Composition du jury :
- M. Jean Vanderdonckt, prof. des universités, Université Catholique de Louvain (Rapporteur)
- M. Christian Bastien, prof. des universités, Université de Lorraine (Rapporteur)
- M. Patrick Reigner, prof. des universités, Grenoble INP (Examinateur)
- M. Laurent D’Orazio, prof. des universités, Université de Rennes (Examinateur)
- M. Guillaume Rivière, maître de conférences, ESTIA (Examinateur)
- M. Yann Laurillau, maître de conférences, HDR, Université Grenoble Alpes (Directeur de thèse)
- Mme Gaëlle Calvary, prof. des universités, Grenoble INP (Co-directrice de thèse)
This thesis is about persuasive interactive systems, a subclass of interactive systems aiming at supporting sustainable behaviors. Persuasive technology is a very recent and emerging topic, which relies on persuasion to support human behavior change voluntarily without coercion and deception.
In this research, we focus the design of persuasive system for promoting behaviors change with energy as application domain. In the interest of preserving the planet, reducing domestic energy use is necessary for fighting against global warming and climate change issues. The approach to save energy by introducing more efficient appliances only tackles partially the current issues. An urgent need is to motivate sustainable behaviors in energy usage. However, changing people behaviors involves a complex and difficult process.
The challenge is to design interactive systems that take into account this process dimension and that keep users involved over time in the process of behavior change. Consequently, based on a review of classifications, design spaces and frameworks targeting the design of persuasive systems, we propose the UP+ framework, our first and conceptual contribution. It considers three dimensions: two related to the process aspect of behavior change, at two levels: micro (cause-effect-causality) and macro (long term)); one related to the psychological aspects of motivation.
Based on the UP+ framework, we conduct a state of the art of existing persuasive systems for energy. It reveals that most of these systems covers a few persuasive functions only (e.g., self-surveillance or eco- feedback). In this work, we particularly focus on decision and causality of behavior. A second contribution is the Mondrian User Interface, a proof-of-concept of UI designed to sustain daily use and to keep inhabitants motivated in the long run. We consider three goals: to catch user’s attention; to offer multiple levels of user interaction depending on the context of use; to allow the integration of interactional bricks supporting persuasion. Particularly, we propose two bricks, Sliders4DM and Plan4Actions.In order to assist inhabitants in preparing a change (decision), Sliders4DM is a novel widget allowing non- specialists to find an appropriate trade-off between (possibly) conflicting criteria in the home via what-if approach. Sliders4DM is evaluated with two experiments: a qualitative one (16 participants) and a quantitative comparative one (177 participants). Plan4Actions is a novel concept of user interface for planning daily actions. The concept is empowered by the co-decision between inhabitants and the home management system. It provides inhabitants with flexibility in planning their daily actions in order to satisfy their objectives in terms of comfort and sobriety. A twofold evaluation presented a favorable assessment from 13 participants about Plan4actions’ comprehension, its usability and potential utility in the domestic context.
We also discuss the limitations related to persuasive interaction covered by current works and the needs for a long-term evaluation to mesure the behavior changes. The thesis opens new perspectives for extending current research of PIS for energy. To conclude, we define our future works, which involves short-term improvements for current prototype and its deployment in real domestic context.