|Abstract:||In this paper, we propose a cooperative localization method based on invariant extended Kalman filter (EKF) robust to initial heading error. The technology for estimating the position of agents (for example, pedestrians, mobile robots) without pre-installed infrastructure can be used for various tasks, such as position mobile robots in factories and position firefighters at disaster sites. The method of estimating the position based on the inertial-measurement units (IMU) mounted on the agent can estimate the position without additional sensors, but the accuracy is reduced due to the accumulation of errors for a long time. Cooperative localization method is a method of correcting the position of the agent estimated based on IMU using the relative distance information between agents. Cooperative localization methods are mainly based on EKF. There is a term related to the estimated attitude in the transition matrix of the EKF. If the estimated attitude error is large, the filter state is propagated incorrectly and the estimation performance is degraded. The proposed method solves this problem by using the invariant EKF so that the attitude error does not affect the transition matrix. The state of the invariant EKF is modeled as a Lie group, and the transition matrix derived from the Lie group becomes independent of the estimated state. That is, even if there is an attitude error, the state can be exactly propagated. Experimental results show that the proposed method is robust to attitude error.|
Proceedings of the 35th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2022)
September 19 - 23, 2022
Hyatt Regency Denver
|Pages:||1624 - 1630|
|Cite this article:||
Lee, Jae Hong, Cho, Seoung Yun, Park, Chan Gook, "Invariant EKF-based Cooperative Localization System Robust to Initial Heading Error," Proceedings of the 35th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2022), Denver, Colorado, September 2022, pp. 1624-1630.
ION Members/Non-Members: 1 Download Credit