Title: Flight Safety Assessment and Management
Speaker: Swee Balachandran
Date: Wednesday, May 11, 2016
Time: 11:00 – 12:00
Location: NIA Room 137
Abstract: Mr. Balachandran will be discussing his dissertation work on the development of a novel Flight Safety Assessment and Management (FSAM) system. FSAM is a decision making system whose goal is to monitor flight conditions for loss of control risk and promptly override a flight crew or a nominal autopilot system with an adaptive (potentially hard to verify) control architecture that can mitigate risk effectively. My dissertation explores the use of decision theoretic approaches – Markov Decision Processes (MDP) and Constrained MDPs to formulate FSAM. Sparse approximate solutions to the MDPs are also explored to support fast online decision making. The use of a cross-entropy based verification technique to quantitatively analyze the performance of the FSAM decision making system will also be discussed.
Bio: Swee Balachandran candidate in the Aerospace Engineering Department at the University of Michigan. He is passionate about robotics. Specifically, interested in using tools in artificial intelligence, machine learning and optimal control to make intelligent decision making systems that can facilitate the autonomous operation of unmanned vehicles.