notesum.ai

Published at November 27

FASIONAD : FAst and Slow FusION Thinking Systems for Human-Like Autonomous Driving with Adaptive Feedback

cs.RO

cs.CV

Released Date: November 27, 2024

Authors: Kangan Qian¹, Zhikun Ma², Yangfan He³, Ziang Luo, Tianyu Shi⁴, Tianze Zhu¹, Jiayin Li⁵, Jianhui Wang⁶, Ziyu Chen⁷, Xiao He⁷, Yining Shi¹, Zheng Fu¹, Xinyu Jiao¹, Kun Jiang¹, Diange Yang¹, Takafumi Matsumaru²

Aff.: ¹Tsinghua University; ²Waseda University; ³University of Minnesota - Twin Cities; ⁴University of Toronto; ⁵Xiamen University Malaysia; ⁶University of Electronic Science and Technology of China; ⁷AI2Robotics

Arxiv: http://arxiv.org/abs/2411.18013v1

Methods	Modality	DS (%) $\uparrow$	RC (%) $\uparrow$
CILRS [13]	C	7.47	13.40
LBC [14]	C	30.97	55.01
Transfuser [10]	C+L	54.52	78.41
ST-P3 [22]	C	55.14	86.74
VAD [24]	C	64.29	87.26
Agent-Driver [38]	C	64.31	87.31
FASIONAD (Ours)	C	64.83	89.04