*********************************
There is now a CONTENT FREEZE for Mercury while we switch to a new platform. It began on Friday, March 10 at 6pm and will end on Wednesday, March 15 at noon. No new content can be created during this time, but all material in the system as of the beginning of the freeze will be migrated to the new platform, including users and groups. Functionally the new site is identical to the old one. webteam@gatech.edu
*********************************
Title: Constructive Barrier Certificates With Applications to Fixed-Wing Aircraft Collision Avoidance
Committee:
Dr. Egerstedt, Advisor
Dr. Coogan, Chair
Dr. Wardi
Dr. Pippin
Abstract:
The objective of the proposed research is to show how to use machine learning to ensure safe fixed-wing aircraft operations. In this proposal we discuss how to construct a barrier certificate for a control affine system subject to actuator constraints and motivate this discussion by examining collision avoidance for fixed-wing aircraft. In particular, we show the theoretical development in this proposal can be used to create a barrier certificate that can ensure that two vehicles will not collide. We then extend this development by discussing how to ensure that multiple safety constraints (e.g., ensure robot distances are above some threshold for all pairwise pairwise combinations of vehicles) can be simultaneously satisfied in a decentralized manner. To motivate the discussion, we analyze a fixed-wing collision avoidance scenario with more than two vehicles where the vehicles have limited actuator inputs and communication capabilities. We then develop a general method for ensuring multiple safety constraints can be satisfied in a decentralized way and validate the theoretical developments with a simulation of 20 vehicles that maintain safe distances from each other even though their nominal paths would otherwise cause a collision. Having shown that safety can be achieved for fixed-wing aircraft, the proposed work seeks to build on these results by showing how machine learning can be used to learn optimal evasive maneuvers that can be applied without losing safety guarantees
