飞机驾驶舱特定声音信号提取

摘要

At the present stage,the identifica-tion and analysis of the acoustic signals in Cockpit Voice Recorder (CVR)cannot identify the specific voice in airplane efficiently which ignore the char-acteristic of the fixed acoustic source in cabin.A method based on Independent Component Analysis (ICA )is proposed to separate acoustic signals in this paper.The simulation is done by Matlab, many groups of experiment data are analyzed by Monte Carlo Method (MCM)to confirm the stabil-ity of this algorithm,and the results show that the specific voice is extracted efficiently.%针对现阶段国内对舱音记录仪（CVR ），所记录声音的辨别和分析方法没有充分利用机舱内声音的定源特性，不能准确地分辨出飞机上特定声音，提出了一种基于独立成分分析（ICA ）对舱音信号进行分离的方法，用 Matlab 软件进行了仿真，并通过蒙特卡洛方法（MC M ）计算多组实验数据的性能指标，以验证算法的稳定性。结果表明，该方法可以比较高效地提取出飞机机舱中特定语音信号。