The manuscript “Driver Modelling for Automated Vehicles: The Importance of Multimodal Fusion” has been accepted for publication at the Revista Iberoamericana de Automática e Informática Industrial (RIAI). Congrats to the authors: Raul Fernandez-Matellan, David Martin Gomez, Arturo de la Escalera Hueso.
In this work, different multimodal fusion strategies based on representations obtained by autoencoders that project each modality to a common space have been evaluated in the TD2D dataset, and latent fusion and late fusion have been compared with unimodal baselines.
