Designing for Intelligibility and Control in Ubiquitous Computing Environments

Ph.D. Computer Science, Hasselt University

Jo Vermeulen
Prof. Dr Karin Coninx
Prof. Dr Kris Luyten
11th December, 2014

2015 FWO-IBM Innovation Award

2015 FWO-IBM Innovation Award

My dissertation was awarded the 2015 FWO–IBM Innovation Award. This prize is awarded each year by Research Foundation Flanders (FWO) and IBM in recognition of an outstanding PhD thesis that presents an original contribution to informatics or its applications. It is Belgium's oldest and most prestigious scientific award for research in computer science and its applications.

© FNRS/Aude Vanlathem



Dissertation (low-res, 9 MB PDF)
Dissertation (high-res, 67 MB PDF)


In this dissertation, I explore designing for intelligibility and control in ubiquitous computing applications, in particular, in context-aware systems. Earlier work has identified a number of interaction challenges that users face when dealing with context-aware systems, such as being unable to understand why the system is responding in a certain way, or being unable to intervene when the system makes a mistake. These challenges impact users’ feelings of trust and of being in control, and could eventually lead to users disengaging from interaction with the system altogether. It is important to address these interaction challenges to allow for smooth integration of ubicomp technologies into our lives. For this purpose, two general principles have been proposed that should be supported by context-aware systems: intelligibility and control. Intelligibility is the ability of a context-aware system to present itself and its behaviour to its users, while control deals with allowing users to step in and intervene to correct the system. Although a number of techniques have been explored to improve intelligibility and control for context-aware systems, it is not yet clear how ubiquitous computing researchers, developers and interaction designers should design for intelligibility and control.

This dissertation serves as a design space exploration to inform the design of future ubiquitous computing applications that provide support for intelligibility and control. I aim for this work to be both generative by guiding designers in exploring various ways to support intelligibility and control, and generalizable by exploring techniques that can be applied by interaction designers in a wide range of ubiquitous computing scenarios. In particular, I provide the following three major contributions:

First, I present a design space that captures different decisions that designers face when adding support for intelligibility and control. This design space consists of six dimensions and can be used both as an analytical tool to classify and compare different techniques, and can help designers explore alternative designs.

Second, I present general design principles and techniques that can be applied in a wide range of ubicomp scenarios. I contribute three general techniques that can be used at three different times during the interaction: feedforward (before actions), slow-motion feedback (during actions), and why questions (after actions).

Third, I describe an in-depth case study of supporting intelligibility and control in proxemic interactions. This can serve as inspiration for designers and researchers looking to consider these principles in different ubicomp applications. In particular, I propose the use of a secondary, assisting floor display that informs users about the tracking status, indicates action possibilities, and invites and guides users through- out their interaction with the primary display.