Smart aging utilises innovative approaches and technology to improve older adults' quality of life, increasing their prospects of living independently. One of the major concerns the older adults to live independently is "serious fall", as almost a third of people aged over 65 having a fall each year. Dementia, affecting nearly 9% of the same age group, poses another significant issue that needs to be identified as early as possible. Existing fall detection systems from the wearable sensors generate many false alarms; hence, a more accurate and secure system is necessary. Furthermore, there is a considerable gap to identify the onset of cognitive impairment using remote monitoring for self-assisted seniors living in their residences. Applying biometric security improves older adults' confidence in using IoT and makes it easier for them to benefit from smart aging. Several publicly available datasets are pre-processed to extract distinctive features to address fall detection shortcomings, identify the onset of dementia system, and enable biometric security to wearable sensors. These key features are used with novel machine learning algorithms to train models for the fall detection system, identifying the onset of dementia system, and biometric authentication system. Applying a quantitative approach, these models are tested and analysed from the test dataset. The fall detection approach proposed in this work, in multimodal mode, can achieve an accuracy of 99% to detect a fall. Additionally, using 13 selected features, a system for detecting early signs of dementia is developed. This system has achieved an accuracy rate of 93% to identify a cognitive decline in the older adult, using only some selected aspects of their daily activities. Furthermore, the ML-based biometric authentication system uses physiological signals, such as ECG and Photoplethysmogram, in a fusion mode to identify and authenticate a person, resulting in enhancement of their privacy and security in a smart aging environment. The benefits offered by the fall detection system, early detection and identifying the signs of dementia, and the biometric authentication system, can improve the quality of life for the seniors who prefer to live independently or by themselves.
Date of Award | 2021 |
---|
Original language | English |
---|
- aging
- falls (accidents) in old age
- dementia
- diagnosis
- technology and older people
- Internet of things
- machine learning
Smart aging : utilisation of machine learning and the Internet of Things for independent living
Ahamed, F. (Author). 2021
Western Sydney University thesis: Doctoral thesis