Ελληνικά   English  
Home  |   News  |   Events  |   Proclamations  |   Job Opportunities  |   My Certh  |   Contact Us
Print this page
NewsNews

Two laboratories of CERTH accomplish the 1st and 3rd place, respectively, in a NATO contest
[4 June 2021]


Friday, 4 June 2021

 

The laboratories of Visual Analytics, Virtual and Augmented Reality (VARLab) and Visual Computing (VCL) of Information Technologies Institute of CERTH achieved the 1st and 3rd place respectively, in the data challenge, within the International Conference on Military Communication and Information Systems (ICMCIS 2021).

The competition was organized by the NATO Communications and Information Agency. Contestants were asked to track, categorize and identify unmanned aerial vehicles (UAVs) as they flew over a specified area, based on data from available sensors.

In particular, the drones had to be tracked to show their position and speed, categorized to unveil they are actually drones, and identified to indicate the type of unmanned aerial system in question.

The research team of Visual Analytics, Virtual and Augmented Reality Laboratory (VARLab) consisting of Dimitrios Tsiktsiris, Theoktisti Marinopoulou, Eftychia Makri, Nikolaos Sakellariou, Eleni Diamantidou, Anastasios Vafiadis, Ilias Kalamaras, Antonios Lalas, Konstantinos Votis, and Dimitrios Tzovaras took the 1st place, by achieving to develop a new machine learning model for the simultaneous solution of the triple problem of tracking-categorization-identification of unmanned aerial systems with increased accuracy compared to the individual approaches, as well as the best score in terms of tracking, but also the overall solution, providing an easy expansion capability.

The research team of the laboratory of Visual Computing Lab (VCL) consisting of Paschalina Medentzidou, Anastasios Dimou, Dimitrios Zarpalas and Petros Daras took the 3rd place, managing to solve the drone tracking problem using a custom data association cost metric and calculating an optimal assignment at each second using the Hungarian algorithm. With an efficient management of tracklet lifespan the team managed to track UAS in real time in a range of flight scenarios with ascending difficulty.

You can find here the announcement of   the NATO Communications and Information Agency: https://www.ncia.nato.int/about-us/newsroom/agency-announces-winners-of-drone-data-challenge.html