VDM Waslander, S: Multi-Agent System Design

Engineering systems with multiple independent decision makers are becoming increasingly prevalent and present many challenging coordination problems. This work investigates two such systems in aerospace engineering: air traffic flow control and autonomous aerial vehicle teams. The air traffic flow control problem is investigated from the point of view of the airlines, which bear the often large costs of flight delays and reroutes made to ensure safe operation in adverse weather conditions. By allowing airlines to be directly involved in traffic flow decision making through resource markets, overall system performance can be improved while safety is ensured by the central authority. Airlines decision making is modeled as cooperative, competitive and strategic, and efficient, equitable mechanisms are designed for each situation. The autonomous aerial vehicle team problem is investigated through the development of the Stanford Testbed of Autonomous Rotorcraft for Multi-Agent Control (STARMAC). This world-leading testbed of quadrotor helicopters is used to demonstrate many novel coordination algorithms, including multi-vehicle autonomous conflict resolution.