FS25-Talk1: Dr. Olga Vysotska, Robotics and augmented reality localization in construction sites

Abstract:
Robotics and augmented reality (AR) can significantly improve construction productivity by assisting workers with mapping, inspection, and documentation tasks. Robots are well-suited for repetitive jobs such as mapping and updating site information, while AR devices allow users to view and interact with 3D holograms directly in the physical environment. For effective collaboration, both robots and AR devices must be precisely localized in a shared map of the site, a task that is especially challenging in construction, where environments are “feature-poor” and constantly changing. Unlike static indoor spaces, construction sites evolve rapidly, making traditional localization methods less effective. To address this, this project incorporates prior knowledge from Building Information Modeling (BIM) or previous scans of the building to enhance localization by leveraging geometric layout information. Additionally, I explore how using sequences of observations, rather than single images, can reduce ambiguities in environments with many perceptually similar spaces, enabling robust long-term localization for construction documentation.

Short Bio:
Olga Vysotska is a Postdoctoral Fellow with Design++. She is also a PostDoc with the Robotics System Lab (RSL) and Computer Vision and Geometry Lab (CVG) at the ETH Zurich. Olga’s research focuses on enabling the localization and autonomous operation of mobile robotics systems and augmented reality devices on construction sites. Olga holds a PhD from the University of Bonn, Germany, in the topic of “Visual place recognition in changing environments”. Before joining ETH, Olga worked as a research scientist in the Mixed Reality and AI Lab at Microsoft and several robotics startups. In this talk, she will present the results of a 2-year research project about the robot and augmented reality localization in construction sites.