Learning pose estimation for UAV autonomous navigation and landing using visual-inertial sensor data

Title	Learning pose estimation for UAV autonomous navigation and landing using visual-inertial sensor data
Authors	Francesca Baldini, Animashree Anandkumar and Richard M. Murray
Source	2020 American Control Conference (ACC)
Abstract	In this work, we propose a robust network-in-the-loop control system for autonomous navigation and landing of an Unmanned-Aerial-Vehicle (UAV). To estimate the UAV's absolute pose, we develop a deep neural network (DNN) architecture for visual-inertial odometry, which provides a robust alternative to traditional methods. We first evaluate the accuracy of the estimation by comparing the prediction of our model to traditional visual-inertial approaches on the publicly available EuRoC MAV dataset. The results indicate a clear improvement in the accuracy of the pose estimation up to 25% over the baseline. Finally, we integrate the data-driven estimator in the closed-loop flight control system of Airsim, a simulator available as a plugin for Unreal Engine, and we provide simulation results for autonomous navigation and landing.
Type	Conference paper
URL	https://authors.library.caltech.edu/100568/1/1912.04527.pdf
DOI
Tag	BAM20-acc
ID	2019i
Funding	NSF VeHICaL
Flags