快速实现盲序列估计的一种新算法

当信道响应未知时,盲序列估计(BSE)利用接收到的有噪声信号直接恢复出源信息。作为一种最优信息检测方法,基于Viterbi算法实现的BSE需要搜索与观察序列长度成指数关系的状态树。当观察序列的长度不断增长时,运算消费的急剧上升将严重制约算法的实时实现。本文提出的盲序列估计新算法仅利用前几个信息的估值来恢复当前信息,它与观察序列的长度无关,因而能大大降低运算消费,仿真结果表明了该方法的可行性。

A NOVEL FAST ALGORITHM FOR BLIND SOURCE SEQUENCE ESTIMATION

Under the situation of unknown channel response, blind sequence estimation(BSE) recovers source symbols from received noisy data directly. As an optimal information detection approach, BSE based on Viterbi algorithm needs searching a trellis whose state number is exponential to the observation length. It is not practical when observation length increases unceasingly. A novel BSE scheme is proposed in this paper, which uses several first estimated results to recover current symbol. The computation burden is lower because it is irrelevant to observation length. The feasibility has been checked by simulation results.