Prototyping methods are commonly employed iteratively throughout the design and product development, typically ranging from early low-fidelity to later high-fidelity prototypes. We present a case study focusing on prototyping a receptionist agent migrating between three platforms (a monitor on the wall, a mobile phone, and a physical robot). More specifically, we compare virtual reality (VR) and physical (real-world) prototyping methods. The two methods are compared in terms of fidelity and usability. The breadth of features, the degree of functionality, and the interactivity were similar. However, the aesthetic refinement differed. The VR prototyping method also had much higher prerequisites in terms of equipment and skills, and the learning curve for the designer was steep. Both methods were equally efficient in user testing, but the VR method revealed more usability issues in the efficiency category, while the physical space method revealed more issues in the effectiveness category.

A migrating receptionist is an intelligent virtual agent (IVA) designed to transfer from one embodiment to another. The embodiment can range from a virtual human-like form to a physical robot. We present a study of the prototyping and evaluation of a migrating virtual receptionist interface. One prototype was made in virtual reality (VR) and one in physical space. The study looks at how to design a migrating virtual receptionist for a department of computer science. The virtual receptionist was designed to migrate from a screen by the entrance, to a mobile phone used while walking, and then to a robot in the coffee area. The evaluation was conducted with 23 participants who interacted with the virtual receptionist using both prototypes. Data was collected with the System Usability Scale (SUS) and Microsoft Reaction Cards. Although the physical space and the virtual space were different, the results of the evaluation are similar. The SUS mean was 83.9 for the physical space and 78.3 for the virtual space. The reaction cards also showed more positive than negative reactions for both prototypes.

Early user involvement in the design of intelligent virtual agents (IVAs) is fundamental for usability and otherwise good user experience. In this paper, we present a combination of methods used in the remote conceptual design of a virtual receptionist for a university department. The study builds on two workshops with potential users. The first was a bodystorming workshop in virtual reality (VR) with four researchers at the department, and the second was a desktop walkthrough workshop on an online whiteboard with five students at the department. Proposed solutions from the workshops were deconstructed using a morphological chart into a pentad of parameters: agent, act, scene, agency, and purpose. New design concepts were then composed by combining solutions. Sketching was used to further detail and present the generated concepts. Our analysis of the workshops indicates that the bodystorming workshop had an aesthetic perspective on embodied interaction while the desktop walkthrough workshop had a more instrumental perspective on usability. The combination of embodied but remote ideation methods with morphological chart structured by the pentad is novel to not only the IVA field, but also to interaction design in general. Finally, the conceptual design of a novel cross-platform IVA is proposed.

Cognitive Work Analysis (CWA) is an appropriate approach in design for high-stakes domains, such as air traffic management (ATM) since it focuses on human expert performance in regular and contingency situations. However, CWA is not suitable for the design of a first-of-a-kind system since there is nothing to analyse before the start of the design process. In 2017, unmanned air traffic management (UTM) for intense drone traffic in cities was such a system. Making things worse, the UTM system has to be in place before the traffic, since it provides basic safety. In this research-through-design study, we present conceptual designing as a bootstrapping approach to CWA in the design of a first-of-a-kind UTM system. In a series of co-design workshops, we identified future services, traffic patterns, and regulations that framed the design of UTM system concepts. They were based on combinations of four basic building blocks: points, lines, planes, and volumes. Concepts of point-based control, airport geofences, grid squares, layers, and tubes were discussed. Throughout the conceptual designing, results were documented in an evolving Work Domain Analysis (WDA), which is a cornerstone of CWA. This approach allowed us to bootstrap the CWA for a first-of-a-kind-system.

Several Wizard-of-Oz techniques have been developed to make robots appear autonomous and more social in human-robot interaction. Many of the existing solutions use control interfaces that introduce significant time delays and hamper the robot operator's ability to produce socially appropriate responses in real time interactions. We present work in progress on a novel wizard control interface designed to overcome these limitations:a motion tracking-based system which allows the wizard to act as if he or she is the robot. The wizard sees the other through the robot's perspective, and uses his or her own bodily movements to control it. We discuss potential applications and extensions of this system, and conclude by discussing possible methodological advantages and disadvantages.

The designed product is often assessed in interaction design education, but there are also courses that focus on learning the design process. It is then necessary to develop criteria for grading in such courses. To make a successful transfer from theory to practice, students also need to learn the criteria practitioners use, rather than the criteria that academically oriented teachers use. To do this, one approach is to align criteria with the conceptions practicing interaction designers have of process quality in design. Therefore, the research questions for this study are what those conceptions are, and how they can be utilized in grading criteria for interaction design projects in education. Interviews were made with 10 interaction designers. The interviews were qualitatively analyzed. The results demonstrate that practicing interaction designers conceptualize the quality of the design process in three ways: it is good if established methods are used and the design is managed within resource constraints, and within organizational and technological limitations, while also meeting stated objectives; it is even better if the design has a thought-through rationale; and ideally, the design should also be inspirational. These conceptions were transferred to points on a criteria-referenced grading scale which was used to develop course specific grading criteria. The criteria were evaluated in terms of comprehensibility and reliability. The evaluation showed that most of the students who also attended lectures understood the criteria. A high and significant covariation and a high level of agreement between the two teachers who graded the projects were shown. Further, the developed criteria should be generalizable to other process-centered interaction design courses and to assessment in other design disciplines.