The Technology Researches Of Sound Source Localization Based On WSN

In recent years, sound source localization techniques have become a popularresearch. Single microphone arrays applications are already used on severaloccasions, such as battlefield anti-sniper, acoustic camera and voice enhancementareas. But it has the shortcoming of limited coverage and not fit for wide range ofsound source detection. To compensate for the lack of a single microphone arrayscoverage range shortcomings, this thesis use a multi-node distributed sound sourcelocalization system based on wireless sensor networks. Each node complete DOAestimation and coordinate estimation, it can share data between nodes throughwireless sensor networks, this can improve the coverage of the sound sourcelocalization, but also can improve the positioning accuracy.Analysis the advantages and disadvantages of normal used time delayestimation algorithms, using GCC algorithm to estimate time delay. Use thehyperbolic positioning principles to complete plane positioning. Hyperbolicequations are nonlinear; direct calculation of the intersection of all hyperbolicequations will take great amount of computation. This thesis presents a solutiononly requires compute one pair hyperbolic equations to reach DOA estimation, itgreatly reduces the calculated amount.Design a practical multi-node localization algorithm. Design microphonearrays and amplifier circuit. Use a digital oscilloscope to accomplish data collection,accomplish algorithm in MATLAB, set up an experimental platform to verify thedesigned algorithm. DOA angle error is less than3°in the range of0to360°.Coordinate positioning error not more than10cm in a rectangular area of threemeters long and two meters wide.