Secure planning against stealthy attacks on robot actuators in unknown stochastic environments is achieved by modeling the problem as a stochastic game and solving it with model-free RL to satisfy a combined LTL formula.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.RO 2years
2020 2verdicts
UNVERDICTED 2representative citing papers
A model-free RL methodology is developed to maximize the probability of LTL satisfaction in unknown stochastic games when the derived DRA has a single Rabin pair, with a generalization providing lower bounds for multiple pairs.
citing papers explorer
-
Secure Planning Against Stealthy Attacks via Model-Free Reinforcement Learning
Secure planning against stealthy attacks on robot actuators in unknown stochastic environments is achieved by modeling the problem as a stochastic game and solving it with model-free RL to satisfy a combined LTL formula.
-
Model-Free Reinforcement Learning for Stochastic Games with Linear Temporal Logic Objectives
A model-free RL methodology is developed to maximize the probability of LTL satisfaction in unknown stochastic games when the derived DRA has a single Rabin pair, with a generalization providing lower bounds for multiple pairs.