Alpamayo-R1: Bridging Reasoning and Action Prediction for Generalizable Autonomous Driving in the Long Tail

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: NVIDIA, :, Wang, Yan, Luo, Wenjie, Bai, Junjie, Cao, Yulong, Che, Tong, Chen, Ke, Chen, Yuxiao, Diamond, Jenna, Ding, Yifan, Ding, Wenhao, Feng, Liang, Heinrich, Greg, Huang, Jack, Karkus, Peter, Li, Boyi, Li, Pinyi, Lin, Tsung-Yi, Liu, Dongran, Liu, Ming-Yu, Liu, Langechuan, Liu, Zhijian, Lu, Jason, Mao, Yunxiang, Molchanov, Pavlo, Pavao, Lindsey, Peng, Zhenghao, Ranzinger, Mike, Schmerling, Ed, Shen, Shida, Shi, Yunfei, Tariq, Sarah, Tian, Ran, Wekel, Tilman, Weng, Xinshuo, Xiao, Tianjun, Yang, Eric, Yang, Xiaodong, You, Yurong, Zeng, Xiaohui, Zhang, Wenyuan, Ivanovic, Boris, Pavone, Marco
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918274749431808
author NVIDIA
:
Wang, Yan
Luo, Wenjie
Bai, Junjie
Cao, Yulong
Che, Tong
Chen, Ke
Chen, Yuxiao
Diamond, Jenna
Ding, Yifan
Ding, Wenhao
Feng, Liang
Heinrich, Greg
Huang, Jack
Karkus, Peter
Li, Boyi
Li, Pinyi
Lin, Tsung-Yi
Liu, Dongran
Liu, Ming-Yu
Liu, Langechuan
Liu, Zhijian
Lu, Jason
Mao, Yunxiang
Molchanov, Pavlo
Pavao, Lindsey
Peng, Zhenghao
Ranzinger, Mike
Schmerling, Ed
Shen, Shida
Shi, Yunfei
Tariq, Sarah
Tian, Ran
Wekel, Tilman
Weng, Xinshuo
Xiao, Tianjun
Yang, Eric
Yang, Xiaodong
You, Yurong
Zeng, Xiaohui
Zhang, Wenyuan
Ivanovic, Boris
Pavone, Marco
author_facet NVIDIA
:
Wang, Yan
Luo, Wenjie
Bai, Junjie
Cao, Yulong
Che, Tong
Chen, Ke
Chen, Yuxiao
Diamond, Jenna
Ding, Yifan
Ding, Wenhao
Feng, Liang
Heinrich, Greg
Huang, Jack
Karkus, Peter
Li, Boyi
Li, Pinyi
Lin, Tsung-Yi
Liu, Dongran
Liu, Ming-Yu
Liu, Langechuan
Liu, Zhijian
Lu, Jason
Mao, Yunxiang
Molchanov, Pavlo
Pavao, Lindsey
Peng, Zhenghao
Ranzinger, Mike
Schmerling, Ed
Shen, Shida
Shi, Yunfei
Tariq, Sarah
Tian, Ran
Wekel, Tilman
Weng, Xinshuo
Xiao, Tianjun
Yang, Eric
Yang, Xiaodong
You, Yurong
Zeng, Xiaohui
Zhang, Wenyuan
Ivanovic, Boris
Pavone, Marco
contents End-to-end architectures trained via imitation learning have advanced autonomous driving by scaling model size and data, yet performance remains brittle in safety-critical long-tail scenarios where supervision is sparse and causal understanding is limited. We introduce Alpamayo-R1 (AR1), a vision-language-action model (VLA) that integrates Chain of Causation reasoning with trajectory planning for complex driving scenarios. Our approach features three key innovations: (1) the Chain of Causation (CoC) dataset, built through a hybrid auto-labeling and human-in-the-loop pipeline producing decision-grounded, causally linked reasoning traces aligned with driving behaviors; (2) a modular VLA architecture combining Cosmos-Reason, a vision-language model pre-trained for Physical AI, with a diffusion-based trajectory decoder that generates dynamically feasible trajectories in real time; (3) a multi-stage training strategy using supervised fine-tuning to elicit reasoning and reinforcement learning (RL) to enforce reasoning-action consistency and optimize reasoning quality. AR1 achieves up to a 12% improvement in planning accuracy on challenging cases compared to a trajectory-only baseline, with a 35% reduction in close encounter rate in closed-loop simulation. RL post-training improves reasoning quality by 45% and reasoning-action consistency by 37%. Model scaling from 0.5B to 7B parameters shows consistent improvements. On-vehicle road tests confirm real-time performance (99 ms latency) and successful urban deployment. By bridging interpretable reasoning with precise control, AR1 demonstrates a practical path towards Level 4 autonomous driving. Model weights are available at https://huggingface.co/nvidia/Alpamayo-R1-10B with inference code at https://github.com/NVlabs/alpamayo.
format Preprint
id arxiv_https___arxiv_org_abs_2511_00088
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Alpamayo-R1: Bridging Reasoning and Action Prediction for Generalizable Autonomous Driving in the Long Tail
NVIDIA
:
Wang, Yan
Luo, Wenjie
Bai, Junjie
Cao, Yulong
Che, Tong
Chen, Ke
Chen, Yuxiao
Diamond, Jenna
Ding, Yifan
Ding, Wenhao
Feng, Liang
Heinrich, Greg
Huang, Jack
Karkus, Peter
Li, Boyi
Li, Pinyi
Lin, Tsung-Yi
Liu, Dongran
Liu, Ming-Yu
Liu, Langechuan
Liu, Zhijian
Lu, Jason
Mao, Yunxiang
Molchanov, Pavlo
Pavao, Lindsey
Peng, Zhenghao
Ranzinger, Mike
Schmerling, Ed
Shen, Shida
Shi, Yunfei
Tariq, Sarah
Tian, Ran
Wekel, Tilman
Weng, Xinshuo
Xiao, Tianjun
Yang, Eric
Yang, Xiaodong
You, Yurong
Zeng, Xiaohui
Zhang, Wenyuan
Ivanovic, Boris
Pavone, Marco
Robotics
Artificial Intelligence
Machine Learning
End-to-end architectures trained via imitation learning have advanced autonomous driving by scaling model size and data, yet performance remains brittle in safety-critical long-tail scenarios where supervision is sparse and causal understanding is limited. We introduce Alpamayo-R1 (AR1), a vision-language-action model (VLA) that integrates Chain of Causation reasoning with trajectory planning for complex driving scenarios. Our approach features three key innovations: (1) the Chain of Causation (CoC) dataset, built through a hybrid auto-labeling and human-in-the-loop pipeline producing decision-grounded, causally linked reasoning traces aligned with driving behaviors; (2) a modular VLA architecture combining Cosmos-Reason, a vision-language model pre-trained for Physical AI, with a diffusion-based trajectory decoder that generates dynamically feasible trajectories in real time; (3) a multi-stage training strategy using supervised fine-tuning to elicit reasoning and reinforcement learning (RL) to enforce reasoning-action consistency and optimize reasoning quality. AR1 achieves up to a 12% improvement in planning accuracy on challenging cases compared to a trajectory-only baseline, with a 35% reduction in close encounter rate in closed-loop simulation. RL post-training improves reasoning quality by 45% and reasoning-action consistency by 37%. Model scaling from 0.5B to 7B parameters shows consistent improvements. On-vehicle road tests confirm real-time performance (99 ms latency) and successful urban deployment. By bridging interpretable reasoning with precise control, AR1 demonstrates a practical path towards Level 4 autonomous driving. Model weights are available at https://huggingface.co/nvidia/Alpamayo-R1-10B with inference code at https://github.com/NVlabs/alpamayo.
title Alpamayo-R1: Bridging Reasoning and Action Prediction for Generalizable Autonomous Driving in the Long Tail
topic Robotics
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2511.00088