Analysis of approaches and methods to acoustic sources localization

This article provides an overview of traditional methods to acoustic sources localization based on signal processing, as well as modern methods based on the use of deep neural networks. The advantages and disadvantages of the above methods are analyzed and discussed. Although some traditional methods can adapt to observed signals, they all depend on assumptions made about the nature of the environment, the properties of the signals, etc. Deep learning models do not explicitly require any of these assumptions, but instead efficiently adapt to the training data provided. However, this is also a major disadvantage of modern methods, as they are less generalizable and less versatile than traditional methods. A justification is given for the need to develop new localization methods, as well as the integration of traditional and intelligent modern localization methods to combine the advantages of each of these groups of methods.