GASP : Unifying Geometric and Semantic Self-Supervised Pre-training for Autonomous Driving
Ljungbergh, William ; Lilja, Adam ; Tonderski, Adam LU
; Ling, Arvid Laveno
; Lindstrom, Carl
; Verbeke, Willem
; Fu, Junsheng
; Petersson, Christoffer
; Hammarstrand, Lars
and Felsberg, Michael
(2026)
2026 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2026
In Proceedings - 2026 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2026
p.3077-3087
- Abstract
Self-supervised pre-training based on next-token prediction has enabled large language models to capture the underlying structure of text, and has led to unprecedented performance on a large array of tasks when applied at scale. Similarly, autonomous driving generates vast amounts of spatiotemporal data, alluding to the possibility of harnessing scale to learn the underlying geometric and semantic structure of the environment and its evolution over time. In this direction, we propose a geometric and semantic self-supervised pre-training method, GASP, that learns a unified representation by predicting, at any queried future point in spacetime, (1) general occupancy, capturing the evolving structure of the 3D scene; (2) ego occupancy,... (More)
Self-supervised pre-training based on next-token prediction has enabled large language models to capture the underlying structure of text, and has led to unprecedented performance on a large array of tasks when applied at scale. Similarly, autonomous driving generates vast amounts of spatiotemporal data, alluding to the possibility of harnessing scale to learn the underlying geometric and semantic structure of the environment and its evolution over time. In this direction, we propose a geometric and semantic self-supervised pre-training method, GASP, that learns a unified representation by predicting, at any queried future point in spacetime, (1) general occupancy, capturing the evolving structure of the 3D scene; (2) ego occupancy, modeling the ego vehicle path through the environment; and (3) distilled high-level features from a vision foundation model. By modeling geometric and semantic 4D occupancy fields instead of raw sensor measurements, the model learns a structured, generalizable representation of the environment and its evolution through time. We validate GASP on multiple autonomous driving benchmarks, demonstrating significant improvements in semantic occupancy forecasting, online mapping, and ego trajectory prediction. Our results demonstrate that continuous 4D geometric and semantic occupancy prediction provides a scalable and effective pre-training paradigm for autonomous driving. For code and additional visualizations, see our project page.
(Less)
- author
- Ljungbergh, William
; Lilja, Adam
; Tonderski, Adam
LU
; Ling, Arvid Laveno
; Lindstrom, Carl
; Verbeke, Willem
; Fu, Junsheng
; Petersson, Christoffer
; Hammarstrand, Lars
and Felsberg, Michael
- organization
- publishing date
- 2026
- type
- Chapter in Book/Report/Conference proceeding
- publication status
- published
- subject
- host publication
- Proceedings - 2026 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2026
- series title
- Proceedings - 2026 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2026
- pages
- 11 pages
- publisher
- IEEE - Institute of Electrical and Electronics Engineers Inc.
- conference name
- 2026 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2026
- conference location
- Tucson, United States
- conference dates
- 2026-03-06 - 2026-03-10
- external identifiers
-
- scopus:105041302017
- ISBN
- 9798331555115
- DOI
- 10.1109/WACV61042.2026.00301
- language
- English
- LU publication?
- yes
- id
- 0461cf55-44f0-460f-a076-60bd001b9ea3
- date added to LUP
- 2026-09-17 15:37:51
- date last changed
- 2026-09-17 15:38:09
@inproceedings{0461cf55-44f0-460f-a076-60bd001b9ea3,
abstract = {{<p>Self-supervised pre-training based on next-token prediction has enabled large language models to capture the underlying structure of text, and has led to unprecedented performance on a large array of tasks when applied at scale. Similarly, autonomous driving generates vast amounts of spatiotemporal data, alluding to the possibility of harnessing scale to learn the underlying geometric and semantic structure of the environment and its evolution over time. In this direction, we propose a geometric and semantic self-supervised pre-training method, GASP, that learns a unified representation by predicting, at any queried future point in spacetime, (1) general occupancy, capturing the evolving structure of the 3D scene; (2) ego occupancy, modeling the ego vehicle path through the environment; and (3) distilled high-level features from a vision foundation model. By modeling geometric and semantic 4D occupancy fields instead of raw sensor measurements, the model learns a structured, generalizable representation of the environment and its evolution through time. We validate GASP on multiple autonomous driving benchmarks, demonstrating significant improvements in semantic occupancy forecasting, online mapping, and ego trajectory prediction. Our results demonstrate that continuous 4D geometric and semantic occupancy prediction provides a scalable and effective pre-training paradigm for autonomous driving. For code and additional visualizations, see our project page.</p>}},
author = {{Ljungbergh, William and Lilja, Adam and Tonderski, Adam and Ling, Arvid Laveno and Lindstrom, Carl and Verbeke, Willem and Fu, Junsheng and Petersson, Christoffer and Hammarstrand, Lars and Felsberg, Michael}},
booktitle = {{Proceedings - 2026 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2026}},
isbn = {{9798331555115}},
language = {{eng}},
pages = {{3077--3087}},
publisher = {{IEEE - Institute of Electrical and Electronics Engineers Inc.}},
series = {{Proceedings - 2026 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2026}},
title = {{GASP : Unifying Geometric and Semantic Self-Supervised Pre-training for Autonomous Driving}},
url = {{http://dx.doi.org/10.1109/WACV61042.2026.00301}},
doi = {{10.1109/WACV61042.2026.00301}},
year = {{2026}},
}