Speech analysis for conceptual CAD modeling using multi-modal interfaces: An investigation into Architects' and Engineers' speech preferences. Issue 3 (14th March 2019)

• Record Type: Journal Article
• Authors: Khan, Sumbul; Tunçer, Bige
• Is Part Of: AI EDAM. Volume 33:Issue 3(2019)
• Page Start: 275
• Page End: 288
• Publication Date: 2019-03-14
• Subjects: Computational linguistics, -- computer-aided design, -- conceptual design, -- human–computer interaction, -- natural interfaces, -- speech analysis; Engineering design -- Data processing -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; 620.00420285
• DOI: 10.1017/S0890060419000015
• Languages: English
• ISSNs: 0890-0604

Abstract

Abstract: Speech- and gesture-based interfaces for computer-aided design (CAD) modeling must employ vocabulary suitable for target professional groups. We conducted an experiment with 40 participants from architecture and engineering backgrounds to elicit their speech preferences for four CAD manipulation tasks: Scale, Rotate, Copy, and Move. We compiled speech command terms used by participants and analyzed verbalizations based on three analytic themes: the exactness of descriptions, the granularity of descriptions, and the use of CAD legacy terms. We found that participants from both groups used precise and vague expressions in their verbalizations and used a median of three parameters in their verbalizations. Architects used CAD legacy terms more than Engineers in the tasks Scale and Rotate. Based on these findings, we give recommendations for the design of speech- and gesture-based interface for conceptual CAD modeling.