A Co-Creative Drawing Robot for Freehand Sketching and Analogical Thinking in Early-Stage Design Ideation

This thesis explores how a collaborative sketching robot can better support designers during early-stage ideation. When a robot draws alongside a designer, its possible directions remain hidden until it commits strokes to the canvas. The designer can only react to what the robot has already done. They cannot see what it might do next. To address this, I designed and built the Spatial Ideation Canvas (SIC), a tablet-based system that sits alongside a CoFRIDA-based sketching robot. Before the robot draws anything, the SIC shows the designer what it could draw. The system offers four tools for this: previews of the robot's candidate directions, bio-inspired analogical search that surfaces unexpected framings, reflective questions that ask the designer to step back and think, and a side-by-side comparison view. I tested the system myself across nine sketching sessions in a single day. Each session followed a design brief drawn from Bruno Munari's philosophy. Three conditions were compared: sketching alone, sketching with CoFRIDA, and sketching with CoFRIDA plus the SIC. The order rotated across morning, midday, and afternoon blocks. The results showed a clear pattern. Scores for creativity support, collaboration, and design quality went up from Solo to CoFRIDA to CoFRIDA + SIC. More importantly, the session artifacts told a different story than the numbers alone. In CoFRIDA sessions, I kept refining the same idea. In SIC sessions, I moved across ideas. The concepts shifted visibly from one iteration to the next. The main takeaway is simple. When a designer can see where the robot might go before it goes there, the collaboration changes. It stops being a back-and-forth reaction and starts feeling like working together. This suggests that making the robot's process visible, not just its output, is a useful principle for designing future co-creative sketching systems.