The field of movement ecology has taken a huge leap forward with the use of biologging techniques, especially tri-axial accelerometers (ACC) for studying animal behaviour. These devices, measuring acceleration along three axes, have broadened the horizons of ecological research. However, hurdles like the weight of tags on birds and limited battery life have pushed us to find creative solutions.
This research focused on creating a real-time processing pipeline for ACC data using on-board artificial neural networks. Using European pied flycatchers as first test subjects, the researchers tagged the birds, gathered raw ACC data, and developed a deep learning model for edge computing. This model, embedded in new loggers, identifies and records bird behaviours straight from the raw data. The system successfully passed the tests, offering a promising method for continuous, automated behavioural monitoring in birds. This breakthrough boosts data collection efficiency and paves the way for exciting new opportunities in wildlife research and conservation.
Scientific output
Yu, H., F.T. Muijres, J.S. te Lindert, A. Hedenström. And P. Henningsson. 2023. Accelerometer sampling requirements for animal behaviour classification and estimation of energy expenditure. Animal Biotelemetry 11, 28.
Yu, H., G.J. Amador, A. Cribellier A., M. Klaassen, H.J. de Knegt, M. Naguib, R. Nijland, L. Nowas, H.H.T. Prins, L. Snijders, C. Tyson, F.T. Muijres. 2023. Edge computing in wildlife behaviour and ecology. Trends in Ecol Evo 39, 2.
Michiel P Boom, Hui Yu, Roeland A Bom, Arne Hegemann, Åke Lindström, Bart A Nolet, Thomas K Lameris. 2024. Migrating shorebird killed by raptor at 3000 m above ground as revealed by high‐resolution tracking. Hui Yu, Shujie Liang, Florian T Muijres, Jan Severin Te Lindert, Henrik J de Knegt, Anders Hedenström, Koosje P Lamers, Per Henningsson. 2024. Flight activity and effort of breeding pied flycatchers in the wild, revealed with accelerometers and machine learning.
Presentations
Yu, H. 2022. 2022. Continuous on-board behaviour classification using accelerometry, a case study and ecological insights. Association for the Study of Animal Behaviour (ASAB) 2022 winter conference, Edinburgh, Scotland, United Kingdom, 6-7 December, 2022.
Yu, H. 2023. Accelerometer sampling requirements for animal behaviour classification and estimation of energy expenditure. International Wader Study Group 2023 conference, Sylt, Germany, 29 September – 3 October 2023.
