Intelligent Controll of an Autonomous Flying Vehicle

IRI-9528115蒂尔伯里，道恩密歇根大学$18，000 - 12个月 自主飞行器的智能控制 自主飞行器的智能控制该奖项为为期一年的研究规划拨款提供资金。 本研究计画的目的是探讨智能控制与机器人学的研究问题。 为此，最终将建造一个完全自主的飞行机器人，它可以在其环境中执行简单的任务。 正在调查的问题包括开发一个理论框架的两个级别的控制器之间的接口，将学习算法内的任务规划器，以提高其性能，因为它变得更有经验的环境中，并集成传感器数据与不同的控制器，以实现自动避障，同时确保目标的实现。 飞行机器人的额外自由度既增加了它的机动性，也使它更难控制。 与地面机器人相比，直升机必须始终处于自动控制之下。 它永远不可能停在原地，评估它当前的环境，计算一条新的路径，然后继续它的使命。 它不是静态稳定的，并且已知动态稳定问题是困难的。 如果这些困难可以通过智能控制来克服，结果将是一个高度灵活，非常可扩展的自主系统。

IRI-9528115 Tilbury, Dawn University of Michigan $18,000 - 12 mos Intelligent Controll of an Autonomous Flying Vehicle Intelligent Control of an Autonomous Flying Vehicle This award funds a one-year research planning grant. The goal of this research project is to explore research problems in intelligent contrml and robotics. For this purpose, a completely autonomous flying robot will ultimately be built which can carry out simple tasks within its environment. The issues under investigation include developing a theoretical framework for the interface between the two levels of controllers, incorporating a learning algorithm within the task planner to improve its performance as it becomes more experienced in its environment, and integrating the sensor data with the different controllers to achieve automatic obstacle avoidance while insuring that the goal is reached. The extra degrees of freedom of a flying robot both increase its maneuverability and make it more difficult to control. In contrast to a terrestrial-based robot, a helicopter must always be under automatic control. It can never stop in place, evaluate its current surroundings, compute a new path, and then continue its mission. It is not statically stable, and the dynamic stabilization problem is known to be hard. If these difficulties can be overcome through intelligent control, the result will be a highly-flexible, extremely maneuverable autonomous system.