Modeling touch gestures to propose optimal design guidelines based on human performance measures

Abstract

Touch interface has evolved into dominant interface system for smartphones over the last 10 years. This evolutionary process has been applicable not only to the smartphone, but also to small hand-held smart devices like portable game consoles and tablet devices. Even further, the most recent Microsoft Windows operating system supports both traditional point and click interface as well as touch interface for broader coverage of OS on digital devices. Identifying factors contributing the human performance on touch interface system has been studied by wide range of researchers globally. Designers and manufacturers of smart devices with touch interface system could benefit from the findings of these studies since they may provide opportunities to design and implement better performing and more usable product with competitive edge over competitors. In this study, we investigated factors affecting human performance on touch interface systems to establish practical design guidelines for designers and manufacturers of smart devices with touch interface system. The first group of factors is demography related variables such as gender, regions and age. The second group of factors is interaction related variables such as number of hands involved in interacting with touch system – one handed versus two handed postures. Finally and most importantly, design-related variables such as sizes, shapes or locations of touch targets are investigated. Our main goal of this study is to identify what are the most affecting factors to human performance of touch interface systems and establish mathematical modeling among them. Developed performance modeling will be leveraged to estimate expected human performance without conducting usability testing on given touch interface system. Once demography, interaction and design related variables are given, we will be able to propose expected performance level by inputting those variables into the established model, thus will contribute to the optimal design practice. Touch gestures considered in this study are tap touch, move touch and flick touch, which are the most widely used touch gestures in designing and implementing touch interface system. We have recruited 259 subjects from 4 major metropolitan areas across 3 different countries – New York, San Francisco, London and Paris and conducted controlled laboratory experiment. In order to assess human performance of each touch gesture, we have defined individual performance measures of each gesture such as task completion time, velocity, throughput introduced by Fitts law (Fitts, 1954), variance/accuracy ratio introduced by Chan & Childress (1990), accuracy or offset tendency from a desired line of target. By investigating these performance measures, we could come up with design guidelines about design specifications such as size and movement direction as well as qualitative insights on how touch gestures are different across all the factors we have gathered from the experimental setup. Design strategies and guidelines as well as human performance modeling will contribute to develop effective and efficient touch interface systems.