FoSS: Modeling Long-Range Dependencies and Multimodal Uncertainty in Trajectory Prediction via Fourier–State Space Integration
Yizhou Huang, Gengze Jiang, Yihua Cheng, Kezhi Wang · 2026
Read FoSS: Modeling Long-Range Dependencies and Multimodal Uncertainty in Trajectory Prediction via Fourier–State Space Integration by Yizhou Huang, Gengze Jiang, Yihua Cheng, Kezhi Wang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-01284.txt.
Description
Huang and three colleagues present FoSS, a dual framework for driving trajectory prediction, joining frequency-domain reasoning and linear-time sequence modeling. One branch applies a discrete Fourier transform, splitting trajectories into amplitude parts for intent and phase parts for variation. It reorders spectra step by step and refines them with two picked state-space submodules. A time-domain picked module rebuilds self-attention behavior in linear time and fuses both views with a cross-attention layer. Learnable queries plus a weighted fusion head produce rival trajectories with motion doubt.
Catalog record
| CREATOR | Yizhou Huang, Gengze Jiang, Yihua Cheng, Kezhi Wang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37742 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.01284