With improvements in living standards and innovations in medical care, life expectancy has increased. However, although people are living longer, particularly in developed countries, they are not necessarily healthier during the additional years of life, with a rising number of people with long-term physical and mental health conditions that require supported living, for example, within a care home or hospital environment. In response to the rising economic costs of managing long term conditions, successive governments have developed policies to reduce the use of institutional environments (e.g., care homes) and of unplanned hospital admissions, and are encouraging the development of systems which aim to monitor, support and manage people’s health in their own home.
These developments have lead to increased research on using remotely monitored, sensor-based technologies to provide relatives, carers and health care professionals with timely data about the well-being of older people living independently, and so provide timely and appropriate support effectively, thus helping them remain in their own homes, especially when they have long-term health problems.
The aim of the research described in this thesis was to investigate the use of an electricity monitor to recognise and monitor changes in resident’s daily activities. This was achieved using two phases; the first conducted a survey to gather information about which activities and features that carers and relatives would like to have access to, so as to be reassured about their relative’s health and well being. The second phase collected and analysed electricity consumption data from four households for a one-week period, to develop models to identify when specific activities had been undertaken, e.g., using the shower, using a kettle.
This research concluded that the monitoring of general and some specific activities is important to the relatives and carers, although the best form of reassurance about their relative’s situation was felt to be human contact. Following the analysis of the electricity consumption data, it was concluded that while it is possible to recognise appliance usage from whole house electricity consumption data, the variability and lack of transferability between houses and appliances would mean that the large-scale use of this type of monitoring would require considerable further development.
