Reproduce “Inspection planning under execution uncertainty” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Shmuel David Alpert, Kiril Solovey, Itzik Klein, Oren Salzman

IRIS-U^2^ is an algorithm developed for autonomous inspection planning that directly addresses execution uncertainty by extending the IRIS framework with Monte Carlo sampling and modified graph search operations. This approach provides statistical guarantees on expected Points of Interest (POI) coverage, path length, and collision probability, demonstrating superior coverage and lower collision risk compared to deterministic methods in both synthetic and realistic bridge inspection scenarios.

Key insights

  • Integrating Monte Carlo simulations directly into an incremental roadmap search allows for probabilistic reasoning about execution uncertainty during inspection planning, yielding paths with quantifiable statistical guarantees.
  • By modifying node definitions and extension operations to account for uncertainty estimates, the search algorithm can effectively navigate a trade-off between maximizing coverage, minimizing path length, and limiting collision risk.
  • Statistical bounds, such as confidence intervals for coverage and collision probability, can guide the selection of algorithm parameters (like the number of MC samples `m`) to optimize planning time and achieve desired confidence levels.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.