The deadline for submitting the results is postponed to Monday 14th June 2021.
A new Challenge on self-awareness in heterogeneous multi-robot systems has been organized within the first International Conference on Autonomous Systems (ICAS 2021). The research field of this competition is the unsupervised anomaly detection through self-aware autonomous systems, which is an active topic involving IEEE Signal Processing Society (https://signalprocessingsociety.org/) , also through the Autonomous Systems Initiative (https://ieeeasi.signalprocessingsociety.org/) , and Intelligent Transportation Systems Society (https://www.ieee-itss.org/) .
The competition will allow all participating teams to create intelligent and autonomous unsupervised algorithms that will be capable of determining the normal or non-normal behavior of a ground vehicle that interplays with the environment. So, the challenge is focused to discover anomalies automatically in a common dataset that will be delivered to all teams who participate in the challenge.
The goal of the competition is to motivate the academic and industrial researchers to create and innovate in the field of the autonomous systems by developing unsupervised algorithms, where future intelligent systems will be capable to understand the environment and detects anomalies automatically. The autonomous understanding by the observation of the environment is an unresolved field in present autonomous systems. So, the aim of this challenge is to detect own and non-own behaviors and anomalies in the navigation of the ground and aerial systems based on sensor data in real-time. This task is particularly challenging because of several factors: first of all, the high variability of situations that the autonomous system has to cope with, make practically impossible the application of classical data fusion, signal processing and supervised techniques. Moreover, available data are acquired by real autonomous systems moving in complex environments, thus they suffer from noise and possible
synchronization problems that have to be considered for solving the required tasks. Finally, proposed solutions must be able to work in real time in embedded architecture with low computational capabilities and very limited power resources.
The competition consists of the following stages: an open competition that any eligible team can participate in, and a closed final competition. Teams participating in the open competition must submit their results no later than Friday, 14 May 2021. Then, the three teams with the highest performance in the open competition will be selected as finalists and will be invited to participate in the final competition. Finalists will be announced on Friday, 28 May 2021. The three teams invited to participate in the final competition will be judged at ICAS 2021 (https://2021.ieee-icas.org/) , which will be held on 11-13 August 2021 as a Virtual Conference.
Prizes: At the final competition at ICAS 2021, the three finalist teams will be awarded with prizes thanks to the sponsoring offered by IEEE Signal Processing Society through the Autonomous Systems Initiative. The champion team will receive a grand prize of $2,500. The first and the second runner-up will receive a prize of $1,500 and $1,000, respectively. Moreover, the best teams will be invited to submit a challenge paper for publication at ICAS.
So, the organizers are proud to announce this challenge: an exciting unsupervised anomaly detection challenge, where autonomous systems interplay with the environment to discover anomalies automatically.
On the behalf of the organizers,
David Martín and Lucio Marcenaro
Access The Webinars
A series of webinars is organized to give more details on the challenge. For accessing the webinars the following link can be used; this will be a virtual room for the preparation of the ICAS challenge:
The next webinar will be on Friday 7th May at 4:30 pm (CET time) to give more details about the new dataset available for the competition.
All the webinars will be recorded, to be seen offline by all the potentially interested participants.
All the teams participating to the challenge, please register by using the following form