Multi-Model GPS Vector Tracking Architecture for Ground-Based Navigation

Oren D. Miller and Scott M. Martin

Peer Reviewed

Abstract:	Vector tracking receivers address several limitations of conventional scalar tracking architectures by jointly processing all satellite channels through a centralized navigation filter, allowing strong signals to aid in the tracking of weaker ones. While this tight coupling improves performance, conventional vector tracking architectures typically rely on a single motion model within the navigation filter. Mismatch between the assumed and true receiver dynamics can cause navigation errors to propagate into signal replica generation, resulting in tracking instability and potential loss of lock. This paper investigates a vector tracking architecture that incorporates a multiple-model navigation solution to address the limitations of traditional single-model assumptions. Specifically, an Interacting Multiple Model (IMM) filter is implemented within the centralized navigator. The IMM framework operates several Kalman filters in parallel, each representing a distinct receiver dynamic model, and adaptively fuses their outputs using innovation-based likelihoods. This enables real-time adaptation to changes in receiver motion while maintaining tight coupling between navigation state estimation and signal tracking. The performance of the proposed IMM-based vector tracking architecture is evaluated using a correlator-level GPS L1 C/A simulation under varying dynamic and signal-degraded conditions. Results demonstrate that the IMM-enabled receiver improves tracking robustness and increases the ability to estimate receiver replicas through line-of-sight blockages compared to conventional single-model vector tracking.
Published in:	Proceedings of the ION 2026 Pacific PNT Meeting April 13 - 16, 2026 Hilton Waikiki Beach Honolulu, Hawaii
Pages:	350 - 364
Cite this article:	Miller, Oren D., Martin, Scott M., "Multi-Model GPS Vector Tracking Architecture for Ground-Based Navigation," Proceedings of the ION 2026 Pacific PNT Meeting, Honolulu, Hawaii, April 2026, pp. 350-364. https://doi.org/10.33012/2026.20580
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In