Skip to content

Commit

Permalink
Correct typos and update video
Browse files Browse the repository at this point in the history
  • Loading branch information
wljungbergh committed Mar 28, 2024
1 parent e107651 commit d29e114
Show file tree
Hide file tree
Showing 3 changed files with 1 addition and 1 deletion.
Binary file modified _publications/neuro-ncap/assets/metrics-transparent.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file modified _publications/neuro-ncap/assets/videos/teaser-video.mp4
Binary file not shown.
2 changes: 1 addition & 1 deletion _publications/neuro-ncap/neuro-ncap.md
Original file line number Diff line number Diff line change
Expand Up @@ -21,7 +21,7 @@ n_equal_contrib: 2


# Core idea
The core idea in NeuroNCAP is to **leverage NeRFs** for autonomous driving <d-cite key="tonderski2023neurad"></d-cite> **to realistically simulate safety-critical scenarios** from a sequence of real-world data **and evaluate entire AD systems in closed-loop**. To this end, we release an **open-source framework**, and evaluate two state-of-the-art end-to-end planners, UniAD <d-cite key="hu2023planning"></d-cite> and VAD <d-cite key="jiang2023vad"></d-cite>, in our safety-critical NeuroNCAP scenarios. We find that they **fail to avoid collisions**, even though they accurately perceive the safety-critical actor. We hope that NeuroNCAP will be used as a **benchmark for evaluating the safety of autonomous driving models**.
The core idea in NeuroNCAP is to **leverage NeRFs** for autonomous driving <d-cite key="tonderski2023neurad"></d-cite> **to realistically simulate safety-critical scenarios** from a sequence of real-world data **and evaluate end-to-end AD systems in closed-loop**. To this end, we release an **open-source framework**, and evaluate two state-of-the-art end-to-end planners, UniAD <d-cite key="hu2023planning"></d-cite> and VAD <d-cite key="jiang2023vad"></d-cite>, on our safety-critical NeuroNCAP scenarios. We find that they **fail to avoid collisions**, even though they accurately perceive the safety-critical actor. We hope that NeuroNCAP will be used as a **benchmark for evaluating the safety of autonomous driving models**.

<div style="display: flex; align-items: center; justify-content: center;">
<video controls loop muted playsinline style="width: 70%;" poster="assets/neuro-ncap-logo.png">
Expand Down

0 comments on commit d29e114

Please sign in to comment.