TY - GEN
T1 - Perceive, Attend, and Drive
T2 - 2021 IEEE International Conference on Robotics and Automation, ICRA 2021
AU - Wei, Bob
AU - Ren, Mengye
AU - Zeng, Wenyuan
AU - Liang, Ming
AU - Yang, Bin
AU - Urtasun, Raquel
N1 - Publisher Copyright:
© 2021 IEEE
PY - 2021
Y1 - 2021
N2 - In this paper, we propose an end-to-end self-driving network featuring a sparse attention module that learns to automatically attend to important regions of the input. The attention module specifically targets motion planning, whereas prior literature only applied attention in perception tasks. Learning an attention mask directly targeted for motion planning significantly improves the planner safety by performing more focused computation. Furthermore, visualizing the attention improves interpretability of end-to-end self-driving.
AB - In this paper, we propose an end-to-end self-driving network featuring a sparse attention module that learns to automatically attend to important regions of the input. The attention module specifically targets motion planning, whereas prior literature only applied attention in perception tasks. Learning an attention mask directly targeted for motion planning significantly improves the planner safety by performing more focused computation. Furthermore, visualizing the attention improves interpretability of end-to-end self-driving.
UR - http://www.scopus.com/inward/record.url?scp=85120520414&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85120520414&partnerID=8YFLogxK
U2 - 10.1109/ICRA48506.2021.9561904
DO - 10.1109/ICRA48506.2021.9561904
M3 - Conference contribution
AN - SCOPUS:85120520414
T3 - Proceedings - IEEE International Conference on Robotics and Automation
SP - 4875
EP - 4881
BT - 2021 IEEE International Conference on Robotics and Automation, ICRA 2021
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 30 May 2021 through 5 June 2021
ER -