Graphical user interface customization for the elderly in ambient assisted living environments

Abstract

With the help of the graphical user interfaces (GUI), young people or everyday-computer-users can easily get accustomed to the various digital devices. However, acceptability of GUIs by elderly, as the rapidly growing group of todaýs population, is a challenging issue due to their little experience with technology and experienced impairments at different levels, such as visual, audial, psychomotor and cognitive impairments. Providing simplicity for user interfaces in order to be acceptable by elderly people is the main concern of the studies that try to address this challenge. Thus, designing simple and hence acceptable user interfaces for elderly requires considering their preferences and impairments. In this thesis, we study GUIs for elderly people with different impairments in the context of in-home healthcare system that involves daily living monitoring, focusing on the relation between two main design parameters of a graphical user interface: page complexity which is the number of interface elements on each page and the page hierarchy which is the number of the pages to be traced in order to finish a task. In order to study the dichotomy between these two phenomena, we designed two versions of the interface: One version is the flat version which has less page hierarchy with more complex pages and the other version is the deep version which has deeper page hierarchy with less complex pages. We conducted experiments with 18 elderly people having different impairments in order to understand whether the flat version or the deep version would be more acceptable. We asked them to complete some tasks such as making a video-call to a friend and setting a reminder to the agenda. Using the errors, time, inter-touch time between two consequent interactions, key strokes and clicks as the experiment metrics, the results revealed that the flat version with a reasonable complexity level would be more acceptable by elderly.