In a first stage, the project will be focussed on analyzing techniques in the tasks of detection, localization and classification of sound sources, with two focuses: 1) On the one hand, the study of novel techniques for classification, separation and localization, which are able to improve the performance of actual systems in terms of quality anon the other hand, the search for integration methodologies that allow to improve the integration of the different subsystems in terms of fiability, quiality and performance. For instance, an aditional knowledge about the typology of the sound source could allow an improvement in the localization and separation algorithms, and aditional knowledge about the relative position of a source could help to classify the sound event.

With these two items, we are determined to obtain a set of different algorithms with more or less complexity that allow to solve the five big problems in the analysis of acoustiv scenes: localization, separation, classification, tracking and environment analysis.

In a second stage of the project, we will focus on the implementation of technical solutions in smart sensor networks, bearing in mind with the design of the autonomous nodes for acoustic applications, among others. A heterogeneous smart network with low consume nodes will be implemented, including high level information extraction systems in order to tackle decisions by fusion information and cognitive systems.

The third stage of the project will be focussed on the application of the techniques to two different cases of study: the field of body Area Networks, for audiological applications, and the detection of violence and endangering events.