Committee:

Dr. Magnus Egerstedt, ECE, Chair , Advisor

Dr. Samuel Coogan, ECE

Dr. Yorai Wardi, ECE

Dr. Charles Pippin, GTRI

Dr. Zsolt Kira, IC

Abstract: Robotics is now being applied to a diversity of real-world applications where safety is a critical consideration for continued adoption. In this thesis we therefore investigate how to develop algorithms that improve the safety of a system without unduly limiting overall performance. We motivate the work with the example of fixed wing collision avoidance and find that using a model allows for collision avoidance to be guaranteed for arbitrarily many vehicles even under limited communication and sensing constraints. We then employ model free approach and show that this enables the system to be more performant while still improving safety over the original system.