Konferenzbeitrag
AI in the Wild: Challenges of Remote Deployments
Vorschaubild nicht verfügbar
Volltext URI
Dokumententyp
Text/Conference Paper
Dateien
Zusatzinformation
Datum
2023
Autor:innen
Zeitschriftentitel
ISSN der Zeitschrift
Bandtitel
Verlag
Gesellschaft für Informatik e.V.
Zusammenfassung
The effect of humanity on the earth becomes more and more apparent. Besides the publicly discussed climate change and overpopulation, also the number of conflicts with wildlife increases. The technological progress of the past years helped to understand these challenges better. Monitoring solutions, known to the public as the Internet of Things (IoT), increase the amount of collected data, whereas artificial intelligence (AI) supports analyzing and gathering a deeper understanding. Most projects in the area of wildlife try to achieve a more sustainable usage of natural resources and a better coexistence with our environment.
The mAInZaun project focuses on the conflict between wolves and livestock. It aims to introduce these new technologies into grazing management and foster non-lethally coexistence between livestock and predators. Artificial intelligence (AI) analyzes images and videos of the areas surrounding the pasture. The algorithms detect possible attackers or predators, such as wolves, stray dogs, bobcats, etc. In the second step, these animals are scared away using adaptive technologies. These can be sound, ultrasound, scent, light, etc.
These systems are usually operated in remote environments, raising challenges like hardware design, power requirements, and maintenance. This paper will discuss these challenges and how we address them in the mAInZaun project.