Cooperative UAV Formation Flying with Stochastic Obstacle Avoidance
In this paper, the problem of navigation of UAVs in formation in an unknown environment is investigated. The vehicles are commanded to make a formation and move as required. During the flight, the UAVs are also required to avoid obstacles and collisions between them. To achieve these tasks, a two mode control strategy is proposed. The two modes are the Safe and the Danger modes. The safe mode is used when there are no obstacles in the environment and the danger mode is activated whenever there is a chance of collision or when there are obstacles in the path. The UAVs keep their formation in the Safe Mode and in the Danger Mode, they can break the formation and rejoin again once there are no obstacles. The control architecture is a two layered hierarchical structure in both modes. In the Safe mode, a controller with relative motion dynamics generates the path for the UAVs and in Danger mode, a decentralized algorithm using a slightly modified Grossberg Network is proposed for obstacle/collision avoidance. This algorithm uses the spatial geometry of the UAVs to generate trajectories. Note that this approach is scalable. The bottom layer in both the architectures uses a Model Predictive Control (MPC) based tracking controller. This controller tracks the reference generated by the upper layers. Numerical results are presented that demonstrate the potential of the approach.
X. Wang et al., "Cooperative UAV Formation Flying with Stochastic Obstacle Avoidance," AIAA Guidance, Navigation, and Control Conference and Exhibit, American Institute of Aeronautics and Astronautics (AIAA), Jan 2005.
Mechanical and Aerospace Engineering
Keywords and Phrases
UAV; Obstacle Avoidance; Cooperative Control; Optimal Control; Hierarchal Control; Collision Avoidance; Grossberg Neural Network; Visibility Graph; Model Predictive Control
Article - Conference proceedings
© 2005 American Institute of Aeronautics and Astronautics (AIAA), All rights reserved.