PhD Studentship: Integrating Machine Learning with One Health Frameworks to Quantify Disease Transmission Pathways between Wildlife and the Environment

The behaviour of animals determines their responses to environmental change and ultimately shapes whether populations persist or decline. Tools to record and analyse the behaviour of individual animals under natural conditions are therefore vital not only for fundamental research to understand why animals behave in the ways they do, but also for applied work addressing urgent challenges in animal conservation and welfare.

However, existing technologies have mostly been developed for controlled laboratory settings and are often unsuitable for field conditions. For instance, cameras used to monitor nesting birds typically require mains power or are too bulky to fit in nests without disturbing occupants. Thermal imaging is emerging as a vital tool for non-invasive monitoring of stress, but thermal cameras are typically large, expensive and power-hungry. Similarly, animal-borne microphones can provide insights into vocal communication and responses to anthropogenic disturbance, but have limited battery life and are too heavy for most species. These limitations highlight the need for integrated sensing systems that are miniaturised, low-power, robust and capable of autonomous operation in challenging field environments. At the heart of this PhD is an engineering challenge: how can sensing hardware and intelligent data processing be designed together to obtain biologically meaningful information from wild animals while operating under constraints on device mass, size, power and environmental robustness?

Advances in AI will be crucial, for example by enabling individual recognition from visual and vocal characteristics and automated quantification of behavioural and physiological responses. Rather than treating hardware and AI as separate components, they will be co-designed to acquire only the information needed to answer biologically meaningful questions, at appropriate spatial and temporal resolution, thereby reducing energy, storage and processing demands. Candidate systems will first be quantitatively benchmarked under controlled conditions before deployment and validation with established wild-animal study populations. The project brings together complementary expertise in behavioural field biology (Dr Alex Thornton), passive monitoring and ecological applications of AI (Dr Kim Hockings), and electronic engineering and sensing (Prof Monica Craciun). Thornton will provide expertise in behavioural ecology, field experimentation and biological validation; Hockings will contribute Engineering and Physical Sciences Research Council Doctoral Landscape Award expertise in passive acoustic and visual monitoring and AI for ecological applications; and Craciun will provide expertise in sensor and electronic-system design, including miniaturisation and low-power sensing.

The project is particularly suited to an interdisciplinary candidate interested in applying engineering, computer science and/or AI to real-world biological problems. The Cornish Jackdaw Project – Thornton’s long-term wild jackdaw study population – provides the ideal setting to test and validate methods before applying them to field sites in more remote locations, including Hockings’ study sites in West Africa. The student will gain experience and transferable skills spanning ecological research, engineering and data science, alongside opportunities for technology translation, stakeholder engagement and commercialisation. Given the urgent need for robust, affordable tools to evaluate behavioural responses to environmental change, the resulting technologies could have applications well beyond the focal study systems, including conservation, wildlife management, animal welfare and environmental monitoring.

Funded by: EPSRC Engineering and Physical Sciences Research Council