铁路行车事故分析中的语音压缩方法

A Method of Speech Compression in Railway Accident Analysis

在铁路运输中，安全行车是第一位的。一旦出现行车事故，对事故原因的分析往往十分棘手。采用黑匣子技术能够准确地记录列车运行中的各种信息，但巨大的语音存储量需求是黑匣子技术的一个难关。若采用传统的语音压缩技术，则理论和实现都很复杂，并且成本太高。采用神经网络不仅简单易行，还有提高信噪比的功能。

Black box techneque is capable of recording messages during running of a train for accident analysis,but a huge storage needed for speeches of drivers is a difficult problem in the technique.A method using neural networks is proposed to compress speech signals with a ratio of 4∶1 and with an increase in signal/noise ratio.