Automatic Detection of Magnetic Disturbances in Magnetic Inertial Measurement Unit Sensors Based on Recurrent Neural Networks

Elkyn Alexander Belalcazar-Bolaños; Diego Torricelli; José L Pons

doi:10.3390/s23249683

Automatic Detection of Magnetic Disturbances in Magnetic Inertial Measurement Unit Sensors Based on Recurrent Neural Networks

Sensors (Basel). 2023 Dec 7;23(24):9683. doi: 10.3390/s23249683.

Authors

Elkyn Alexander Belalcazar-Bolaños^{1

2}, Diego Torricelli¹, José L Pons^{3

4

5

6}

Affiliations

¹ Neural Rehabilitation Group, Cajal Institute, Spanish National Research Council (CSIC), 28002 Madrid, Spain.
² Department of Automation and Systems Engineering, Carlos III University, 28911 Madrid, Spain.
³ Legs and Walking AbilityLab, Shirley Ryan AbilityLab, Chicago, IL 60611, USA.
⁴ Department of Physical Medicine and Rehabilitation, Feinberg School of Medicine, Northwestern University, Evanston, IL 60208, USA.
⁵ Department of Biomedical Engineering, McCormick School of Engineering and Applied Science, Northwestern University, Evanston, IL 60208, USA.
⁶ Department of Mechanical Engineering, McCormick School of Engineering and Applied Science, Northwestern University, Evanston, IL 60208, USA.

Abstract

This paper proposes a new methodology for the automatic detection of magnetic disturbances from magnetic inertial measurement unit (MIMU) sensors based on deep learning. The proposed approach considers magnetometer data as input to a long short-term memory (LSTM) neural network and obtains a labeled time series output with the posterior probabilities of magnetic disturbance. We trained our algorithm on a data set that reproduces a wide range of magnetic perturbations and MIMU motions in a repeatable and reproducible way. The model was trained and tested using 15 folds, which considered independence in sensor, disturbance direction, and signal type. On average, the network can adequately detect the disturbances in 98% of the cases, which represents a significant improvement over current threshold-based detection algorithms.

Keywords: deep neural; inertial measurement; magnetic disturbance; magnetometer.

Grants and funding

This work has been partially funded by the Grant MIA.2021.M02.0007, which has been funded by the Ministry of Economic Affairs and Digital Transformation and by the “European Union NextGenerationEU/PRTR”. E.A.B.B acknowledges support from the Community of Madrid through the Grant DDII2017-IND2017/TIC-7705, and the Colombian Ministry Minciencias.