Maneuvering Characteristics Control Systems using Discrete-Time MDPs

Abstract: An effective Maneuvering Characteristics Augmentation System (MCAS) will mitigate the risk of stalls during takeoff for Boeing 737 MAX 8 aircraft. In this paper, we develop a flight simulator for modeling the flight trajectory of an aircraft throughout takeoff. We then proceed to model the MCAS system aboard these aircraft as a discrete-time Markov Decision Process and subsequently compute the MDP’s optimal policy. Using this optimal policy, the MCAS system is able to prevent 95% of the stalls that would have otherwise occurred in absence of the MCAS system’s intervention. Read our paper here.