Background: Animal-borne data loggers today often house several sensors recording simultaneously at high frequency. This offers opportunities to gain fine-scale insights into behaviour from individual-sensor as well as integrated multi-sensor data. In the context of behaviour recognition, even though accelerometers have been used extensively, magnetometers have recently been shown to detect specific behaviours that accelerometers miss. The prevalent constraint of limited training data necessitates the importance of identifying behaviours with high robustness to data from new individuals, and may require fusing data from both these sensors. However, no study yet has developed an end-to-end approach to recognise common animal behaviours such as foraging, locomotion, and resting from magnetometer data in a common classification framework capable of accommodating and comparing data from both sensors.
Methods: We address this by first leveraging magnetometers’ similarity to accelerometers to develop biomechanical descriptors of movement: we use the static component given by sensor tilt with respect to Earth’s local magnetic field to estimate posture, and the dynamic component given by change in sensor tilt with time to characterise movement intensity and periodicity. We use these descriptors within an existing hybrid scheme that combines biomechanics and machine learning to recognise behaviour. We showcase the utility of our method on triaxial magnetometer data collected on ten wild Kalahari meerkats (Suricata suricatta), with annotated video recordings of each individual serving as groundtruth. Finally, we compare our results with accelerometer-based behaviour recognition.
Results: The overall recognition accuracy of >94% obtained with magnetometer data was found to be comparable to that achieved using accelerometer data. Interestingly, higher robustness to inter-individual variability in dynamic behaviour was achieved with the magnetometer, while the accelerometer was better at estimating posture.
Conclusions: Magnetometers were found to accurately identify common behaviours, and were particularly robust to dynamic behaviour recognition. The use of biomechanical considerations to summarise magnetometer data makes the hybrid scheme capable of accommodating data from either or both sensors within the same framework according to each sensor’s strengths. This provides future studies with a method to assess the added benefit of using magnetometers for behaviour recognition.
Labelled triaxial magnetometer data
Paper: "Behavioural compass: animal behaviour recognition using magnetometers"
Authors: Pritish Chakravarty, Maiki Maalberg, Gabriele Cozzi, Arpat Ozgul, Kamiar Aminian
-----------------------------------------------------------
A guide to accessing and using the labelled, calibrated, windowed (sliding window of length two seconds with 50% overlap between successive windows) magnetometer data supplied:
A .mat file is supplied, which can be opened with MATLAB. There is one variable within named "sessionWiseMagData_fourBehaviours". It is a cell with eleven entries, corresponding to the eleven recording sessions done for this study.
Let "rs" denote the recording session number (rs = 1, 2, 3...11). sessionWiseMagData_fourBehaviours{rs} is a cell with four entries. Let "b" denote the behaviour number (b = 1,2,3,4 where 1 corresponds to vigilance, 2 to resting, 3 to foraging, and 4 to running). Then sessionWiseMagData_fourBehaviours{rs}{b} contains all the two-second bouts of calibrated triaxial magnetometer data recorded for behaviour "b" during recording session "rs".
Since magnetometer data was resampled to 100 Hz, each entry sessionWiseMagData_fourBehaviours{rs}{b}{i}, which denotes the i-th two-second bout of video-labelled behaviour "b" recorded during recording sessions "rs", is a matrix of size 200x3, where 200 denotes 2s*100 entries/axis/second, and 3 denotes the number of axes of the magnetometer).
Note that feature computation was done on each of these 82550 two-second bouts of labelled triaxial magnetometer data for the four behaviours of interest.
Labelled3DmagData_Chakravarty_et_al_2019_MovEco.rar
Feature matrices
Paper: "Behavioural compass: animal behaviour recognition using magnetometers"
Authors: Pritish Chakravarty, Maiki Maalberg, Gabriele Cozzi, Arpat Ozgul, Kamiar Aminian
-----------------------------------------------------------
A note on the feature matrices supplied:
- There are two Excel files. "allFeatures.xlsx" contains all nine features (Table 1 in main manuscript) considered in this study. "top3features.xlsx" contains the top three features (feature selection details in Appendix 3, Additional File 1), and is supplied for convenience.
- The first row of each Excel file contains the header, arranged as: {feature names, behaviour label, recording session number). Each subsequent row contains feature values computed over a two-second window of triaxial magnetometer data. Both Excel files have the same number of rows, and the same row number refers to the same two-second window of triaxial magnetometer data across both files.
- The behaviour labels are as follows. 1: vigilance; 2: resting; 3: foraging; 4: running.
- The last column contains the recording session number corresponding to each two-second window of magnetometer data. Data were collected over 11 recording sessions (Table 2 in main manuscript), so the last column contains integer values from 1 to 11.
FeatureMatrices_Chakravarty_et_al_2019_MovEco.rar