Cited 0 time in
Cited 0 time in
A stochastic approach for attack resilient UAV motion planning
- A stochastic approach for attack resilient UAV motion planning
- Bezzo, Nicola; Weimer, James; Du, Yanwei; Sokolsky, Oleg; Son, Sang H.; Lee, Insup
- DGIST Authors
- Son, Sang H.
- Issue Date
- 2016 American Control Conference, ACC 2016, 2016-July, 1366-1372
- Article Type
- Conference Paper
- In this paper we propose a stochastic strategy for motion planning of unmanned aerial vehicles (UAVs) subject to malicious cyber attacks. By injecting erroneous information and compromising sensor data, an attacker can hijack a system driving it to unsafe states. In this work we bear on the problem of choosing optimal actions while one or more sensors are not reliable. We assume that the system is fully observable and one or more measurements (however unknown) return incorrect estimates of a state. We build an algorithm that leverages the theory of Markov decision processes (MDPs) to determine the optimal policy to plan the motion of a UAV and avoid unsafe regions of a state space. We name Redundant Observable MDPs (ROMDPs) this class of markovian processes that deal with redundant attacked measurements. A quadrotor case study is introduced and simulation and experimental results are presented to validate the proposed strategy. © 2016 American Automatic Control Council (AACC).
- Institute of Electrical and Electronics Engineers Inc.
- Related Researcher
Son, Sang Hyuk
RTCPS(Real-Time Cyber-Physical Systems Research) Lab
There are no files associated with this item.
- ETC2. Conference Papers
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.