betaFeedback

WorksEngineering

Concentration-Based Probabilistic Reachability for Linear Systems under Multiplicative Uncertainty and Additive Noise

Abstract

This paper investigates probabilistic reachability for discrete-time linear stochastic systems subject to bounded multiplicative uncertainty and potentially unbounded additive noise. To address the conservatism of reachable sets constructed solely from covariance information, we develop two complementary approaches based on sub-Gaussian and sub-exponential matrix moment-generating-function characterizations. The sub-Gaussian approach relies on a worst-case envelope over the uncertainty support, whereas the sub-exponential approach exploits expectation-based propagation of the matrix variance proxy. Both approaches are used to construct probabilistic reachable sets with explicit probability guarantees, considering ellipsoidal confidence regions and half-space bounds. For the latter, we derive a necessary and sufficient condition under which the sub-exponential chance-constraint margin is no larger than its sub-Gaussian counterpart. Numerical results show that the proposed bounds can be less conservative than those based on the covariance, and illustrate the complementary nature of the two approaches across probability levels.

Paper details

vlnc:W_nr3z23mz
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
arXiv nonexclusive license
Usage

views · 12

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Engineering

ChatGPT (OpenAI)

View statement
Looking for related papers…