Context-Aware Indoor Positioning for Detailed Mobility Pattern Analysis in Aging Populations
| dc.contributor.advisor | Fang, Qiyin | |
| dc.contributor.author | Wang, Haixin | |
| dc.contributor.department | Biomedical Engineering | en_US |
| dc.date.accessioned | 2024-10-24T15:48:23Z | |
| dc.date.available | 2024-10-24T15:48:23Z | |
| dc.date.issued | 2024 | |
| dc.description.abstract | This thesis presents the development and evaluation of an enhanced turn-key indoor positioning system (IPS) for tracking the mobility patterns of older adults in residential settings. The design of the IPS hardware and software focused on usability in the context of aging-in-place, while maintaining high data quality, reduced incidences of missing data, and elevated room detection accuracy, with the highest accuracy reaching 99.47%. By integrating positional data with IMU sensors, this system not only captures precise locations but also identifies activity states and contextual information, establishing a detailed profile of mobility patterns. A 'floor filter' in the data processing models was developed to address vertical alignment challenges commonly encountered in multi-story dwellings. This adjustment improved prediction accuracies, with an average accuracy increase of 3.33% to 6.28% across various models. Among these, the Multi-Layer Perceptron Neural Network (MLP NN) and Shallow Neural Network (SNN) exhibited the highest accuracies for user room location predictions. Furthermore, we demonstrated the practical application of these technologies in a real-world setting through pilot clinical studies involving older adults. This study not only validated the integration of IPS and IMU data but also facilitated the establishment of behavioral trends that are crucial for context-aware analysis. The system's ability to adapt to different indoor environments without extensive setup, alongside its proven accuracy and reliability in capturing detailed mobility and activity information, underscores its potential to enhance elderly care and support aging in place. By leveraging advanced machine learning models and innovative data processing techniques, this work contributes to the field by offering a robust, scalable solution for monitoring the mobility patterns of the elderly, thus paving the way for future healthcare applications designed to accommodate the complexities of aging populations. | en_US |
| dc.description.degree | Master of Applied Science (MASc) | en_US |
| dc.description.degreetype | Thesis | en_US |
| dc.description.layabstract | This project develops a system that helps track the location and movement of older adults in their homes to support their independence and reduce stress on healthcare services. It improves on current technology by providing more accurate tracking inside the home. The system uses sensors to monitor how active someone is and understands the context of their movements—like whether they are resting or moving around—which helps in assessing both their physical and mental well-being. The results show that this technology is effective in tracking daily activities and can help in providing better care for older adults. | en_US |
| dc.identifier.uri | http://hdl.handle.net/11375/30464 | |
| dc.language.iso | en | en_US |
| dc.subject | Mobility | en_US |
| dc.subject | Aging-in-place | en_US |
| dc.subject | Context-aware | en_US |
| dc.title | Context-Aware Indoor Positioning for Detailed Mobility Pattern Analysis in Aging Populations | en_US |
| dc.type | Thesis | en_US |