Future CAD systems could speed up the design process by tracking designers’ eye movements with a camera as they work.
Researchers from Leeds University and the Open University are investigating how identifying what part of a drawing a designer is focusing on could enable CAD packages to suggest suitable shapes for them to use.
Alison McKay, professor of design systems at Leeds University, told The Engineer that the idea was for the software to act like a design assistant.
‘It wouldn’t be doing the creative part of design but it would say: “If you’re taking the shape transformations in this direction then have you thought of these other ways in which you could take it?”
‘[To do this] we need to know which bits of the shape the designer is interested in. So one of the big challenges in implementing these systems is finding out what the designer is looking at.’
The team has spent the last two years researching how designers observe and identify shapes, using a camera to record parameters such as where the designer’s gaze falls and how frequently and then predicting which shapes they were most interested in.
The researchers envision that an intelligent CAD system could collect this information and feed it into a set of rules that would use the different elements of what a designer was drawing — the equivalent of grammar in language to suggest new shapes.
‘The sorts of places that would help is early on in product development, for example, where people are trying to design in the style of a brand,’ said McKay.
This is because in the early creative stages of work, designers often develop shapes that they didn’t intend to draw but begin manipulating once they have noticed them within the design.
The team is now developing a proposal for an EU-funded project that may include an eye-tracking CAD system, depending on interest from commercial partners.
McKay said that a system for developing 2D abstract shapes incorporating eye-tracking could be developed in around four to five years.
The two-year ‘Designing with Vision’ project was funded by a grant of around £200,000 from the Leverhulme Trust.