Assured Vision Aided Inertial Localization

Andrey Soloviev, Chun Yang, Michael Veth, Clark Taylor

Abstract:	Imaging systems represent a popular approach for opportunistic navigation. While the literature tends to focus on developing new navigation observation techniques, or on improving accuracy, meeting the user's full navigation requirements usually depends on a combination of both accuracy and integrity. Compounding the issue is that these signals of opportunity typically have error statistics that are highly non-stationary, non-white, and non-Gaussian in nature. This can result in estimates that are overly optimistic and possibly divergent. This paper develops a methodology of assured vision-aided estimation and validates its performance using simulation and experimental results. As a specific case study, the feasibility is demonstrated for Assured Vision-Aided Inertial Localization (AVAIL). AVAIL utilizes a batch-processing routine that adopts a multi-pose constrained estimation (MPCE) approach. The most difficult case of monocular vision (unknown depth) is considered, but the approaches presented herein can be readily extended to cases of stereo-vision.
Published in:	Proceedings of the 27th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2014) September 8 - 12, 2014 Tampa Convention Center Tampa, Florida
Pages:	2160 - 2173
Cite this article:	Soloviev, Andrey, Yang, Chun, Veth, Michael, Taylor, Clark, "Assured Vision Aided Inertial Localization," Proceedings of the 27th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2014), Tampa, Florida, September 2014, pp. 2160-2173.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In