Project
Analysis of tracking data and sensor platforms
Animal movement behaviour has been an interesting topic for many years. Researchers have focused on many aspects, including animal migration and dispersal, and the biomechanics, physiology and energetics of animal locomotion. Important model species for this research have been various species of birds because aerial locomotion is among the most energy-consuming behaviours in nature, and birds are the longest-distance migrants.
In this project, researchers develop novel high-tech sensors and onboard processing algorithms that allow the use of these next-generation sensors for studying the movement dynamics of small passerines in unprecedented detail. Specifically, they are developing ultra-lightweight accelerometer loggers with onboard data processing techniques for identifying detailed behaviours like flying, resting, and foraging in passerines in the wild.
Using detailed analysis of the flight dynamics of these freely flying animals, the researchers also aim to test whether high sampling rate accelerometer data (~100 Hz) can be used to identify different flight behaviours such as commuting, hunting, gleaning and eating, and estimate bird flight performance metrics such as aerodynamic force production and the energetic cost of flight.
During the field season in 2022, the researchers performed control and calibration experiments on pied flycatchers in large-cage outdoor aviaries. Here, they tracked flying and hunting flycatchers using stereoscopic videography and light weight accelerometer loggers. Using these combined videography-accelerometer data, the researchers developed a novel machine learning model for classifying a range of behaviours of flying pied flycatchers using lightweight accelerometer data.
After analysing the data collected in 2022, the researchers organised the results and findings into a scientific paper, which was published in the journal of Animal Biotelemetry.
During the field season in 2023, the researchers used new loggers (< 0.6g) developed by industrial partner Druid Technology on wild pied flycatchers to track their behaviour patterns using an on-board accelerometer and light sensor. Other stationary sensors, such as RFIDs and cameras, were also used in the study to provide complementary data for research.
The researchers submitted a research paper (currently under review) to Journal of Experimental Biology in early 2024. Here is a short summary of the paper: The paper investigates the aerodynamic and behavioral aspects of flight in pied flycatchers (Ficedula hypoleuca) during their chick rearing period. The study analyzed flight performance, activity patterns, and sexual dimorphism in wing loading relative to flight metrics using accelerometers and supervised machine learning. Flight activity was recorded using miniaturized accelerometers on 26 birds, with flight behavior annotated via machine learning (XGBoost) trained on aviary-based data. Flight effort and performance were measured by dynamic body acceleration (VeDBA). Findings reveal no significant differences in flight activity and performance between sexes despite males having lower wing loading. The research indicates equivalent parental investment with reduced effort from males, possibly due to better flight efficiency related to morphology. The preprint of the paper can be found in the right-hand column.
The researchers will work on two more scientific papers in 2024.