How to Assess Augmented Sensor Model for Robust Data Fusion with an INS? A Case Study with a Dual Antenna GNSS

Thomas Brunner, Jean-Philippe Michel, Marie-Lise Duplaquet, Florian Jerram, Jean-Baptiste Lacambre, Pierre d’Harcourt

Abstract:	Performance and robustness of sensor data fusion with an INS depend heavily on the quality of the sensor model. This might be hard to determine in real-life situations, with possibly no ground truth available. We present a methodology to determine and assess the quality of a sensor model, that we apply to the fusion of dual antenna GNSS heading measurements with an INS. In particular, we design a statistical validation test based on the Allan variance of Kalman filter innovation. A high-end Phins Exail INS (with Fiber Optic Gyroscopes) is used as heading reference in dynamic motion. The dual antenna measurement statistical analysis is performed using cumulative distribution function and Allan variance. It leads to a sensor model including state augmentation to deal with error correlation over time. Validation is done with a real-life navigation setup, using only a MEMS INS and the dual antenna GNSS receiver. The validation method is stringent, only the full model can pass the test, while a naive approach fails.
Published in:	Proceedings of the 2024 International Technical Meeting of The Institute of Navigation January 23 - 25, 2024 Hyatt Regency Long Beach Long Beach, California
Pages:	1048 - 1063
Cite this article:	Brunner, Thomas, Michel, Jean-Philippe, Duplaquet, Marie-Lise, Jerram, Florian, Lacambre, Jean-Baptiste, d’Harcourt, Pierre, "How to Assess Augmented Sensor Model for Robust Data Fusion with an INS? A Case Study with a Dual Antenna GNSS," Proceedings of the 2024 International Technical Meeting of The Institute of Navigation, Long Beach, California, January 2024, pp. 1048-1063. https://doi.org/10.33012/2024.19532
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In