Awareness and monitoring of personal environment to improve quality of living at home for the elderly

Jones, Nathan 2022. Awareness and monitoring of personal environment to improve quality of living at home for the elderly. PhD Thesis, Cardiff University. Item availability restricted.

Preview	PDF (Nathan Jones, Phd, Thesis) - Accepted Post-Print Version Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (2MB) | Preview
	PDF (Cardiff University Electronic Publication Form) - Supplemental Material Restricted to Repository staff only Download (101kB)

Abstract

As the average age of the population in the UK gets higher caution needs to be taken to ensure that elderly and vulnerable people are not left behind and can access the care and support that they need. This study aims to assess whether basic household sensors paired with a machine learning system can be used to act as an early warning system for carers, reducing carer workload while ensuring an equal level of care for their patients. To achieve this we have designed an anomaly detection system using a threshold system that aims to detect unusual behaviour in the property. Partnering with the Safehouse company we have access to two properties that have been outfitted with a range of environmental sensors. For the purpose of this study focusing on the temperature, humidity, light and motion in these property. Three time series prediction methods have been compared, these are LSTM, Arima and Autoencoder. A mixture of several important metrics were used to measure performance, namely Mean Absolute Error (MAE), Root Mean Square Error (RMSE) and R2, as well as expert feedback from the team at Safehouse. While we found the LSTM had the highest accuracy when making predictions on the data, particularly when using a combined sensor approach, due to the potential time taken to process data regularly it is not suited for a real world system. For this reason we suggest the use of an Autoencoder to make predictions on the data, with extra parameters to assist in identifying important anomalies. Further research involving direct contact with carers is needed to discover whether this system is sufficient to their needs. Stricter monitoring of the people within the properties is also necessary in order to decide whether the anomalies detected identify all relevant real world problems.

Item Type:	Thesis (PhD)
Date Type:	Completion
Status:	Unpublished
Schools:	Computer Science & Informatics
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Funders:	KESS 2
Date of First Compliant Deposit:	7 October 2022
Date of Acceptance:	24 August 2022
Last Modified:	07 Oct 2023 01:30
URI:	https://orca.cardiff.ac.uk/id/eprint/153142

Actions (repository staff only)

Edit Item