Robust Kalman filter for heavy-tailed process and measurement noises
Qianxin Wang, Chenwang Ye, Guobin Chang, Conghao Tang, Xun Zhang · 2026
Read Robust Kalman filter for heavy-tailed process and measurement noises by Qianxin Wang, Chenwang Ye, Guobin Chang, Conghao Tang, Xun Zhang (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/pmc-13341795.txt.
Description
Five authors led by Qianxin Wang present a 2026 method to track moving targets when sensor noise is wild. Standard Kalman filters guess a target path from a model plus noisy reads, but they assume the noise is mild and bell-shaped. Real noise often has heavy tails, with rare large spikes that wreck the guess. Past fixes dealt with sensor noise but left noise in the motion model itself largely alone.
The new filter keeps the same simple form as the standard one and can start from its output. It is built to shrug off heavy tails in both kinds of noise at once. It can also take in prior facts about the noise when those are known and trusted. The team checked it on a simulated tracking task and on real data, and report that it held up well in both.
Catalog record
| CREATOR | Qianxin Wang, Chenwang Ye, Guobin Chang, Conghao Tang, Xun Zhang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 79 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |