aiqbench analysis
This paper presents a diagnostic-driven approach to solving the Sim2Real gap in autonomous driving by addressing three specific data distribution mismatches between synthetic training sets and real-world LiDAR sensors. The authors demonstrate that correcting coordinate frame misalignments, optimizing non-maximum suppression per object class, and modeling null return points significantly boost detection accuracy without altering the core network architecture. These findings offer critical insights for researchers aiming to deploy robust 3D perception systems trained solely on simulated environments.
Coordinate Frame Recovery, Per Class NMS and Null Return Modelling for Sim2Real V2X LiDAR Detection
arXiv:2611.03940v1 Announce Type: cross Abstract: We describe our approach to the V2X-Real track of the UrbanTwin Sim2Real LiDAR Challenge, in which a 3D object detector must be trained exclusively on synthetic data and evaluated on unseen real roadside LiDAR from an Ouster OS1-64 unit, while the synthetic point clouds are simultaneously scored for distributional realism. Our final entry scores 0.7457 combined, from a detection mAP of 0.4382 and a realism score of 0.9251. Three findings account for most of our result, and all three were diagnostic rather than architectural. First, a coordinate frame mismatch between our predictions and the evaluation ground truth was suppressing detection AP entirely; recovering it took our score from 0.3520 to 0.5147 in a single submission. Second, a single global non maximum suppression threshold was destroying genuine pedestrian detections, because pedestrians occur in groups and a 0.3 IoU threshold applied to a 0.7 m box suppresses distinct neighbouring people; correcting this per class raised overall mAP from 0.0751 to 0.4295. Third, approximately 31 percent of the real reference point cloud consists of null returns encoded as the literal point (0, 0, 0), a format feature entirely absent from our synthetic clouds, which pinned Earth Mover's Distance near its floor until we modelled it. Negative results are reported at equal length, and Section 9 gives a full account of data provenance and compliance. That section also accounts for every mechanism contributing to the submitted score.
Abstract:We describe our approach to the V2X-Real track of the UrbanTwin Sim2Real LiDAR Challenge, in which a 3D object detector must be trained exclusively on synthetic data and evaluated on unseen real roadside LiDAR from an Ouster OS1-64 unit, while the synthetic point clouds are simultaneously scored for distributional realism. Our final entry scores 0.7457 combined, from a detection mAP of 0.4382 and a realism score of 0.9251. Three findings account for most of our result, and all three were diagnostic rather than architectural. First, a coordinate frame mismatch between our predictions and the evaluation ground truth was suppressing detection AP entirely; recovering it took our score from 0.3520 to 0.5147 in a single submission. Second, a single global non maximum suppression threshold was destroying genuine pedestrian detections, because pedestrians occur in groups and a 0.3 IoU threshold applied to a 0.7 m box suppresses distinct neighbouring people; correcting this per class raised overall mAP from 0.0751 to 0.4295. Third, approximately 31 percent of the real reference point cloud consists of null returns encoded as the literal point (0, 0, 0), a format feature entirely absent from our synthetic clouds, which pinned Earth Mover's Distance near its floor until we modelled it. Negative results are reported at equal length, and Section 9 gives a full account of data provenance and compliance. That section also accounts for every mechanism contributing to the submitted score.
8 pages. 1st place, Track 7 (UrbanTwin Sim2Real LiDAR, V2X-Real), DriveX Grand Challenge, ECCV 2026 Workshop. Code: this https URL
Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
arXiv:2611.03940 [cs.CV]
(or arXiv:2611.03940v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2611.03940 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 352 words
Related coverage
An Attempt at Top-down Interpretability
LessWrong · Oct 24, 2026
Exposure is Optional: Learning Unlike Coordination in Language Models
arXiv · Jul 23, 2026
AI swarms are starting to pose indirect takeover risk
AI Alignment Forum · Aug 12, 2026
Mixture of Wide Experts: Sparse Routing for Interpretable Native Coordinates
arXiv · Nov 5, 2026