Thesis defence: "Sensor and the Beast: Generalised Methods to Recognise Animal Behaviour and Quantify Energy Expenditure Using Inertial Sensors, and Applications"

Miniature sensors like accelerometers and magnetometers can be affixed to wild animals to remotely understand how they lead their otherwise unobservable lives. But how do we reliably process sensor data, and how much can we infer?

In this thesis, I present 'biomechanically aware' algorithms to accurately recognise animal behaviour and estimate energy expenditure from sensor data. I then apply these algorithms to understand how meerkats - group-living carnivores found in the Kalahari Desert (the same animal as Timon in Lion King) - use different strategies to look for food based on their group 'rank'.

This thesis was supervised by Prof. Kamiar Aminian, head of the Laboratory of Movement Analysis and Measurement, and was done in collaboration with the Population Ecology Group at the University of Zurich. Data for this SNSF-funded project (grant no. CR32I3_159743) were collected at the Kalahari Meerkat Project in South Africa.