Supervised control of cooperative multi-agent robotic vehicles

In this paper, we address some of technical challenges associated with supervised tactical mobility behaviors modeling and control of a group of cooperative multi-agent robotic vehicles. Cooperative task planning of robotic systems is a dynamic and complex problem and very challenging. The challenge comes from many avenues including task decomposition, assignment, resource allocation, and task execution and monitoring. Furthermore, the behavior-based control system requires a deliberative and reactive task planning capability for execution of cooperative tasks of mobile robots as well as adequate adaptability to learn new cooperative schemes to improve system performance in over-time. Another challenge of this problem is associated with sensory data gathering, distributed data processing, centralized/decentralized data fusion, and intelligent world perception modeling and comprehension of cooperative mobile robots. We propose a multilayered supervisory control architecture for coordinated task planning of a group of multi-agent robots. The supervisory controller allows control of robotic vehicles in either teleoperation, semi-autonomous, and/or autonomous modes.