Swift, Reactive Aerial and Ground Inspection of Industrial Facilities (SWIFT)
There is strong demand for autonomous inspection of industrial structures by aerial and ground robotic systems. A key goal for such robots is that the basic functions of mapping, localisation, exploration and path planning be carried out autonomously using onboard sensor payloads (LIDARs and camera). Such a mission involves autonomously taking off, carrying out a mapping or re-mapping mission and then returning to a base station. Additionally, for this technology to be advantageous to companies in the offshore industry, multiple robots will need to be deployed simultaneously by a single operator, which requires high levels of robot autonomy.
The project’s primary objective is to develop real-time visual mapping and inspection capability. This research strand will integrate Dr. Stefan Leutenegger’s world-leading work on dense, highly accurate object-level 3D reconstruction with ongoing ORCA work on large scale mapping. Aerial mapping systems will be provided by Imperial College London which will be paired with the University of Oxford’s terrestrial mapping capabilities. The sensor payloads and algorithms which will be used will overlap (thus driving collaboration) and the final demonstration will show coordinated mapping between the aerial and
A secondary objective involves Dr. Fallon’s contribution to the Cerberus Team currently competing in the international DARPA Subterranean Challenge (“SubT”, running until 2021). Cerberus is developing a team of six simultaneously-exploring aerial and legged robots to inspect and map in hazardous underground environments. Specifically Dr. Fallon’s team is contributing with their legged platform, ANYmal. Our intention is that alignment with the DARPA competition will help direct and accelerate the progress of the SWIFT project’s UK collaboration.
Lead Investigators: Dr. Maurice Fallon & Dr. Stefan Leutenegger