Robust perception-action models should be learned from training data with diverse visual appearances and realistic behaviors, yet current approaches to deep visuomotor policy learning have been generally limited to in-situ models learned from a single vehicle or a simulation environment. We advocate learning a generic vehicle motion model from large scale crowd-sourced video data, and develop an end-to-end trainable architecture for learning to predict a distribution over future vehicle egomotion from instantaneous monocular camera observations and previous vehicle state. Our model incorporates a novel FCN-LSTM architecture, which can be learned from large-scale crowd-sourced vehicle action data, and leverages available scene segmentation side tasks to improve performance under a privileged learning paradigm.

@inproceedings{Xu:2017:ELO,
   author = "Huazhe Xu and Yang Gao and Fisher Yu and Trevor Darrell",
   title = "End-to-end Learning of Driving Models from Large-scale Video Datasets",
   booktitle = "IEEE Conference on Computer Vision and Pattern Recognition (CVPR) oral
      presentation",
   year = "2017",
   month = jul
}