MATEC Web of Conferences (Jan 2023)

Reliability of visual inertial odometry on an unmanned aerial vehicle

  • Setati Tiro,
  • Maweni Thabisa,
  • Botha Natasha

DOI
https://doi.org/10.1051/matecconf/202338804006
Journal volume & issue
Vol. 388
p. 04006

Abstract

Read online

There is a need to ensure that an Unmanned Aerial Vehicle (UAV) can fly safely in indoor environments. In this paper, the visual-inertial odometry (VIO) implemented in a commercial UAV is tested to determine its reliability to keep its position in confined indoor spaces. The experimental results indicated that the UAV was able to track the subject from an average distance of 3915.23 mm in the y-direction, with RMSEx and RMSEy of 452.47 mm and 332.66 mm, respectively. This shows that the UAV can track objects in confined environments, such as the aisles of an indoor warehouse.