Final Thesis

Machine Learning for pass-by vehicle sound classification

Master Thesis of Tegetmeyer-Kleine, Thorsten

Can the computer tell us which car has passed by? Imagine a scenario where it would be possible to tell what kind of car it is just by listening to a recording without having seen it! A suitable technology used to answer this question is machine learning. The training of the analysis algorithm plays a crucial role and decides about the quality of the result. As in sports, not only the precision but also the frequency of repetitions decides on the success of the training. This work includes the training of a machine learning algorithm with the open-source software VA (http://www.virtualacoustics.org) developed at ITA. Subsequently, the result shall be evaluated in practical tests.