Large Deviations in Safety-Critical Hamiltonian Systems with Probabilistic Initial Conditions

We address the problem of determining the least improbable deviations leading to an unsafe rare event in a weakly perturbed mechanical system with probabilistic initial conditions. These deviations are obtained as the solution to a variational problem formulated using rigorous approximation techniqu...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:arXiv.org 2024-05
Hauptverfasser: Gomez, Aitor R, Bujorianu, Manuela L, Wisniewski, Rafal
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:We address the problem of determining the least improbable deviations leading to an unsafe rare event in a weakly perturbed mechanical system with probabilistic initial conditions. These deviations are obtained as the solution to a variational problem formulated using rigorous approximation techniques grounded in the principles of large deviations theory. These types of results have been extended to accommodate stochastic uncertainty in the initial states, which is a common assumption in mechanical systems. Furthermore, we demonstrate the applicability of the method by solving the problem for a rare collision event between two space objects, i.e. a high-dimensional and non-linear problem, resulting in the most likely sample paths leading to the realization of the unsafe rare event. The solution is validated against the necessary conditions for optimality derived from the maximum principle. Access to these unsafe sample paths offers relevant information regarding the dangerous configurations of rare events and can be used to design control strategies to reduce the probability of realization.
ISSN:2331-8422