Orbis: Overcoming Challenges of Long-Horizon Prediction in Driving World Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Mousakhan, Arian, Mittal, Sudhanshu, Galesso, Silvio, Farid, Karim, Brox, Thomas
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909955678797824
author Mousakhan, Arian
Mittal, Sudhanshu
Galesso, Silvio
Farid, Karim
Brox, Thomas
author_facet Mousakhan, Arian
Mittal, Sudhanshu
Galesso, Silvio
Farid, Karim
Brox, Thomas
contents Existing world models for autonomous driving struggle with long-horizon generation and generalization to challenging scenarios. In this work, we develop a model using simple design choices, and without additional supervision or sensors, such as maps, depth, or multiple cameras. We show that our model yields state-of-the-art performance, despite having only 469M parameters and being trained on 280h of video data. It particularly stands out in difficult scenarios like turning maneuvers and urban traffic. We test whether discrete token models possibly have advantages over continuous models based on flow matching. To this end, we set up a hybrid tokenizer that is compatible with both approaches and allows for a side-by-side comparison. Our study concludes in favor of the continuous autoregressive model, which is less brittle on individual design choices and more powerful than the model built on discrete tokens. Code, models and qualitative results are publicly available at https://lmb-freiburg.github.io/orbis.github.io/.
format Preprint
id arxiv_https___arxiv_org_abs_2507_13162
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Orbis: Overcoming Challenges of Long-Horizon Prediction in Driving World Models
Mousakhan, Arian
Mittal, Sudhanshu
Galesso, Silvio
Farid, Karim
Brox, Thomas
Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
Existing world models for autonomous driving struggle with long-horizon generation and generalization to challenging scenarios. In this work, we develop a model using simple design choices, and without additional supervision or sensors, such as maps, depth, or multiple cameras. We show that our model yields state-of-the-art performance, despite having only 469M parameters and being trained on 280h of video data. It particularly stands out in difficult scenarios like turning maneuvers and urban traffic. We test whether discrete token models possibly have advantages over continuous models based on flow matching. To this end, we set up a hybrid tokenizer that is compatible with both approaches and allows for a side-by-side comparison. Our study concludes in favor of the continuous autoregressive model, which is less brittle on individual design choices and more powerful than the model built on discrete tokens. Code, models and qualitative results are publicly available at https://lmb-freiburg.github.io/orbis.github.io/.
title Orbis: Overcoming Challenges of Long-Horizon Prediction in Driving World Models
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2507.13162