Cooperative Control via Task Load Balancing for Networked Uninhabited Autonomous Vehicles

In this paper we first define a mathematical model of the "plant" for the cooperative control problem for multiple uninhabited autonomous vehicles (UAVs). This includes a representation for the vehicles, environment, and communication network. Next, we define an approach to cooperative control that uses local UAV task planning and multi-UAV coordination via task load balancing over the communication network. Our approach is motivated by our desire to cope with significant imperfections in the communication network and uncertainty in the environment, and yet provide a scalable strategy which can be implemented in real time across a network of UAVs, each of which only has relatively low processing power. Our focus in this paper is on defining the cooperative controller problem in a mathematical framework that is familiar to a control engineer, studying properties of a load balancing strategy for cooperation, and then via simulations to identify key challenges when network influences dominate the problem.