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问世，它们代表着我们将在英国各地运送货物的方式发生了范式转变。然而，为了从这些自主飞行中受益，还必须有自主的地面系统，否则操作很快就会变得不可行，因为两端都需要团队。Inteliports and Motion Robotics正在开发自动化整个地面操作的机器人，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Inteliport机器人系统协作的可行性，以最大限度地提高效率、准确性和无人机端口之间的互操作性。AIDA旨在从物流循环中消除人类经理经历的认知复杂性，允许人类在更高级别的监督下进行管理。将枯燥的重复性任务交给机器人，从而提高生产率、可扩展性和安全性。AIDA在操作复杂性变得指数级复杂的多个无人机端口网络内协作时尤其有用。先进的技术水平，这项研究将测试模拟深度强化学习训练环境的特定解决方案的可行性，该解决方案与真实世界的无人机端口完美匹配，并可以通过使用统计建模高效实现。此外，一旦经过充分训练，我们将展示训练的网络可以在现实世界中驱动系统。

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.