Study On Acoustic Localization Method Based On The Time Delay Estimation Of Mobile Robot

The research and development of sound source localization system is a research focus ofindustrial control and signal processing. One of the main contents of the study is the locationmethod.In this paper, microphone array is used to study the sound source localization system. It isdiscussed that the necessity and the main difficulties of sound source localization based on themicrophone array. And it introduces several common methods of sound source localization, aswell sums up and compares their advantages and disadvantages. This paper mainly studied themethod of time delay estimation, including the time delay estimation method of generalizedcross-correlation and adaptive time delay estimation method. Analyses of performance of variousmethods are presented in principle. Also it shows applicable occasions, merits and demerits ofvarious algorithms. A method of generalized cross-correlation time delay estimation based onneural network filtering is proposed on this basis.This paper discusses the basic theory ofartificial neural network and its superiority on detail, this method has better anti-noise andadaptive performance than that before. Moreover, we summarized several common localizationmethod based on time delay estimation and those occasions of application, which includingspatial mathematic geometric method, ball interpolation method, linear interpolation method, andsearch positioning method based on the Tabu.Finally, a practical3D acoustic source localization method is presented. We do experimentand analyze its performance.The experimental data shows that the positioning system is real-timeand feasible, positioning accuracy is basicly satisfactory.