Aiming at the difficulty of accurately obtaining time-varying channel state information in moving underwater acous-tic communication with low complexity,a low complexity bidi-rectional underwater acoustic channel estimation algorithm based on fast information collection and virtual training (FIC-VT) was proposed. A superimposed training (ST) scheme was adopted to linearly superimpose a symbol sequence and training sequence to ensure continuous transmission of the training sequence and improve the time-varying channel track-ing capability. Based on belief propagation,the FIC-VT algo-rithm was proposed,which divided a block of data into multi-ple short blocks,each of which was divided into multiple sub segments,and by a fast information collection algorithm,the channel information of multiple sub segments was fused to ob-tain low complexity local channel estimates for each short block. Turbo equalization was employed to virtualize the esti-mated symbol sequence into a virtual training sequence (VT),and through iterative calculation,the proposed algo-rithm eventually achieves low-complexity and accurate estima-tion of time-varying underwater acoustic channels. The pro-posed algorithm is implemented by using fast Fourier transform (FFT),with computational complexity per tap at a logarith-mic level. The effectiveness of the proposed algorithm is veri-fied through computer simulation,pool motion communication experiments,and Jiaozhou Bay motion communication experi-ments.
关键词
时变水声信道/低复杂度双向信道估计/快速信息收集/虚拟训练序列/叠加训练/快速傅里叶变换(FFT)
Key words
time-varying underwater acoustic channel/low complexity bidirectional channel estimation/fast information collection/virtual training sequence/superimposed training/fast Fourier transform(FFT)
维普
万方数据