CBL - Campus del Baix Llobregat

Projecte llegit

Títol: Evaluation of Remain Well Clear and Collision Avoidance for Drones

Estudiants que han llegit aquest projecte:

Director/a: PASTOR LLORENS, ENRIC

Departament: DAC

Títol: Evaluation of Remain Well Clear and Collision Avoidance for Drones

Data inici oferta: 27-07-2020     Data finalització oferta: 27-03-2021


Estudis d'assignació del projecte:
    GR ENG SIST AEROESP
    GR ENG TELEMÀTICA
Tipus: Coordinat    Títol: Cloud-based Remain Well Clear and Collision Avoidance for Drones - (CA)
 
Lloc de realització: EETAC
 
Paraules clau:
Remain Well Clear, Collision avoidance, drones, U-Space, Detect and Avoid, Encounter Model
 
Descripció del contingut i pla d'activitats:
The integration of drones in non-segregated airspace requires
the implementation of two key functions: the remain well clear
and the collision avoidance functions. Remain well clear
intended to keep vehicles separated a safe distance from each
other, while collision avoidance is intended as a last ditch
system providing directions to avoid imminent collisions.

Integrating this functions into drones is complicated, as
their capacity to carry additional avionics is limited.

This project intends to address the problem from a different
perspective; that is, moving the responsibility of
implementing those functions to the cloud. With existing
technology, drones can be connected to internet through 4G/5G
technology, thus being able to create a continuous flow of
positioning information. Services on the cloud can determine
if any breach on the RWC or CA functions may happen, thus
providing the necessary traffic advisory to avoid the
conflict.

The project will implement a prototype of both functions based
on the already existing 4G capabilities and the exploitation f
miniaturized ADS-B devices.
 
Overview (resum en anglès):
One of the cornerstones that should enable inserting unmanned aircraft into the airspace is the development of Detect and Avoid (DAA) systems. DAA systems will improve the Remote Pilot (RP) situational awareness by means of electronic conspicuity devices, providing them with the necessary means to Remain Well Clear (RWC) from other traffic and, if necessary, avoid Mid-Air collisions (MAC). DAA systems will compensate for the loss of a pilot on board, which drastically reduces the capacity to keep a safe separation from traffic, making current Rules of the Air very challenging to achieve.

Given the growing popularity of drone operations for commercial and recreational purposes, new standards should include them in the not-too-distant future. Since current DAA standards and algorithms (DO-365 and ED-258) are being developed targeting large, mostly military Remotely Piloted Aircraft Systems (RPAS), this project proposes a new set of detection volumes and alert thresholds for U-Space users according to an aircraft type classification. This will allow adapting the existing DAA algorithms to small drones, complying with the new European framework of services and applications for drones (U-Space).

Because testing new safety nets (such as new DAA algorithms) on real aircraft would be dangerous and inadequate, radar reports and computer-based simulations allow for a risk-free and faster evaluation of safety net performances. Due to the current lack of real drone radar tracks, this project has developed a multi-rotor drone encounter generator tool (called DEG). This software is able to generate a large number of synthetic pairwise quadcopter drone conflict tracks, simulating the instant prior to a MAC. The way trajectories are generated by DEG strongly depends on the type of operation being flown (inspection/surveillance flights and logistic flights) and the aircraft type (including a DJI F450 and a faster version called DJI F450 FAST).

The results of this project include a drone conflict trajectory example generated with DEG and an investigation of the performance and effectiveness of the DEG tool using a tailored existing DAA algorithm (DAIDALUS).

© CBLTIC Campus del Baix Llobregat - UPC