Abstracts Track 2022

Area 1 - Human Factors and Information Systems

Nr:	1
Title:	Windshield Display: Visually Augmented Driving at Low Cost
Authors:	Karlheinz Blankenbach and Matthias Eisenhardt
Abstract:	Interior lighting is prospering in modern cars. This enables new chances for ergonomic visualization of important data such as speed, warnings, wayfinding support and comfort functions (e.g. incoming calls). Today's systems are build by a single pixelated lightguide on top of the dashboard and the bottom of the windschield. This is called "Schweibenwurzel-Display". In a first project we evaluated the required optical parameters such as luminance and RGB luminance ratio. Eg. a luminance larger than 3,300 cd/m² is required. This paper describes thge next step with 8 "lines" for such a display so that numbers (speed) icons (simple traffic signs), trajectory etc. can be reproduced. Our evaluations in a mock-up provide this approach as very ergonic with outstanding feedback. The time of "eves of the road" is significantly reduced and warning in preipheral vision is perceived.

Nr:	2
Title:	Empowered Operators: How to Foster the Acquisition of Robotic Programming Skills?
Authors:	Charly Blanc, Julien Nembrini and Sarah Degallier Rochat
Abstract:	Humans and machines speak different languages. In order to make the interaction possible, interfaces for operators typically translate the machine language in a more abstract, task-oriented one. The consequence is that the operator has little understanding in the process, which can lead to a lack of ownership and a feeling of dehumanization (Kinzel, 2016). In order to empower operators in their interaction with robots, we are developing a three-layer interface that enable them to progressively gain control over the system, as proposed in the literature (Schou et al., 2018). The top layer (tasks on Figure 1) is based on the usual, task-oriented language, while the intermediate layer (skills) is based on the procedure-oriented language typically used in intuitive interfaces for collaborative robotics. The bottom layer (device primitives) corresponds to the commonly available robotic commands. The top layer is the most intuitive, but also the less flexible. The intermediate layer requires some basic, high-level understanding in robotics and programming, and the bottom one, a deeper knowledge in these fields. Our goal is to encourage the user to gain control over the system by progressively learning the different languages, while maintaining a feeling of safety. In order to do so, we take a Blockly-like approach (Weintrop et al., 2017) where the user can easily switch between the different languages and combine them. In addition we provide some tutorials that aim at developing the required understanding to master the lower languages. A technology probe proposing only the top layer was showcased in the New Museum of Biel in Switzerland during an exhibition on robotics. It is also available for testing online (see Appendix A). The user was guided through different steps to program the robot, as shown in the provided video (see Appendix B). A short survey (N=15) showed that the ease of use of the interface provided a feeling of safety and control over the machine, but also that most users had a feeling of frustration and wanted to understand more how the system was working. A set of workshops is planned to test the whole architecture with target users. Two modes will be tested: one where the tutorials are mainly done in simulation and the user only interacts with the robot once (s)he is familiarized with the interface, and another one where the user directly interacts with the physical system. Our hypothesis is that simulation may induce more epistemic behaviors (the user can test ideas without fearing to break the robot), while real interaction may lead to a better understanding of the robot constraints. We will study the perceived self-efficacy and the stress level in these two situations.