If Amazon were to deliver its 40 million Amazon Prime subscribers via drones, they are 130,000 flights of unmanned aerial vehicles that the U.S. should provide every day. A colossal traffic knowing that it takes already 15,000 air traffic controllers to manage 87,000 flights over the United States. Unable to hire tens of thousands of controllers to handle the incessant flea jumps of these drones of delivery. Stanford researchers are working on algorithms that will coordinate this deluge of drones but Airborne Concept french and Egis will bring a brick that is essential to the management of this traffic.
A one hundred grams only ADS-B transponder
The demonstration took place on the airport of Toulouse-Francazal, end September 2015. A multidirectional less drone 10 kg was equipped with a transponder of ADS-B type. It is this type of equipment that allows air traffic controllers to have on their screens the spots of each airliner. The chosen transponder was small enough to do not too grever drone performance but nevertheless allowed to very closely follow the movements of the drone. From the french Airborne Concept and Egis submitted to the Directorate General of Civil Aviation (DGCA) will allow the supply air controls of the future software algorithms. Nasa had already installed an ADS-B transponder on a drone in 2012 but it was a “Predator”, a drone otherwise more imposing than the small octoptere used in Toulouse. The transponder chosen by Nasa, type Garmin GDL-90 ADS-B was a standard equipment of civil aviation while the french used a transponder of one hundred grams only. It is a track on which Google is also working on his side. Now, these are potentially all drones that could be placed under the authority of air traffic control.
Stanford testing algorithms collision for UAVs
Once all the drones more than 10 kg equipped or even less, how to handle the tens of thousands of devices that will potentially across the sky? They are going to have a few minutes flights only much will converge to the same areas of supply such as Amazon or Google warehouses. Manage traffic in the order of tens of thousands of devices must be automated and it is the subject of research of a Stanford team that develops software of air traffic control (Unmanned Aircraft System (UAS) Traffic Management (UTM) that will allow the drones to move without risk of collision. Just as captains have on their dashboard of a screen that indicates the position of nearby (TCAS) flights and possibly those whose path between potentially collide with that of their device, this is no longer an air traffic controller or the pilot last spring. NASA had imagined a system where the drones were to follow air corridors GPS. The approach is nonetheless little adapted to urban areas where the density of drones was going to cause multiple alerts. PUT it imagined the ACAS X system where calculations of trajectory of each drone to generate an alert in case of possible collision. At a human operator to divert one of two drones to prevent the accident. The Stanford Intelligent Systems Laboratory team went further by imagining automatic avoidance algorithms. The software calculates the trajectories of two drones avoidance and it seems to work. They carried one million simulations of encounter between 2 up to 10 devices successfully. Calculations by pairs seem more effective than simple calculations of avoidance drone by drone but developments continue. Researchers particularly want to integrate additional variables such as communications, rogue drones failures in the system or the weather in their algorithms. They hope to see them implemented in the UTM system that NASA is finalizing for 2019.
Translator : Bing Translation
“The drone equipped with a transponder miniature object of the demonstration which is carried out in Toulouse”, Egis press release, Sepember 24, 2015
“Google targets low-cost ADS-B Out avionics market”, FlightGlobal, March 24 2015
“NASA Flight Tests New ADS-B Device on Ikhana UAS”, Communiqué de la Nasa, March 23, 2012