Driving Through Ghosts: Behavioral Cloning with False Positives

@article{Buhler2020DrivingTG,
  title={Driving Through Ghosts: Behavioral Cloning with False Positives},
  author={Andreas Buhler and Adrien Gaidon and Andrei Cramariuc and Rares Ambrus and G. Rosman and W. Burgard},
  journal={2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  year={2020},
  pages={5431-5437}
}
Safe autonomous driving requires robust detection of other traffic participants. However, robust does not mean perfect, and safe systems typically minimize missed detections at the expense of a higher false positive rate. This results in conservative and yet potentially dangerous behavior such as avoiding imaginary obstacles. In the context of behavioral cloning, perceptual errors at training time can lead to learning difficulties or wrong policies, as expert demonstrations might be… Expand
1 Citations

Figures and Tables from this paper

PillarFlow: End-to-end Birds-eye-view Flow Estimation for Autonomous Driving
  • 2
  • PDF

References

SHOWING 1-10 OF 58 REFERENCES
Exploring the Limitations of Behavior Cloning for Autonomous Driving
  • 74
  • PDF
ChauffeurNet: Learning to Drive by Imitating the Best and Synthesizing the Worst
  • 252
  • PDF
Learning by Cheating
  • 58
  • PDF
Conditional Affordance Learning for Driving in Urban Environments
  • 74
  • PDF
Uncertainty-Aware Data Aggregation for Deep Imitation Learning
  • 6
  • PDF
PRECOG: PREdiction Conditioned on Goals in Visual Multi-Agent Settings
  • 113
  • PDF
End-to-End Driving Via Conditional Imitation Learning
  • 369
  • PDF
Uncertainty-Aware Driver Trajectory Prediction at Urban Intersections
  • 30
  • PDF
...
1
2
3
4
5
...