Foundation-model-guided world models and predictive control for autonomous remote handling in extreme environments

About the Project

University of Sheffield (Top 100 QS, Russell Group)

Eligibility: UK (home) and exceptional international candidates

Deadline: 31 May 2026

Start Date: October 2026

Project Overview

We are seeking an exceptional and highly motivated PhD/EngD candidate to join a cutting-edge research project at the intersection of robotics, artificial intelligence, and control.

The goal of this project is to develop intelligent robotic systems capable of autonomous operation in extreme and hazardous environments, such as fusion energy systems and nuclear facilities. These environments present fundamental challenges: limited sensing, high uncertainty, and complex physical interactions.

This project addresses one of the central challenges in modern robotics: enabling robots to reason, predict, and act safely in contact-rich, partially observable environments.

Research Focus

The project will explore how foundation models (e.g. large-scale vision/language/action models) can guide:

• World models for learning predictive representations of system dynamics
• Model Predictive Control (MPC) for robust decision-making under uncertainty
• Robotic manipulation systems for contact-rich tasks

Key research challenges include:

• Learning reliable predictive models from sparse and noisy sensory data
• Incorporating semantic priors into planning and control
• Achieving safe and robust manipulation in constrained and hazardous environments

The outcomes of this research will contribute to the next generation of autonomous robotic systems for extreme industrial applications.

Research Environment

You will join a world-class, interdisciplinary ecosystem including:

• Intelligent Manipulation Lab (intmanlab.com)
• Sheffield Robotics
• Centre for Machine Intelligence (CMI), University of Sheffield

The project is embedded within the Fusion Engineering Centre for Doctoral Training (CDT), offering:

• A 4-year fully funded programme
• A cohort-based doctoral training environment
• A 3-month intensive training programme in fusion engineering
• Collaboration with leading academic and industrial partners

Who We’re Looking For

We are looking for candidates who are:

• Passionate about robotics, AI, and control systems
• Equipped with a strong background in robotics, machine learning, or control
• Strong programming skills in Python and experience with modern ML frameworks (e.g. PyTorch)
• Familiar with modern AI architectures (e.g. Transformers)
• Eager to tackle challenging, open-ended research problems

Highly desirable:

• A strong research track record, with evidence of excellence such as publications in leading venues (e.g. NeurIPS, ICML, ICLR, AAAI, ICRA, IROS, CoRL, CVPR) or equivalent research output

What You’ll Gain

• A 4-year EngD studentship starting October 2026
• Opportunity to publish in leading venues (e.g. RSS, CoRL, IEEE T-RO, ICML, NeurIPS)
• Access to cutting-edge robotic platforms and real-world applications
• Collaboration with leading researchers and industry partners
• Training within a vibrant, interdisciplinary doctoral cohort - In carrying out the work, you will also be part of the Fusion Engineering CDT, and, along with your cohort of other doctoral students from universities across the UK, will receive training on cutting edge topics in fusion energy from academic and industry experts. Students will receive a 3-month training programme in fusion engineering at the start of the course, delivered across the CDT partner universities. For further information about the CDT programme, please visit the CDT website at https://www.fusion-engineering-cdt.ac.uk/training-fusioneers/programme/ or send an email to hello@fusion-engineering-cdt.ac.uk.

Funding

• Full studentship (4 years)
• UKRI stipend (~£21,805/year for 2026/27)
• £25k Research Training Support Grant (RTSG)

How to Apply

Apply via:

https://www.sheffield.ac.uk/postgraduate/phd/apply/applying

(Select UKAEA Fusion Engineering CDT)

For informal enquiries: email the supervisor at a.ghalamzan@sheffield.ac.uk

Funding Notes

Students will receive a 4-year studentship including home tuition fees, UKRI stipend (indicated as £21,805 in 26-27) and a £25k RTSG budget for the project. All costs associated with attending CDT training will be met by the RTSG budget.

Please note that this advert will be withdrawn when a suitable candidate is identified, we recommend that you apply early.