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

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Qian, Kangan, Ma, Zhikun, He, Yangfan, Luo, Ziang, Shi, Tianyu, Zhu, Tianze, Li, Jiayin, Wang, Jianhui, Chen, Ziyu, He, Xiao, Shi, Yining, Fu, Zheng, Jiao, Xinyu, Jiang, Kun, Yang, Diange, Matsumaru, Takafumi
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866915036519202816
author Qian, Kangan
Ma, Zhikun
He, Yangfan
Luo, Ziang
Shi, Tianyu
Zhu, Tianze
Li, Jiayin
Wang, Jianhui
Chen, Ziyu
He, Xiao
Shi, Yining
Fu, Zheng
Jiao, Xinyu
Jiang, Kun
Yang, Diange
Matsumaru, Takafumi
author_facet Qian, Kangan
Ma, Zhikun
He, Yangfan
Luo, Ziang
Shi, Tianyu
Zhu, Tianze
Li, Jiayin
Wang, Jianhui
Chen, Ziyu
He, Xiao
Shi, Yining
Fu, Zheng
Jiao, Xinyu
Jiang, Kun
Yang, Diange
Matsumaru, Takafumi
contents Ensuring safe, comfortable, and efficient navigation is a critical goal for autonomous driving systems. While end-to-end models trained on large-scale datasets excel in common driving scenarios, they often struggle with rare, long-tail events. Recent progress in large language models (LLMs) has introduced enhanced reasoning capabilities, but their computational demands pose challenges for real-time decision-making and precise planning. This paper presents FASIONAD, a novel dual-system framework inspired by the cognitive model "Thinking, Fast and Slow." The fast system handles routine navigation tasks using rapid, data-driven path planning, while the slow system focuses on complex reasoning and decision-making in challenging or unfamiliar situations. A dynamic switching mechanism based on score distribution and feedback allows seamless transitions between the two systems. Visual prompts generated by the fast system enable human-like reasoning in the slow system, which provides high-quality feedback to enhance the fast system's decision-making. To evaluate FASIONAD, we introduce a new benchmark derived from the nuScenes dataset, specifically designed to differentiate fast and slow scenarios. FASIONAD achieves state-of-the-art performance on this benchmark, establishing a new standard for frameworks integrating fast and slow cognitive processes in autonomous driving. This approach paves the way for more adaptive, human-like autonomous driving systems.
format Preprint
id arxiv_https___arxiv_org_abs_2411_18013
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle FASIONAD : FAst and Slow FusION Thinking Systems for Human-Like Autonomous Driving with Adaptive Feedback
Qian, Kangan
Ma, Zhikun
He, Yangfan
Luo, Ziang
Shi, Tianyu
Zhu, Tianze
Li, Jiayin
Wang, Jianhui
Chen, Ziyu
He, Xiao
Shi, Yining
Fu, Zheng
Jiao, Xinyu
Jiang, Kun
Yang, Diange
Matsumaru, Takafumi
Robotics
Computer Vision and Pattern Recognition
Ensuring safe, comfortable, and efficient navigation is a critical goal for autonomous driving systems. While end-to-end models trained on large-scale datasets excel in common driving scenarios, they often struggle with rare, long-tail events. Recent progress in large language models (LLMs) has introduced enhanced reasoning capabilities, but their computational demands pose challenges for real-time decision-making and precise planning. This paper presents FASIONAD, a novel dual-system framework inspired by the cognitive model "Thinking, Fast and Slow." The fast system handles routine navigation tasks using rapid, data-driven path planning, while the slow system focuses on complex reasoning and decision-making in challenging or unfamiliar situations. A dynamic switching mechanism based on score distribution and feedback allows seamless transitions between the two systems. Visual prompts generated by the fast system enable human-like reasoning in the slow system, which provides high-quality feedback to enhance the fast system's decision-making. To evaluate FASIONAD, we introduce a new benchmark derived from the nuScenes dataset, specifically designed to differentiate fast and slow scenarios. FASIONAD achieves state-of-the-art performance on this benchmark, establishing a new standard for frameworks integrating fast and slow cognitive processes in autonomous driving. This approach paves the way for more adaptive, human-like autonomous driving systems.
title FASIONAD : FAst and Slow FusION Thinking Systems for Human-Like Autonomous Driving with Adaptive Feedback
topic Robotics
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2411.18013