Presented at IROS 2022

A study on “Localization for Robot Navigation Using Camera Images” by Takahiro Niwa et al. was accepted for presentation at The 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

With the developments in deep learning technology, robot navigation using only an onboard camera has become possible. However, navigation can be challenging in an environment with numerous similar landscapes because localization methods may select the wrong node. To overcome this issue, we proposed a novel learning-based localization method utilizing spatial and temporal information obtained from maps and camera images, respectively. We implemented this method in a mobile robot with an omnidirectional camera and demonstrated navigation in indoor environments with numerous similar landscapes. This result is expected to contribute to the realization of various services based on mobile robot navigation.

Title: Spatio-temporal Graph Localization Networks for Image-based Navigation
Authors: Niwa, T., Taguchi, S., Hirose, N.
Appears in: IEEE/RSJ International Conference on Intelligent Robots and Systems
Presented: October 24, 2022
https://doi.org/10.1109/IROS47612.2022.9981958