Ophélie Jobert

Abstract:

Shape-changing user interfaces (SCIs) have emerged as a new way of interacting with computers, using dynamic changes in physical shape as both output and input. In order to achieve a wide range of possible shape changes, one type of SCI is of particular interest: modular robotic interfaces. Modular robotic interfaces consist of several modules that can rearrange themselves, autonomously or not, enabling them to achieve a multitude of shapes and to overcome the lack of porosity and modularity of other types of SCIs. However, to achieve high precision in haptic and visual feedback, the modules tend to become as small as possible (miniaturization), leading to their multiplication. On the one hand, miniaturization leads to technical challenges. Prototyping modular robotic interfaces, as with many other SCIs, becomes more complex and time-consuming, delaying the availability of usable prototypes. Yet these interfaces still offer a large space to propose novel interaction techniques. To benefit from innovation, initial user studies must be conducted as early as possible. This thesis explores early-stage prototyping techniques that can be applied before actionable physical prototypes are available. On the other hand, the multiplication of modules leads to manipulation challenges for users. Robotic modules can help overcome these challenges. The system is no longer passive and can assist users in their tasks. The interface then shifts from being a tool to being a partner to accomplish users’ tasks. This assistance, coupled with direct manipulation of the modules, could affect the user experience: too much assistance can lead to a loss of users’ control and lower their sense of agency, which refers to the feeling of control over one’s own actions and their effects on the environment. A lower sense of agency can prevent users from accepting the system and from taking responsibility for the consequences of the system’s use. This thesis identifies the possible types of assistance and their impact on users’ sense of agency. 

This thesis therefore addresses two main research questions : (1) "How can we explore the interaction techniques of an SCI when it is not easily physically available?", and (2) "What are the criteria for optimal agency while benefiting from the robotization of modular robotic interfaces?". To answer these questions, we first conducted a state-of-the-art review of SCIs, and more specifically of modular robotic interfaces, then of the different types of prototypes and their limitations, and finally of the impact of different types of assistance and cooperation on agency. We then conducted two types of studies: First, we explored the eXtended reality (XR) prototyping approach for exploring interaction techniques. We compared the interaction techniques proposed by participants on a physical prototype and on its XR counterpart. Then, we conducted a study exploring the links between users’ control, depending on system assistance, and the users’ sense of agency, as well as the potential effects of task difficulty and system predictability.