This PhD project at Cardiff University focuses on developing MorphoFusionAssist, a modular, reconfigurable rolling robot designed to assist with inspection and maintenance in modern fusion plants. The robot is composed of sealed spherical units capable of rolling through narrow, winding ducts and reconfiguring on-site into compact manipulators or load-sharing clusters. This approach offers significant advantages over traditional fixed-base arms or continuum robots by minimizing collision risk and setup time, and maintaining functionality even if some modules fail. Each module is engineered for decontamination and rated to withstand heat, dust, and residual magnetic fields, ensuring reliable operation in hazardous environments. The project builds on existing prototypes and research in rolling robots, motion planning, and compliant control. The successful candidate will develop robust, internally actuated modules with magnetic shielding and reliable docking mechanisms, plan safe robot routes using advanced curve-aware and multi-contact models, and implement minimal sensing strategies for contact estimation. Control strategies will include underactuated, energy-shaping, and compliant controllers, with operator-in-the-loop shared autonomy for robust human guidance. The research will also address the fusion of mobility and manipulation, enabling the robot to switch between rolling and local manipulation, perform grasp-less tasks using rolling contact, and coordinate multiple modules for precise load sharing and fixture placement. Testing will be conducted in Cardiff’s gm²R lab, RAIM group, and potentially on UKAEA RACE mock-ups. The anticipated outcome is a versatile, reconfigurable platform capable of replacing multiple specialist systems, reducing costs and expanding remote maintenance capabilities in fusion and other hazardous, confined settings. The project will contribute new theory and algorithms for reconfigurable rolling manipulation, with broad applicability. Full funding is available for Home and International students, and applicants should have a strong background in engineering or a related field, with proficiency in English. Applications are open until January 9, 2026, with the possibility of early closure if a suitable candidate is found.