Multisensor Active Reverse Engineering

Abstract

No single sensor sees everything: each modality has its own reach, resolution and blind spots, and integrating several is usually treated as a post-processing problem. This talk treats it as an active one. Xiaoping Qian sets out a sensing-and-modelling loop for multisensor reverse engineering – an initial reconstruction, a seek step that decides where the next measurement should be taken, and an update that folds it back into the model – places it in the wider context of digital models in product realisation, and reports experimental results from the Computational Design and Manufacturing Lab at Illinois Institute of Technology.