AS-EKF : a delay aware state estimation technique for telepresence robot navigation

Das, Barnali and Dobie, Gordon and Pierce, Stephen Gareth; (2019) AS-EKF : a delay aware state estimation technique for telepresence robot navigation. In: Proceedings - 3rd IEEE International Conference on Robotic Computing, IRC 2019. IEEE, ITA, pp. 624-629. ISBN 9781538692455 (https://doi.org/10.1109/IRC.2019.00126)

[thumbnail of Das-etal-IRC2019-AS-EKF-a-delay-aware-state-estimation-technique-for-telepresence]
Preview
Text. Filename: Das_etal_IRC2019_AS_EKF_a_delay_aware_state_estimation_technique_for_telepresence.pdf
Accepted Author Manuscript

Download (1MB)| Preview

Abstract

State estimation in dynamical telepresence systemsis very important in real-world applications as the true state of the robot is unspecified and sensors provide only a sequence of noisy measurements. In this research, we proposed a new technique of state estimation using delayed sensor measurements of a Telepresence robot for real-time navigation. An Augmented State Extended Kalman Filter (AS-EKF) is introduced to estimate the true position of the robot. The proposed algorithm was successfully tested in a real-environment experimental frameworkusing a state-of-the-art differential-drive telepresence robot. Our results show improvements of an average of more than 34% when compared to traditional EKF.