OccSim: Multi-kilometer Simulation with Long-horizon Occupancy World Models
OccSim: Multi-kilometer Simulation with Long-horizon Occupancy World Models
OccSim is a data-driven autonomous driving simulation framework that eliminates the need for HD maps or pre-recorded logs. It generates over 3000 continuous frames and maps spanning more than 9 kilometers, with the potential to extend beyond 110K frames. The system uses a static occupancy world model and a Layout Generator to synthesize diverse simulation data, achieving up to 74% zero-shot performance on unseen data when scaled.
This paper introduces OccSim, a novel 3D occupancy world model that enables long-horizon simulation without relying on HD maps or pre-recorded logs. It achieves an 80x improvement in stable generation length and can generate over 9 kilometers of synthetic environment maps.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
improvement in stable generation length
80
View original evidence
rst 3D occupancy world model driven simulator that requires neither HD maps nor log-derived layouts. OccSim obviates the requirement for continuous logs or HD maps; conditioned only on a single initial frame and a sequence of future ego-actions. It can stably generate over 3000 continuous frames, an 80x improvement in stable generation length over previous SoTA occupancy world models, and fuses them into maps spannin
Open source S3Frames
110,000 frames
View original evidence
g over 9 kilometers. Further qualitative stress tests show that it can roll out beyond 110K frames. OccSim is powered by two modules: W-DiT based static occupancy world model and the Layout Generator. W-DiT handles the ultra-long-horizon generation of static environments by explicitly introducing known rigid transformations in architecture design, while the Layout Generator populates the dynamic foreground with react
Open source S4data size scaling factor
≤5
View original evidence
aled up to 5x its original size, zero-shot performance increases to 74%.
Open source S6
- paper: not_reported
- code: not_reported
- weights: not_reported
- dataset: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
rst 3D occupancy world model driven simulator that requires neither HD maps nor log-derived layouts. OccSim obviates the requirement for continuous logs or HD maps; conditioned only on a single initial frame and a sequence of future ego-actions. It can stably generate over 3000 continuous frames, an 80x improvement in stable generation length over previous SoTA occupancy world models, and fuses them into maps spannin
Open source S3
g over 9 kilometers. Further qualitative stress tests show that it can roll out beyond 110K frames. OccSim is powered by two modules: W-DiT based static occupancy world model and the Layout Generator. W-DiT handles the ultra-long-horizon generation of static environments by explicitly introducing known rigid transformations in architecture design, while the Layout Generator populates the dynamic foreground with react
Open source S4
ive agents based on the synthesized road topology. With these designs, OccSim can synthesize massive, diverse simulation streams. Extensive experiments demonstrate its downstream utility: data collected directly from OccSim can pre-train 4D semantic occupancy forecasting models to achieve up to 67% zero-shot performance on unseen data, outperforming previous asset-based simulator by 11%. When the collected data is sc
Open source S5
aled up to 5x its original size, zero-shot performance increases to 74%.
Open source S6
Source：arXiv Robotics — research abstracts · arxiv.org