The use of robotic agents, such as unmanned aerial vehicles (UAVs) or unmanned ground vehicles (UGVs), has motivated the development of numerous autonomous cooperative task allocation and planning methods for heterogeneous networked teams. Typically agents within the team have different roles and responsibilities, and ensuring proper coordination between them is critical for efficient mission execution. However, as the number of agents, system components, and mission tasks increase, planning for such teams becomes increasingly complex, motivating the development of algorithms that can operate in real-time dynamic environments.
Given the complexity of the cooperative missions considered, there have been numerous solution approaches developed in recent years. This chapter provides an overview of three of the most common planning frameworks: integer programming, Markov decision processes, and game theory. The chapter also considers various architectural decisions that must be addressed when implementing online planning systems for multi-agent teams, providing insights on when centralized, distributed, and decentralized architectures might be good choices for a given application, and how to organize the communication and computation to achieve desired mission performance. Algorithms that can be utilized within the various architectures are identified and discussed, and future directions for research are suggested.
Cooperative Mission Planning for Multi-UAV Teams
Handbook of Unmanned Aerial Vehicles ; Chapter : 16 ; 1447-1490
2014-08-09
44 pages
Article/Chapter (Book)
Electronic Resource
English
Cooperative mission execution and planning
SPIE | 1998
|Planning to Fail: Mission Design for Modular Repairable Robot Teams
British Library Conference Proceedings | 2005
|