Clustering of data recorded by Distributed Acoustic Sensors to identify vehicle passage and typology

There is an increasing interest in researchers on the use of modern sensor networks deployed in smart cities to collect data that can be used for a better man- agement of the transportation systems, and, more generally, to improve its impact on the environment. A fairly recent technology, Distributed Acoustic Sensors is be- ing used more and more in the field of transportation system, especially in the field of traffic management. In this paper we propose a methodology for exploiting the data of moving vehicles, captured through these sensors to detect and classify the types of passing vehicles. The methodology is based on a data processing pipeline, whose purpose is to exploit signal processing algorithms to clean the data and to use a clustering algorithm to detect the types of vehicle. The data used for testing the methodology is collected through a series of experi- ments with the DAS technology, in a real city environment.