Urban noise classification of active noise control system for residential buildings

In this report, various audio signal features are extracted and combined into different sets to be examined. Feature set that produces highest accuracy is to be chosen as optimal features applied in support vector machine (SVM) classifier to classify noises around residential buildings. The designed noise classification system is the premier approach to provide relevant coefficients for active noise control filter in an active noise control (ANC) system. Based on the experiment conducted in this paper, the ultimate trained SVM classification model classifies noises that can reache an accuracy around 95%.Bachelor of Engineering (Electrical and Electronic Engineering