After several months and two challenging phases, the 2024 edition of ICUAS UAV Competition has concluded with the special session with team presentations in Chania, Greece, held on June 7, 2024.
From the group of six finalist teams (AIRo Lab from The Hong Kong Polytechnic University), AVADER AGH from AGH University of Krakow, AVANT-UFMG from Universidade Federal de Minas Gerais, ITUAV from Istanbul Technical University, UAS-DTU from Delhi Technological University (Formerly DCE), UNIST ASL from Ulsan National Institute of Science and Technology), the team from Ulsan National Institute of Science and Technology, Republic of Korea - UNIST ASL (https://asl.unist.ac.kr/) emerged on top as the winners of the Competition, preparing the best yield estimation for UAVs in both indoor and outdoor agriculture scenarios.
We congratulate the team supervisor, professor Hyondong Oh, team leader Jaemin Seo and team members: Geunsik Bae, Minjae Jung, Taewook Park, Hongro Jang, Myeonggeun Gu, Woojae Shin, Jinwoo Lee, and Hankyoul Ko.
Congratulations to team AIRo Lab from the Hong Kong Polytechnic University for placing second, team UAS-DTU from Delhi Technological University on the third place, and all other finalists and teams that submitted simulation phase solutions. Thank you all for your effort and time, we hope it was fun and educational.
Some stats and fun facts from the 2024 edition of the competition:
36 teams registered from 19 countries
24 teams submitted simulation phase solution
Simulation phase solutions have been evaluated 103 times in total
With the count of manual interventions in the automated evaluation process being less than 103, we can conclude it worked at least once
With the average run in simulation lasting 2 minutes and 50 seconds, the simulated UAVs have flown for almost 5 hours
In the first two simulation runs, the solutions overestimated the fruit count by ~10%, with subsequent 4 runs averaging underestimated count by ~9%
6 teams qualified for the finals
In the final stage on real data provided by Avyonx, solutions by top teams underestimated the fruit count by ~8.5%, matching the performance from the simulation phase
Only one solution did not pass “hallucination” check and reported fruit detections in data from an empty orchard
With the competition concluded, we have to thank the hashtag#Aerostream (https://aerostream.fer.hr/) project and especially prof. Tamara Petrović, ICUAS’24 organizing committee, ICUAS Association, Inc and professor Kimon P. Valavanis, professors Nikos Vitzilaios and Lefteris Doitsidis; FER and Avyonx for their support, availability and help in making the competition happen.
Thank you to everyone that stopped by the competition session in Chania, and hope to see you all next year in Marseille.