Information Flow in Cooperative Control of Multi-Vehicle Systems

As the Air Force moves toward its vision of a mixed force of inhabited and uninhabited air and space vehicles, the role of integrated command, control, communications, and computation will continue to increase in importance. A key element in all future military systems will be the role of communications networks and the increased use of information ow as an integral part of military operations. In many applications, such as coordinated control of unmanned aerial vehicles (UAVs), the presence of a central or supervisory controller cannot be assumed or is undesirable, and the presence and reliability of intervehicle communication channels is itself subject to change. As such, there is a need for decentralized control and decision making strategies which are robust to changes in the underlying inter-vehicle communication topology. This proposal is focused on developing the underlying theory required to achieve integration of information ow into control analysis and design for cooperative tasks of multi-vehicle systems. By making use of tools from control theory, dynamical systems, and graph theory, we will develop a framework for analyzing the e ects of information and sensor topology on feedback systems and develop tools for designing information ow and control algorithms that build on this framework. We will apply these ideas to several test problems involving real-time, distributed control of a set of multiple vehicles performing cooperative tasks. In addition to computational exploration through simulation, we plan to implement our algorithms on a multi-vehicle, wireless testbed for networked control, communications, and computing that is being developed at Caltech under separate funding.