AIDA: - AI based Drone-port Automation for higher productivity, scalability and removing human error

用于物流的自动无人机即将出现，它们代表了我们如何在英国各地运输货物的范式转变。然而，为了从这些自主飞行中受益，还必须有自主地面系统，否则，该业务很快就会变得商业上不...由于每一端都需要团队，因此是可行的。Inteliports和Motion Robotics正在开发自动化整个地面操作的机器人技术，以实现真正的端到端自主物流。AIDA项目将测试使用深度神经网络同步所有Inteliports机器人系统的协作，以最大限度地提高无人机港口之间的效率、准确性和互操作性。AIDA旨在消除物流循环中人类管理者所经历的认知复杂性，使人类能够在更高水平上进行管理监督，将无聊的重复性任务移交给机器人，从而提高生产力，可扩展性和安全性。AIDA在操作复杂度呈指数级复杂的多个无人机端口的网络内协作时特别有用。推进现有技术，这项研究将测试一个特定的解决方案的可行性，用于模拟深度强化学习训练环境，匹配真实的世界无人机端口，并且可以通过使用统计建模来有效地实现。此外，一旦完全训练，我们将展示经过训练的网络可以在真实的世界中驱动系统。

Autonomous drones for logistics are just around the corner and they represent a paradigm shift in how we will move goods around the UK. However, in order to benefit from these autonomous flights, there must be autonomous ground systems too, otherwise the operation quickly becomes commercially un-viable as teams of people are needed at each end.Inteliports and Motion Robotics is developing the robotics that automates the entire ground operation to allow for true end to end autonomous logistics.Project AIDA will test the feasibility of using deep neural networks to synchronise the collaboration of all Inteliports' robotic systems so as to maximise the objective of efficiency, accuracy and interoperability between drone ports.AIDA aims to remove the cognitive complexity experienced by human managers from the logistics loop, allowing humans to manage at a higher level of oversight, devolving boring repetitive tasks to the robots and thereby increasing productivity, scalability and safety.AIDA is especially useful when collaborating within a network of multiple drone-ports where the complexity of operations becomes exponentially complex.Advancing the state of the art, the study will test the feasibility of a specific solution for simulating the Deep Reinforcement Learning Training environment in such a way it perfectly matches the real world drone port and can be achieved efficiently though the use of statistical modelling.Moreover once fully trained we will show that the trained network can drive the systems in the real world.