End-to-End Driving Via Conditional Imitation Learning

  title={End-to-End Driving Via Conditional Imitation Learning},
  author={Felipe Codevilla and Matthias Miiller and Antonio L{\'o}pez and Vladlen Koltun and Alexey Dosovitskiy},
  journal={2018 IEEE International Conference on Robotics and Automation (ICRA)},
Deep networks trained on demonstrations of human driving have learned to follow roads and avoid obstacles. However, driving policies trained via imitation learning cannot be controlled at test time. A vehicle trained end-to-end to imitate an expert cannot be guided to take a specific turn at an upcoming intersection. This limits the utility of such systems. We propose to condition imitation learning on high-level command input. At test time, the learned driving policy functions as a chauffeur… CONTINUE READING

From This Paper

Figures, tables, and topics from this paper.


Publications citing this paper.
Showing 1-3 of 3 extracted citations


Publications referenced by this paper.
Showing 1-10 of 41 references

End to End Learning for Self-Driving Cars

ArXiv • 2016
View 8 Excerpts
Highly Influenced

Autonomous driving

U. Franke
In Computer Vision in Vehicle Technology • 2017
View 1 Excerpt

Dart: Optimizing noise injection in imitation learning

M. Laskey, A. Dragan, J. Lee, K. Goldberg, R. Fox
In Conference on Robot Learning (CoRL), • 2017

Similar Papers

Loading similar papers…