Energy related activities recognition using smartphones

• Main Author: Lim, Sean Dao Chuan
• Other Authors: Soh Yeng Chai
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2020
• Subjects: Engineering::Electrical and electronic engineering
• Online Access: https://hdl.handle.net/10356/140381
• Institution: Nanyang Technological University

In recent years, there is increased ownership and reliance on the smartphone. The increased demand and competition from other market competitor has motivated smartphone producer to include more hardware and functions into the smartphone produced. This project explored the possibility of using microphone embedded within the smartphone to recognise home-related activities through the unique audio signature produced. Research done has shown that while sound recognition, specifically music and speech recognition, has been around for quite some time, the area of environmental sound recognition is relatively untapped. However, the basic methodology and idea to develop a classification model are similar. For this project, Mel-Frequency Cepstral Coefficients (MFCCs) are used for feature extraction and Convolutional Neural Network (CNN) is used as the preferred classification model. The outcome of this project was satisfactory, yet there is still room for improvements to the system in terms of accuracy as well as performance. This report will cover the research and development process of the project as well as suggestions for potential future development.