Validating the Gibsonian Human-Robot Interface Design Taxonomy (GHRID): An Exploratory Study in Autonomous Vehicle Interfaces

This thesis investigates whether Gibsonian Human–Robot Interface Design (GHRID) can improve the way humans understand the perceptual and action-related states of autonomous systems. Building on ecological psychology, the work argues that robot interfaces should not merely communicate internal states through symbolic cues, but should make the robot’s perceived environment, action possibilities, and constraints directly perceivable. The theoretical basis of the thesis is the GHRID taxonomy, which distinguishes between more symbolic and more ecological forms of interface communication and identifies key dimensions such as information type, affordance grounding, Umwelt exposure, and temporal coupling. To empirically explore this framework, a within-participants study was conducted using two autonomous vehicle interface stimuli: one classified as highly Gibsonian according to the taxonomy and one classified as non-Gibsonian. Participants evaluated both interfaces through measures of usability, trust, workload, situation awareness, perceived transparency, affordance understanding, and open-ended interpretations of the vehicle’s intentions. The results suggest that the GHRID-high interface supported a clearer understanding of the vehicle’s perceptual world and possible actions, providing preliminary evidence for the usefulness of Gibsonian principles in the design of transparent human–robot interfaces. Overall, the thesis contributes to HRI by offering both a theoretical framework and an initial empirical validation of ecological interface design as a promising direction for improving human understanding, anticipation, and coordination with autonomous systems.