Enabling Affordable and Efficiently Deployed Location Based Smart Home Systems
By Kelly, Damian; McLoone, Sean; Dishong, Terry; Technology and Health Care, Vol. 17, pp. 221-235Publication Date: 2009
Article outlines the development of a location sensing technique with minimal hardware requirements which can reduce the installation overhead of location based smart home systems for monitoring older residents in their homes. The location sensing technique utilizes the signals from the resident’s mobile phone, arriving at a single Bluetooth enabled base station computer. Two types of mobile devices were compared in a home environment consisting of 4 rooms. It was found that both provided similar levels of accuracy. Where cost is an issue and a prompting system is not necessary, a cheaper device such as a Nokia 6230 can be used with similar localization accuracy. However, when prompting and localization functionality is desirable in the same device, the N95 type device is preferable due to its more stable programming environment. Along with the ability to reliably provide prompts such as medication reminders to the user, the Nokia N95 has further sensors available to it, which can be used for any number of monitoring activities, including detecting whether the mobile device is being carried by the user, confirming the impact of a prompt at any given time. Four location prediction algorithms were compared for localization accuracy: k-Nearest Neighbor (KNN), Linear Discriminant Analysis (LDA), Quadratic Discriminant Analysis (QDA), and Support Vector Machines (SVMs), of which LDA and QDA proved to be the most accurate. Implications for further development of the location sensing technique for smart home systems are discussed.
Published by: IOS Press (Website:http://www.iospress.nl)
Link to text: http://dx.doi.org/10.3233/THC-2009-0549
