Abstract
The Play-Doh to Roombots project is a bachelor semester project done by Radosław Dryzner and supervised by Mehmet Mutlu, Dr. Stéphane Bonardi, Simon Hauser and Prof. Auke Ijspeert. The aim of this project and the software it deliver is to allow the user to construct Roombots structures out of real world objects such as Play-Doh, but also any 3D mesh. This consists of scanning a Play-Doh sculpture into the software and using different approaches to reconstruct it using Roombots. The resulting structure can as well be bounded in real world dimensions if the user wishes to do so.
During the development of the software, multiple approaches were used with varying performance of time and quality of reconstruction. These approaches consist of artificial intelligence search methods such as Breadth-First-Search, Depth-First-Search and Greedy algorithms on one hand, as well as a direct voxelization method on the other. These modular approaches leave the possibility for extension in the future.