Aleatoric Uncertainty Reduction in a Multisensor Navigation System Using Gated Recurrent Unit

Tarafder Elmi Tabassum, Sorin Andrei Negru, Ivan Petrunin

Abstract:	In a degraded global navigation satellite systems (GNSS) environment such as urban canyons, integrating multi-sensor navigation systems are usually utilized in various Urban Air Mobility (UAM) applications to achieve reliable navigation. This paper addresses the major challenges of achieving resilient navigation by aiming to enhance the performance of multi-sensor navigation solution with position uncertainty correction. The novel contribution introduces GRU-based multiple sensor error compensation aided ESKF for the hybrid multi-sensor framework that utilizes GRU-based error compensators networks for sensors to enhance positioning within complex environments. Unlike state-of-the-art systems that lack evidence of providing performance enhancement while reducing uncertainty within multi-sensor architectures, the proposed framework simultaneously reduces position uncertainty and improves accuracy and robustness. Furthermore, training and testing datasets are generated using MATLAB incorporating Unreal Engine simulation environment for UAVs to replicate complex scenarios including environmental conditions, illumination variations, weather effects, drifting, sensor noise and flight dynamics where traditional systems tend to fail. The proposed hybrid architecture has been validated under combinations of complex scenarios including various sources of aleatoric uncertainty such as sensor noise, feature tracking error, environmental dynamics, weather effects and lighting conditions, and the EuRoc dataset. For instance, it captures uncertainty arising from the limited visibility of feature, reflection and improved performance by 90% compared to the state-of-the-art solution for local filter 1, 82% in local filter 2 compared to degraded GNSS, and 86% improvement in overall performance. Additionally, it demonstrates generalization ability over seen and unseen scenarios with complex fault conditions tested on EuRoc datasets for local filter 1, validating the efficiency and resiliency of the proposed approach.
Published in:	Proceedings of the 38th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2025) September 8 - 12, 2025 Hilton Baltimore Inner Harbor Baltimore, Maryland
Pages:	962 - 973
Cite this article:	Tabassum, Tarafder Elmi, Negru, Sorin Andrei, Petrunin, Ivan, "Aleatoric Uncertainty Reduction in a Multisensor Navigation System Using Gated Recurrent Unit," Proceedings of the 38th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2025), Baltimore, Maryland, September 2025, pp. 962-973. https://doi.org/10.33012/2025.20231
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In