A system for cocktail party effect using large sensor array

使用大型传感器阵列实现鸡尾酒会效果的系统

• 批准号: 12450164
• 负责人: TAKAHASHI Kota
• 金额: $ 9.22万
• 依托单位: The University of Electro-Communications
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (B)
• 财政年份: 2000
• 资助国家: 日本
• 起止时间: 2000 至 2002
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/en/grant/KAKENHI-PROJECT-12450164/
• 关键词: array signal processing; digital signal processing; cocktail party effect; 視聴覚融合

A psychological term "cocktail party effect" means biological special ability to extract an important signal from other ambient noises. In this technological research, the cocktail party effect is focused on in order to construct a new type of sensing system.An audio-visual sensor fusion system which integrates audio signals and sequential images on time-space coordinates, is proposed in this research. Furthermore, a large microphone array system is incorporated into the proposed system. Summaries of the research are as follows.A prototype system with the large microphone array of 156 microphones has been constructed. DSP modules for parallel processing have been designed and assembled. Communication modules for DSP-to-DSP communications are also designed and assembled. Furthermore, software tools including an automatic coding software for communication tasks and debugger have been constructed. Using these hardware and software, real-time sound source separation can be demonstrated.Theoretical analysis for the signal separation using the large microphone array are performed. A new signal separation method are proposed, which is a combination of sub-space method and grouping method based on the theory of independent component analysis.Results of experiments using the prototype system demonstrated the advantages of the large microphone array and audio-visual sensor fusion scheme.

心理学术语“鸡尾酒会效应”指的是从其他环境噪声中提取重要信号的生物特殊能力。本研究针对鸡尾酒会效应，提出一种在时空坐标系上融合音频信号与序列图像的视听传感器融合系统，以构建一种新型的感知系统。此外，一个大的麦克风阵列系统被纳入所提出的系统。主要工作如下：构建了一个由156个麦克风组成的大麦克风阵列的原型系统。设计并组装了并行处理的DSP模块。设计并组装了DSP到DSP的通信模块。此外，软件工具，包括一个自动编码软件的通信任务和调试器已经建成。利用这些硬件和软件，可以实现实时声源分离，并对大型麦克风阵列的信号分离进行了理论分析。提出了一种基于独立分量分析理论的子空间法和分组法相结合的信号分离方法，并在原型系统上进行了实验，实验结果表明了大麦克风阵列和音视频传感器融合方案的优越性。