GA-SVR and pseudo-position-aided GPS/INS integration during GPS outage

Tan, Xinglong; Wang, Jian; Jin, Shuanggen; Meng, Xiaolin

Abstract

The performance of Global Positioning System and Inertial Navigation System (GPS/INS) integrated navigation is reduced when GPS is blocked. This paper proposes an algorithm to overcome the condition where GPS is unavailable. Together with a parameter-optimised Genetic Algorithm (GA), a Support Vector Regression (SVR) algorithm is used to construct the mapping function between the specific force, angular rate increments of INS measurements and the increments of the GPS position. During GPS outages, the real-time pseudo-GPS position is predicted with the mapping function, and the corresponding covariance matrix is estimated by an improved adaptive filtering algorithm. A GPS/INS integration scheme is demonstrated where the vehicle travels along a straight line and around a curve, with respect to both low-speed-stable and high-speed-unstable navigation platforms. The results show that the proposed algorithm provides a better performance when GPS is unavailable.

Citation

Tan, X., Wang, J., Jin, S., & Meng, X. (2015). GA-SVR and pseudo-position-aided GPS/INS integration during GPS outage. Journal of Navigation, 68(4), https://doi.org/10.1017/S037346331500003X

• Journal Article Type: Article
• Acceptance Date: Jan 10, 2015
• Online Publication Date: Feb 13, 2015
• Publication Date: Jul 1, 2015
• Deposit Date: Jul 25, 2016
• Publicly Available Date: Jul 25, 2016
• Journal: Journal of Navigation
• Print ISSN: 0373-4633
• Electronic ISSN: 1469-7785
• Publisher: Cambridge University Press
• Peer Reviewed: Peer Reviewed
• Volume: 68
• Issue: 4
• DOI: https://doi.org/10.1017/S037346331500003X
• Keywords: GPS/INS Integrated Navigation; Pseudo-GPS Position; Support Vector Regression; Improved Adaptive Filtering
• Public URL: https://nottingham-repository.worktribe.com/output/983170
• Publisher URL: http://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=9719975&fileId=S037346331500003X