Balancing Mission Requirement for Networked Autonomous Rotorcrafts Performing Video Reconnaissance

Abstract : Current and emerging missions, including active surveillance and tracking, terminal guidance and search and rescue, require image-feedback from camera-equipped vehicles. Certain mission scenarios and sensor restrictions may require the collaboration of assets over an ad-hoc network. The development in this paper extends efforts to balance trade-offs between asset/sensor cone positioning to satisfy mission requirements and network requirements such as maintaining network connectivity. To address the trade-offs between asset positioning and network connectivity, a prioritized task-function based guidance law is developed for a simple scenario containing three assets tracking teams of mobile targets. One developed task-function maintains a communication network by ensuring the distance between the UAV s does not exceed a critical threshold. Additional task-functions enable assets to keep targets of interest in the image cone by regulating image features derived from the camera view. The UAV s are modeled as rotorcraft using an LTI model equipped with a gimbaled camera. Early simulation results are provided to examine the behavior of the assets for different configurations of objects observed by the asset cameras. Future efforts will seek to improve performance by modifying the control law, possibly adding time varying tasks shaped by trajectory planning techniques.

摘要：当前和新兴的任务，包括主动监视与跟踪、末端制导以及搜索与救援，都需要配备摄像头的飞行器提供图像反馈。某些任务场景和传感器限制可能需要在自组织网络上进行资源协作。本文的研究进展致力于在资源/传感器圆锥定位之间进行权衡，以满足任务需求和网络需求，比如维持网络连通性。为了解决资源定位和网络连通性之间的权衡问题，针对一个包含三个追踪移动目标团队的简单场景，制定了一种基于优先任务函数的制导律。一个制定的任务函数通过确保无人机之间的距离不超过临界阈值来维持通信网络。其他任务函数通过调节从摄像头视角得出的图像特征，使资源能够将感兴趣的目标保持在图像圆锥内。无人机被建模为使用配备有万向节摄像头的线性时不变（LTI）模型的旋翼飞行器。提供了早期的模拟结果，以检验资源在其摄像头所观测到的不同物体配置下的行为。未来的研究将寻求通过修改控制律来提高性能，可能会添加由轨迹规划技术塑造的时变任务。