Corpus ID: 225039984

Trajectory Prediction using Equivariant Continuous Convolution

@article{Walters2020TrajectoryPU,
  title={Trajectory Prediction using Equivariant Continuous Convolution},
  author={R. Walters and Jinxi Li and R. Yu},
  journal={ArXiv},
  year={2020},
  volume={abs/2010.11344}
}
  • R. Walters, Jinxi Li, R. Yu
  • Published 2020
  • Computer Science
  • ArXiv
  • Trajectory prediction is a critical part of many AI applications, for example, the safe operation of autonomous vehicles. However, current methods are prone to making inconsistent and physically unrealistic predictions. We leverage insights from fluid dynamics to overcome this limitation by considering internal symmetry in trajectories. We propose a novel model, Equivariant Continous COnvolution (ECCO) for improved trajectory prediction. ECCO uses rotationally-equivariant continuous… CONTINUE READING

    Figures and Tables from this paper

    References

    SHOWING 1-10 OF 47 REFERENCES
    Incorporating Symmetry into Deep Dynamics Models for Improved Generalization
    • 9
    • PDF
    Uncertainty-aware Short-term Motion Prediction of Traffic Actors for Autonomous Driving
    • 54
    • PDF
    Social LSTM: Human Trajectory Prediction in Crowded Spaces
    • 1,021
    • Highly Influential
    • PDF
    VectorNet: Encoding HD Maps and Agent Dynamics From Vectorized Representation
    • Jiyang Gao, Chen Sun, +4 authors C. Schmid
    • Computer Science, Mathematics
    • 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
    • 2020
    • 24
    • Highly Influential
    • PDF
    Human Trajectory Forecasting in Crowds: A Deep Learning Perspective
    • 6
    • Highly Influential
    • PDF
    Generalizing Convolutional Neural Networks for Equivariance to Lie Groups on Arbitrary Continuous Data
    • 25
    • PDF
    Human motion trajectory prediction: a survey
    • 107
    • PDF
    Argoverse: 3D Tracking and Forecasting With Rich Maps
    • 173
    • Highly Influential
    • PDF
    Lagrangian Fluid Simulation with Continuous Convolutions
    • 22
    • Highly Influential
    • PDF