Data Fusion for Deep Learning on Transport Mode Detection: A Case Study

Hugues Moreau (1, 2) , Andrea Vassilev (1) , Liming Chen (3, 2)

1 CEA-LETI - Commissariat à l'énergie atomique et aux énergies alternatives - Laboratoire d'Electronique et de Technologie de l'Information
2 LIRIS - Laboratoire d'InfoRmatique en Image et Systèmes d'information
3 ECL - École Centrale de Lyon

In Transport Mode Detection, a great diversity of methodologies exist according to the choice made on sensors, preprocessing, model used, etc. In this domain, the comparisons between each option are not always complete. Experiments on a public, real-life dataset are led here to evaluate carefully each of the choices that were made, with a specific emphasis on data fusion methods. Our most surprising finding is that none of the methods we implemented from the literature is better than a simple late fusion. Two important decisions are the choice of a sensor and the choice of a representation for the data: we found that using 2D convolutions on spectrograms with a logarithmic axis for the frequencies was better than 1-dimensional temporal representations. To foster the research on deep learning with embedded inertial sensors, we release our code along with our publication.

Mots clés: Transport Mode Detection; Data Fusion for Deep Learning on Transport Mode Detection: A Case Study; spectrogram; inertial sensors; accelerometer; Deep learning; CNN

Domaines: Intelligence artificielle [cs.AI]; Systèmes embarqués

Hugues Moreau, Andrea Vassilev, Liming Chen. Data Fusion for Deep Learning on Transport Mode Detection: A Case Study. 2021. ⟨hal-03242601v2⟩