相干调制水声数字通信中的一种盲符号估计方法

A New Blind Symbol Estimation Algorithm for Coherent Underwater Acoustic Communication

相干调制是近年来水声通信领域中最为活跃的研究课题之一。但水声信道的时变多途效应造成接收数据严重的码间干扰（ Ｉ Ｓ Ｉ），使得高速水声通讯的实现十分困难。盲符号估计不需要利用已知序列对滤波器参数进行周期训练，从而能有效地节省信道带宽。利用盲估计方法克服 Ｉ Ｓ Ｉ的关键在于，对多途接收的有噪信号进行多倍码速率的采样，将信道等效为一个单输入多输出系统，并构造采样输出的循环平稳序列的二阶统计量，去估计原发送序列。本文针对水声信道的实际特点，提出了一种盲符号估计的新方法。与文献［３］的算法相比，它具有更低的运算量，便于数据的实时处理。文中最后利用计算机仿真结果验证了该算法的有效性。

Coherent modulation has been one of the most active research fields in digital acoustic communication recently. Time variant multipath propagation in acoustic channel causes serious intersymbol interference (ISI), which makes high speed data transmission over acoustic channel very difficult. Blind symbol estimation can save bandwidth efficiently because the method needs not predetermine channel response and perform channel identification. The key idea involved in overcoming ISI with blind symbol estimation is to sample the multipath data with noise at a rate several times that of symbol transmission and make the channel equivalent to a single input multi output system. The cyclostationary sequence thus obtained is used to construct second order statistics for estimating the original source symbol. Taking into account the special features of underwater acoustic channel, we propose a new blind symbol estimation algorithm. Compared with the method of Ref.3, our algorthm′s computational overhead is less and it can meet the requirements of real time processing. Simulation results show preliminarily the feasibility of our algorithm.