EngD position: AI for Underground Infrastructure Detection and Characterisation

The University of Twente is offering a two-year EngD position focused on developing machine learning models to improve the detection of underground infrastructure in Ground Penetrating Radar (GPR) radargrams. As part of the ZoARG|ReDUCE programme, the successful candidate will work in a multidisciplinary environment to enhance excavation safety and utility mapping accuracy. The role involves managing large, annotated datasets and translating research findings into practical, industry-ready solutions. Applicants will benefit from a structured programme that balances academic education with hands-on design projects in collaboration with industrial partners.

• Evaluate existing GPR interpretation methods and relevant machine learning software tools.
• Prepare, preprocess, and manage large-scale GPR datasets collected at the Utility Mapping Site.
• Design, train, and validate machine learning models for automated utility characterization.
• Compare developed AI models against current literature and commercial software solutions.
• Collaborate with infrastructure owners, contractors, and technology providers within the ZoARG programme.
• Translate research findings into practical recommendations for measurement practice and technology evaluation.