Hydra-MDP++: Advancing End-to-End Driving via Expert-Guided Hydra-Distillation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Kailin, Li, Zhenxin, Lan, Shiyi, Xie, Yuan, Zhang, Zhizhong, Liu, Jiayi, Wu, Zuxuan, Yu, Zhiding, Alvarez, Jose M.
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916653952925696
author Li, Kailin
Li, Zhenxin
Lan, Shiyi
Xie, Yuan
Zhang, Zhizhong
Liu, Jiayi
Wu, Zuxuan
Yu, Zhiding
Alvarez, Jose M.
author_facet Li, Kailin
Li, Zhenxin
Lan, Shiyi
Xie, Yuan
Zhang, Zhizhong
Liu, Jiayi
Wu, Zuxuan
Yu, Zhiding
Alvarez, Jose M.
contents Hydra-MDP++ introduces a novel teacher-student knowledge distillation framework with a multi-head decoder that learns from human demonstrations and rule-based experts. Using a lightweight ResNet-34 network without complex components, the framework incorporates expanded evaluation metrics, including traffic light compliance (TL), lane-keeping ability (LK), and extended comfort (EC) to address unsafe behaviors not captured by traditional NAVSIM-derived teachers. Like other end-to-end autonomous driving approaches, \hydra processes raw images directly without relying on privileged perception signals. Hydra-MDP++ achieves state-of-the-art performance by integrating these components with a 91.0% drive score on NAVSIM through scaling to a V2-99 image encoder, demonstrating its effectiveness in handling diverse driving scenarios while maintaining computational efficiency.
format Preprint
id arxiv_https___arxiv_org_abs_2503_12820
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Hydra-MDP++: Advancing End-to-End Driving via Expert-Guided Hydra-Distillation
Li, Kailin
Li, Zhenxin
Lan, Shiyi
Xie, Yuan
Zhang, Zhizhong
Liu, Jiayi
Wu, Zuxuan
Yu, Zhiding
Alvarez, Jose M.
Computer Vision and Pattern Recognition
Hydra-MDP++ introduces a novel teacher-student knowledge distillation framework with a multi-head decoder that learns from human demonstrations and rule-based experts. Using a lightweight ResNet-34 network without complex components, the framework incorporates expanded evaluation metrics, including traffic light compliance (TL), lane-keeping ability (LK), and extended comfort (EC) to address unsafe behaviors not captured by traditional NAVSIM-derived teachers. Like other end-to-end autonomous driving approaches, \hydra processes raw images directly without relying on privileged perception signals. Hydra-MDP++ achieves state-of-the-art performance by integrating these components with a 91.0% drive score on NAVSIM through scaling to a V2-99 image encoder, demonstrating its effectiveness in handling diverse driving scenarios while maintaining computational efficiency.
title Hydra-MDP++: Advancing End-to-End Driving via Expert-Guided Hydra-Distillation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2503.12820