About the Project
Lameness is an important health-, welfare-, and production-limiting problem for the livestock industry.
The current “gold-standard” method to assess lameness is by visual locomotion scoring, which essentially involves trained veterinary clinicians carefully observing and scoring tens of animals.
Recent technical advances have enabled the development of objective methods for detecting lameness, although they require each animal to be equipped with costly devices or restrained.
In contrast, radio-frequency sensor methods such as the micro-Doppler radar system (MDRS) we have prototyped offer cost and labour advantages.
We aim to develop an effective MDRS for lameness detection in sheep and cattle. The project will involve the modelling, examination and classification of animals’ locomotion, developing innovative machine-learning methods of analysis that can process the radar data effectively.
The project is funded by the Vet Fund (https://www.gla.ac.uk/connect/supportus/vetfund/), providing a stipend of £20,000 plus sufficient operational costs to complete the project successfully. It is fully funded for Home/EU applicants.
The project is inherently cross-disciplinary, and will be heavily weighted to machine learning with some elements of signal processing. It would suit candidates with backgrounds ranging from veterinary or animal science, through to sensor and electronic engineering, physics, computer science, artificial intelligence, or machine learning.
The main aims of this project are:
• Characterisation: To characterise and validate micro-Doppler radar signatures of sheep and cattle with varying degrees of gait impairment, as an automatic diagnostic tool for “precision veterinary medicine”
• Innovation: To develop innovative machine learning algorithms that can infer the state of locomotion of the animals under test from their radar signatures and support contactless automatic classification
• Exploration: To establish the effect of the many operational parameters on the overall performance of the system, such as location and height of the radar, anterior or posterior view of the animal, minimal amount of data required to reach reliable decisions, and radar parameters (in particular carrier frequency, resolution, and polarisation)
• Validation: To quantify the performance of the micro-Doppler radar sensing method for which we had previously established technical proof of concept for lameness detection in dairy cows and sheep in commercial farm settings.
Work that has led up to this project:
This co-supervisory team provides extensive cross-disciplinary expertise in diverse areas of scientific research, including animal husbandry, farm animal medicine and production, data analysis, radar sensing technologies and machine learning methods. The team has successfully worked together in the past and shown the technical potential of micro-Doppler radar technology for lameness detection in a small number of farm animal and horses.
Below are two papers that are published or in revision for publication, and figures from Shrestha et al. (2018) presenting examples of graphical outputs of typical signals, specifically Doppler (velocity) vs time patterns for animals walking.
1, Shrestha, A., Loukas, C., Le Kernec, J., Fioranelli, F., Busin, V., Jonsson, N., King, G., Tomlinson, M., Viora, L., Voute, L., (2018) ‘Animal lameness detection with radar sensing’, IEEE Geoscience and Remote Sensing Letters, doi: 10.1109/LGRS.2018.2832650
2. Busin, V., Viora, L., King, G., Tomlinson, M., Le Kernec, J., Jonsson, N., Fioranelli, F., ‘Evaluation of lameness detection using radar sensing in ruminants’, Veterinary Record (in revision)
Continue with Facebook
