Skip to content

Open position for a full PhD research contract

3-year PhD contract sponsored by EUROCONTROL
Open position for a full PhD research contract

Research Contract

EUROCONTROL 18-220569-A

Presentation

The ‘Deep Learning for Demand Capacity Balancing Hotspot Detection’ is the research topic of the contract of EUROCONTROL with the ICARUS research group of the UPC. It will fund a student during 3-years, with interest in obtaining a PhD degree in the area of air traffic management and deep learning techniques.

The ICARUS research group is currently investigating deep learning techniques applied to drones navigation, payload processing and conflict classification. Also we have expertise in new operational concepts related to airspace modernization from SESAR (Free Route Airspace, Dynamic Sectorization, Continuous Descend Trajectories, ...) and drones integration into non-segregated airspace.

Air traffic Demand Capacity Balancing (DCB) is a strategical protection filter aiming to prepare to the human controller a traffic situation compatible with his/her mental resources. DCB shall accurately predict the congestion areas to be able then to implement the solutions needed to resolve the imbalance. Currently this is done in two steps: first, predicting the imbalanced sectors by comparing the controller workload with the sector capacity, and second, through expert’s judgments to determine which imbalances are actual hotspots and require measures to adjust the demand. This judgment takes into account several features such as: the indicators used for the workload calculation, the nature of the traffic in sector, the accuracy of the workload prediction, the time horizon or the possibility to resolve the imbalance with capacity measures.

 The challenge of this project is to show the potential of using neural networks to create an automated system able to learn from these human experts and their past decision to quickly determine the hotspot zones in the future. Convolutional (CNN) and Recurrent (RNN) will be tested during the 3-years project duration.

Offering

Three years of full-time funded internship, including the PhD/Master program enrollment, to reach the highest academic degree

Barcelona-Castelldefels working area, with possibility of funded stages at other prestigious research centers

Work on the latest research topics and technologies

Reach the necessary capabilities to manage research projects

Research Planning

The contract will have a duration of 3 years and will have the following planning:

Literature review of the 2 involved topics: deep learning and DCB

Set up of the tested environment: Neural Network libraries and Data preparation

Development and test of new and/or existing deep learning models to use with the data set

Validation and final tuning of the model

The student will meet with the topic area expert from EUROCONTROL at least twice a year to report advances and adjust requirements, and will deliver the following 4 contractual reports:

D1 - Comprehensive literature review (M9) D2 - Workshop paper (M18)

D3 - Conference paper (M27) D4 - Journal paper (M36)

Candidate’s profile

The candidate should be European and have an Engineering or Science degree and, optionally, a Master degree. The referable profile should demonstrate a strong level in computer science, sufficient proficiency with English language and high academic grades. Additionally, competences in at least one of the two topics of the project would be considered: air traffic management and/or deep learning.

During the contract the candidate will have to enroll in a PhD program under the supervision of Dr. Salamí, and, if previously necessary, to its leading Master program.

Availability

Expected starting date: September 2019

Applications

Send an e-mail to the contact person BEFORE 15-06-2019 with:

European citizen passport

Curriculum Vitae

Qualifications certificate (English) - If qualifications are in another language, add official translation and description of the qualification system of your country -

Contact

Esther Salamí

esalami@ac.upc.edu