A study of distributed adaptive node-specific signal estimation (DANSE) algorithm

• Main Author: Goh, Chen Chuan.
• Other Authors: School of Electrical and Electronic Engineering
• Format: Theses and Dissertations
• Language: English
• Published: 2012
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/50953
• Institution: Nanyang Technological University

Conventionally, microphone array with multiple elements are used for the gathering of spatial-temporal information of an acoustic environment in noise reduction. For further improvement, multiple spatial distributed arrays are used to provide a better representation ofthe acoustic field. A class of multichannel Wiener filtering technique known as Speech Distortion Weighted Multichannel Wiener Filter (SDW-MWF) is developed with its performance close to theoretical optimum. Basing on SDW-MWF, a distributed noise reduction technique known as Distributed Adaptive Node-specific Signal Estimation (DANSE) is introduced to address the problem of increasing computational complexity and large communication bandwidth that bundled with the increase in the number of sensor nodes for the centralized SDWMWF implementation. This dissertation is to study the performance of these noise reduction algorithms using adaptive implementations and the influence of several parameters such as DFT size used for frequency domain processing and the acoustic reverberation time.