Double Degree PhD Scholarship : Monitoring of grazing animals using sensors and data

Organisation/Company Université de Liège Department Gembloux Agro-Bio Tech Research Field Agricultural sciences Researcher Profile First Stage Researcher (R1) Positions PhD Positions Final date to receive applications 15 Apr 2026 - 00:00 (Europe/Brussels) Country Belgium Type of Contract Temporary Job Status Full-time Is the job funded through the EU Research Framework Programme? Horizon Europe (other) Is the Job related to staff position within a Research Infrastructure?

No

Context
:
Grasslands cover a significant share of the world’s ice‑free land mass and are at the heart of the most criticized as well as most sensitive livestock farming systems. Adequate management is of utmost importance to maintain pasture health and allow the grasslands to provide these ecosystem services in the best possible way contrariwise, poor management leads to depletion of the forage resource with a whole cascade of negative effects for both the grazier and the environment.

Grazing is a process that has declinations at multiple scales in space and time ranging from the whole paddock over a grazing season to the smallest unit of the grazing process, i.e. the grass‑severing bite, that covers a couple of cm². Herbivores continuously sense the ever‑changing grazing environment in order to adapt their decisions. Short‑term decisions made at the level of each individual bite have consequences on the efficiency of the grazing process, the performance of the herbivores and the health of the grasslands.

As theorised by Charnov & Orians, herbivores are optimal foragers able to consume forage at higher rates than what the average sward structure would allow them to. Hence, starting from a favourable sward structure, the efficiency of the grazing process usually decreases with grazing down level: the lower animals get in the vegetation, the lower the harvest per bite. As a consequence, herbivores will increase the amount and/or the frequency of bites, will change the duration of their meals and will gradually have to cover higher areas during their meal sessions to look for these optimal structures until the sward is so depleted that they don’t waste time looking for better sward structures that they consider are no more present on the paddock.

Therefore, a better continuous monitoring of the perception of the animal behaviour on field is an open door to develop tools to spot animal or grassland health problems or to analyse reactions to specific structural elements in grasslands to innovate in grazing management. Over the past decades, many studies have documented the potential of sensing technology to monitor the grazing behaviour of domestic herbivores which served as the first bases during the implementation of this project.

Accelerometers and Inertial Measurement Units (IMU) combine practicality and good sensing performances for the inference of a wide range of behaviours, activities and postures, possibly down to the level of the bite. Other sensors, such as microphones and pressure sensors have been explored but present some limitations in the range of behaviour they can supply and situations where they can be used.

Global Navigation Satellite Systems (GNSS) combined to real‑time kinematics (RTK) technology give complementary information to accelerometers and have been frequently used to monitor free‑ranging animals. The location of the device on the animal also varies. While most works locate the sensors on halters, close to the jaws or on the neck, farmers are used to put collars on their stock and not halters.

But PhD Position H – Monitoring of grazing animals using sensors and data science most importantly, the weaknesses that have been evoked in the conclusion of most work is the unverified or lack of adaptability of models developed in specific case studies to new environment.

Objectives
: The PhD research will contribute
1) to the analysis methods and metrics for understanding the complex interactions between forage resource and dynamics;
2) to develop Machine Learning methods for analysing sensor data on animal movement and behaviour, adapting methods to different animals and environments;
3)…