Human activities recognition in smart living environment
Smartphones with embedded sensors can track users' movements and corresponding activities in real-time. This serves as the foundation for the recognition of human activity in smart environments where users can accurately monitor their lifestyle and physical health. There are various different w...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
Nanyang Technological University
2023
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/166757 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | Smartphones with embedded sensors can track users' movements and corresponding activities in real-time. This serves as the foundation for the recognition of human activity in smart environments where users can accurately monitor their lifestyle and physical health. There are various different well-known machine learning techniques that can process this sensor data to derive high-level information about users, including their locations, activities they are engaging in, and the associated energy devices they are using.
Through an integration of these sensor data from mobile devices, this project aims to build an integrated information system addressing the analysis of basic and complex behaviors done by the users in a smart environment and their usage of energy-related devices. Long Short-Term Memory (LSTM), one of the more well-known machine learning approaches, is used in this project. With the model built using LSTM, an Android application is developed to verify the model's accuracy in real-time. As a part of Human Activity Recognition (HAR), an audio classification is also implemented, where the model is able to classify the sound generated by the different activities performed. |
|---|