UAV cooperative control

This paper addresses the development of cooperative rendezvous and cooperative target classification agents in a hierarchical distributed control system for unmanned aerospace vehicles. For cooperative rendezvous, a Voronoi based polygonal path is generated to minimize exposure to radar. The rendezvous agent minimizes team exposure from Individual coordination functions while satisfying stringent timing constraints. For cooperative target classification, templates are developed, optimal trajectories are followed, and adjacent vehicles are assigned to view at complementary aspect angles. Views axe statistically combined to maximize the probability of correct target classification over various aspect angles.

1 Introduction

The cooperative control problem is characterized by three attributes: 1) complexity, 2) information structure, and 3) uncertainty. To address complexity and, possibly to accommodate the prevailing information pattern, an innovative hierarchical decomposition is used [1], [2]. Figure 1 illustrates a general architecture for cooperative control and task apportionment among multiple vehicles. One control level and three decision levels are introduced. At decision level 1 is the vehicle agent that does path planning, trajectory generation, and maintains models of terrain, threats, and targets. At decision level 2 is the sub-team agent which coordinates the activities of any tasks that require more than one vehicle. to accomplish. If the sub-team has more than one task, then the agent apportions the vehicles to the tasks. Also, this agent does rendezvous coordination as well as sensor, weapon, decoy, and jamming coordination. The team agent, at decision level 3, is responsible for meeting the mission objective, determining sub-objectives for the sub-teams, and apportioning resources and tasks.