper.The relationship among the received sequence,the sent sequence and the channel transfer matrix of the underwater acoustic OFDM system is deduced firstly.Based on the relationship and motivated by sparse channel characteristics,a method using orthogonal matching pursuit(OMP)algorithm to estimate underwater acoustic OFDM channels is investigated.The effects of pilot arrangement schemes,pilot number and multipath on underwater acoustic channel estimation errors are analyzed.Simulations show that the channel transfer matrix of the underwater acoustic OFDM system can be estimated accurately by OMP algorithm with fewer pilots when they are arranged randomly.When the signal to noise ratio is greater than 10dB,the mean square error(MSE)of OMP algorithm is less than-24dB,18dB less than that of LS algorithm.更多还原