Abstract
Multimodal interaction techniques using gesture and speech offer architects and engineers a natural way to create 3D CAD (computer-aided design) models for conceptual design. Gestures and speech for such interfaces must be based on empirically grounded knowledge instead of arbitrary vocabularies as employed in extant literature. We conducted an experiment with 41 participants from architecture and engineering backgrounds to elicit preferences of gestures and speech for 3D CAD modeling referents such as primitives, manipulations, and navigation. In this paper, we present results from the thematic analysis of the elicited gestures. We present a compilation of gestures, which were evaluated by experts to be suitable for the articulation of 3D CAD modeling referents. We also present a set of speech command terms elicited from participants. Finally, we provide recommendations for the design of a gesture and speech based CAD modeling system for conceptual design.
•We present a compilation of user-elicited gestures and speech terms for CAD modeling.•Most participants used bimanual gestures for navigation and manipulation tasks.•Most participants used unimanual gestures for creating open ended freeform.•Participants rated most tasks easier when they used both gestures and speech.