PRISM: Efficient and Locally Optimal Probabilistic Planning with Reachability Guarantees

Alex Rose, Christopher Jewison, Jonathan P. How · 2026

Read PRISM: Efficient and Locally Optimal Probabilistic Planning with Reachability Guarantees by Alex Rose, Christopher Jewison, Jonathan P. How (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-2606-26413.txt.

Description

Alex Rose, Christopher Jewison, and Jonathan P. How present PRISM, a planner for robots unsure of their exact state. Belief-space planning works with a belief, a probability spread over possible states, instead of one known position. Motion noise and limits on states and controls make guarantees hard. Older planners sample points and link them into roadmaps. Coverage of the belief space often stays patchy, or routes grow costly.

PRISM rests on a new result about controlling state covariance under constraints. Covariance measures how spread out the uncertainty is. The planner splits the job into steering the average state and shrinking that spread. An online step then trims the cost of workable routes. The authors prove full coverage under mild assumptions on start and goal spreads. In hard simulated scenarios, PRISM reached complete coverage where rivals fell short. Its routes also cost less on average, with less variation.

Catalog record

CREATORAlex Rose, Christopher Jewison, Jonathan P. How
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE107 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.26413

Browse the full library · About the Tower of Babel