The Drone Detection and Communication Service is a conceptual solution that will address surveillance and detection capabilities integrating AI models and communication mechanisms for detecting drones flying on restricted areas of urban environments. This solution aims to develop an AI-powered drone detection system in an urban setting, focusing on monitoring and identifying unauthorized drone activity within restricted airspaces around critical infrastructure sites using affordable technology like cameras and arrays of microphones. The operational scope encompasses a diverse group of operators managing a fleet of drones, including those authorized to operate within the monitored airspace and others that deliberately intrude into prohibited zones. The goals of the solution are to use AI models based on multi sensor data such as visual and audio for identifying and differentiating between authorized and unauthorized drone activities within the monitored airspace, and to establish the communication methods and protocols for contacting drone operators. The validation of the solution assesses the real-time performance of the system under various conditions in simulated environments mimicking operation conditions that could hinder its efficacy. The aim is to demonstrate the reliability of artificial intelligence techniques to detect rogue or noncompliant drones, which is of high relevance to protect key infrastructures and identify attacks or access and privacy violations. This solution provides highly accurate drone detection capabilities while reducing costs by using cost-effective sensor data.
Validation scenarios
The scenarios of solution consider urban area with complex layout and simple layout to test the drone detection system’s capabilities across a spectrum of urban environments. These scenarios are designed to test the drone detection system’s capabilities under various conditions, ensuring its potential effectiveness in real-world applications.
Validation Scenario 1: Urban Area with Simple Layout. This scenario simulates an urban environment with a simpler layout, where the focus is on the effectiveness of the drone detection system in less congested areas and a reduced number of flight plans and no complex communication situations such as drones with flight plan authorized to fly over the restricted area, since in these situations it is not necessary to contact the operator when the drone is detected.
Validation Scenario 2: Urban Area with Complex Layout. This scenario simulates an urban environment characterized by a complex layout, including multiple high-rise buildings, busy streets, potential obstructions to drone detection, and complex situations such as drones with an approved flight plan not crossing the restricted area where the drone is being detected, which implies having to contact the operator to correct the situation.
KPAs: Safety, Security, Interoperability
Flight plans for validation. Two flight plans crossing restricted area, and three plans not crossing.
Flight plans for validation. Two flight plans crossing restricted area, and three plans not crossing.
Simulation of complex scenario with audio and video detectors running in parallel, and communication service listening and communicating event.
Simulation of complex scenario with audio and video detectors running in parallel, and communication service listening and communicating event.
Graphical user interface of the drone detection model using optical sensor (camera)
Graphical user interface of the drone detection model using optical sensor (camera)
Contact info:
Enrique Puertas, european University of Madrid, enrique.puertas@universidadeuropea.es
On January 7th 2024, during the first AI4HyDrop Workshop, we invited the participants to contribute with their suggestions for U-plan data and format. We will listen to your feedback along the whole project (around the end on 2025), to reflect your ideas in a final report.
However, during the validations, the final users reported different needs that had to be reflected also in the flight plan, like the possibility to define geo-cages, to fly at a certain altitude, or to define segregated volumes in the origin or destination for the fixed wings to execute their climb/descend manoeuvres. There were also technical limitations and the desire to specify preferences to be allowed if possible.
In Labyrinth, it was selected JSON as the format for the messages to communicate with the U-space.
Same as in the project DACUS, Labyrinth selected the GeoJSON standard to represent the flight plans. It allows to describe the trajectories as a sequence of points and areas or volumes.
In GeoJSON, a large number or elements can be added to the coordinates, which allows to specify information and constraints associated to the waypoints.
However, inspired by proposals that suggest high level flight plan specifications, we could consider the possibility of using plans with the capability to reflect complex parts like conditions or loops.
The article in the slide was suggesting a high-level format that would be translated into a list of waypoints to be uploaded to the drone. In our case, a similar high-level specification could be sent to the Operation Plan Preparation/Optimisation service, which would convert it to a more low-level format to be processed by the Flight Authorization service. In the event that the Operation Plan Preparation/Optimisation service is not available, the option to directly send a simpler format should be possible.