Robust Data-Fusion for Drone Surveillance Systems Using Cross-Domain AI University of Edinburgh in United Kingdom

This PhD opportunity at the University of Edinburgh's School of Engineering focuses on advancing robust data-fusion techniques for drone surveillance systems using cross-domain artificial intelligence. Drones are increasingly deployed for Earth observation, urban surveillance, goods delivery, and humanitarian missions, but their effectiveness depends on the reliability of on-board sensors and the integrated processing of multi-modal data. In challenging environments—such as fog, dust, or radio-frequency congestion—sensor modalities like cameras, LiDAR, and radar may become unreliable or unavailable, necessitating innovative solutions for data fusion and analysis.

The project aims to develop low-shot training methods for cross-domain AI models to: (1) improve the reliability of data analysis when some sensor modalities are absent or contaminated by noise, and (2) enhance the overall accuracy of existing models. Key research questions include bridging the 'accuracy gap' by combining data from different modalities to improve detection, tracking, and classification; addressing the 'reliability gap' by identifying and mitigating unreliable modalities using limited data from others; and evaluating the 'generalisation gap' to ensure robustness against distribution shifts, such as day–night transitions.

Applications of this research span security, defence, climate change monitoring, disaster surveillance, search and rescue, and urban planning. The project is highly relevant to industry and government agencies seeking reliable, adaptive sensor fusion frameworks for drone platforms operating in complex urban environments.

Eligibility requirements include a 1st Class undergraduate degree (or equivalent) and meeting the University’s English language standards. Candidates with backgrounds in computer science, electrical engineering, robotics, or related disciplines are encouraged to apply. Strong analytical and programming skills are advantageous.

Funding for this project is currently pending confirmation and, if ratified, will be available to home applicants only (UK+EU settled/pre-settled). Prospective candidates should contact Dr Mehrdad Yaghoobi ([email protected]) for further information on funding and eligibility. Early applications are advised, as the advert may close once a suitable candidate is found.

For more details and to apply, visit the project page: FindAPhD Project Link.