Safe Unmanned Aviation Operations Using AI-Based Icing Cloud Identification

Icing is a safety challenge in all aviation, but especially in unmanned aviation. When a drone encounters a freezing cloud, the ice that accumulates in the drone's structures causes critical damage within tens of seconds. By detecting freezing conditions quickly and reliably, the safety of flying can be improved and the number of productive flight hours significantly increased. Although the effects of freezing are well known, sensors suitable for drones to identify icing conditions are not currently on the market. In the project, we will implement and test a sensor based on camera and neural network computing to identify icing conditions and determine the solution’s potential for commercialization.