Navigating Deformable Spaces – How to Localise in a Shifting World. This project proposes novel methods that will enable robotic systems to navigate safely and precisely within deformable environments

Description

Navigating Deformable Spaces – How to Localise in a Shifting World. This project proposes novel methods that will enable robotic systems to navigate safely and precisely within deformable environments, where current technologies fall short due to assumptions of static world models. These limitations often lead to unintended collisions, poor navigation behaviours, and increased uncertainty in robotic actions. By advancing the ability to model and interact with a deformable world through machine learning, vision-based foundation models, and rule-based explanations, this project will open new avenues for deploying robotic systems in unconventional scenarios, from environmental disaster recovery to advanced healthcare applications such as robotic-assisted arthroscopy.. Scheme: Discovery Projects. Field: 4603 - Computer Vision and Multimedia Computation. Lead: Prof Will Browne