NeuroNCAP : Photorealistic Closed-Loop Safety Testing for Autonomous Driving
Ljungbergh, William ; Tonderski, Adam LU
; Johnander, Joakim
; Caesar, Holger
; Åström, Kalle
LU
; Felsberg, Michael
and Petersson, Christoffer
(2025)
18th European Conference on Computer Vision, ECCV 2024
In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
15088 LNCS.
p.161-177
- Abstract
We present a versatile NeRF-based simulator for testing autonomous driving (AD) software systems, designed with a focus on sensor-realistic closed-loop evaluation and the creation of safety-critical scenarios. The simulator learns from sequences of real-world driving sensor data and enables reconfigurations and renderings of new, unseen scenarios. In this work, we use our simulator to test the responses of AD models to safety-critical scenarios inspired by the European New Car Assessment Programme (Euro NCAP). Our evaluation reveals that, while state-of-the-art end-to-end planners excel in nominal driving scenarios in an open-loop setting, they exhibit critical flaws when navigating our safety-critical scenarios in a closed-loop... (More)
We present a versatile NeRF-based simulator for testing autonomous driving (AD) software systems, designed with a focus on sensor-realistic closed-loop evaluation and the creation of safety-critical scenarios. The simulator learns from sequences of real-world driving sensor data and enables reconfigurations and renderings of new, unseen scenarios. In this work, we use our simulator to test the responses of AD models to safety-critical scenarios inspired by the European New Car Assessment Programme (Euro NCAP). Our evaluation reveals that, while state-of-the-art end-to-end planners excel in nominal driving scenarios in an open-loop setting, they exhibit critical flaws when navigating our safety-critical scenarios in a closed-loop setting. This highlights the need for advancements in the safety and real-world usability of end-to-end planners. By publicly releasing our simulator and scenarios as an easy-to-run evaluation suite, we invite the research community to explore, refine, and validate their AD models in controlled, yet highly configurable and challenging sensor-realistic environments.
(Less)
- author
- Ljungbergh, William
; Tonderski, Adam
LU
; Johnander, Joakim
; Caesar, Holger
; Åström, Kalle
LU
; Felsberg, Michael
and Petersson, Christoffer
- organization
- publishing date
- 2025
- type
- Chapter in Book/Report/Conference proceeding
- publication status
- published
- subject
- keywords
- Autonomous driving, Closed-loop simulation, Neural rendering, Trajectory planning
- host publication
- Computer Vision – ECCV 2024 - 18th European Conference, Proceedings
- series title
- Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
- editor
- Leonardis, Aleš ; Ricci, Elisa ; Roth, Stefan ; Russakovsky, Olga ; Sattler, Torsten and Varol, Gül
- volume
- 15088 LNCS
- pages
- 17 pages
- publisher
- Springer
- conference name
- 18th European Conference on Computer Vision, ECCV 2024
- conference location
- Milan, Italy
- conference dates
- 2024-09-29 - 2024-10-04
- external identifiers
-
- scopus:85208599732
- ISSN
- 1611-3349
- 0302-9743
- ISBN
- 9783031734045
- 9783031734038
- DOI
- 10.1007/978-3-031-73404-5_10
- language
- English
- LU publication?
- yes
- additional info
- Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.
- id
- d8d58617-68d6-4d50-a474-c1af0392bc51
- date added to LUP
- 2025-01-31 10:21:32
- date last changed
- 2025-07-04 23:09:12
@inproceedings{d8d58617-68d6-4d50-a474-c1af0392bc51,
abstract = {{<p>We present a versatile NeRF-based simulator for testing autonomous driving (AD) software systems, designed with a focus on sensor-realistic closed-loop evaluation and the creation of safety-critical scenarios. The simulator learns from sequences of real-world driving sensor data and enables reconfigurations and renderings of new, unseen scenarios. In this work, we use our simulator to test the responses of AD models to safety-critical scenarios inspired by the European New Car Assessment Programme (Euro NCAP). Our evaluation reveals that, while state-of-the-art end-to-end planners excel in nominal driving scenarios in an open-loop setting, they exhibit critical flaws when navigating our safety-critical scenarios in a closed-loop setting. This highlights the need for advancements in the safety and real-world usability of end-to-end planners. By publicly releasing our simulator and scenarios as an easy-to-run evaluation suite, we invite the research community to explore, refine, and validate their AD models in controlled, yet highly configurable and challenging sensor-realistic environments.</p>}},
author = {{Ljungbergh, William and Tonderski, Adam and Johnander, Joakim and Caesar, Holger and Åström, Kalle and Felsberg, Michael and Petersson, Christoffer}},
booktitle = {{Computer Vision – ECCV 2024 - 18th European Conference, Proceedings}},
editor = {{Leonardis, Aleš and Ricci, Elisa and Roth, Stefan and Russakovsky, Olga and Sattler, Torsten and Varol, Gül}},
isbn = {{9783031734045}},
issn = {{1611-3349}},
keywords = {{Autonomous driving; Closed-loop simulation; Neural rendering; Trajectory planning}},
language = {{eng}},
pages = {{161--177}},
publisher = {{Springer}},
series = {{Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)}},
title = {{NeuroNCAP : Photorealistic Closed-Loop Safety Testing for Autonomous Driving}},
url = {{http://dx.doi.org/10.1007/978-3-031-73404-5_10}},
doi = {{10.1007/978-3-031-73404-5_10}},
volume = {{15088 LNCS}},
year = {{2025}},
}