An Autonomous Override System to Prevent Airborne Loss of Control

Authors

  • Sweewarman Balachandran University of Michigan
  • Ella M. Atkins University of Michigan

DOI:

https://doi.org/10.1609/aaai.v30i2.19074

Abstract

Loss of Control (LOC) is the most common precursor to aircraft accidents. This paper presents a Flight Safety Assessment and Management (FSAM) decision system to reduce in-flight LOC risk. FSAM nominally serves as a monitor to detect conditions that pose LOC risk, automatically activating the appropriate control authority if necessary to prevent LOC and restore a safe operational state. This paper contributes an efficient Markov Decision Process (MDP) formulation for FSAM. The state features capture risk associated with aircraft dynamics, configuration, health, pilot behavior and weather. The reward function trades cost of inaction against the cost of overriding the current control authority. A sparse sampling algorithm obtains a near-optimal solution for the MDP online. This approach enables the FSAM MDP to incorporate dynamically changing flight envelope and environment constraints into decision-making. Case studies based on realworld aviation incidents are presented.

Downloads

Published

2016-02-18

How to Cite

Balachandran, S., & Atkins, E. (2016). An Autonomous Override System to Prevent Airborne Loss of Control. Proceedings of the AAAI Conference on Artificial Intelligence, 30(2), 3991-3996. https://doi.org/10.1609/aaai.v30i2.19074