A Novel Framework for Robot Training and Tutoring Using Natural Language

Applications are invited for a self-funded, 3-years full-time or 6-years part-time PhD project, to commence in October 2020 or February 2021.

The PhD will be based in the School of Computing and will be supervised by Dr Jiacheng Tan and Dr Zhaojie Ju.

Given the advances in robotics and AI, most commercially available robots have been meticulously pre-programmed and trained for the predefined tasks. Such robots can out-perform human in environments where the tasks and work-spaces can be fully controlled and modelled as in the case of industrial applications. In less controlled and hazardous environments, expert operators have to be involved in the control loop to compensate for the inadequate abilities of robots in perceiving, understanding and assessing the surrounding environment and in planning the operations tasks.

Without a breakthrough in AI, this reliance on human operators is unlike to change in the near future, which severely limited the use autonomous robots in hazardous and domestic environments such as underwater rescue, firefighting and healthcare even if the hardware platforms are adequate.

In this project, we aim at a novel approach to enable robots to learn and understand their surround environments and to be tutored to perform diversified tasks by non-expert users using natural language.

By relaxing the stringent requirements on robot perception system and by relieving the demands for the specialised knowledge and skills of robot training, it is anticipated that the work would make it possible to deploy dexterous robots with higher accessibility and at lower costs for domestic and hazardous applications where both task operations and environments are varying and cannot be fully pre-defined.

The research will be based upon and extend our work on robot task and work-space analysis through grounding the spatial relationships and quantifiers in natural language. The work will involve natural language analysis, machine learning, knowledge modelling, intelligent robot control and task scheduling. Candidates with an interest in AI and robotics and a good undergraduate or master’s degree in computer science, engineering, or other relevant subject areas are invited to apply.

Fees and funding

Visit the research subject area page for fees and funding information for this project.

Funding availability: Self-funded PhD students only. 

PhD full-time and part-time courses are eligible for the UK Government Doctoral Loan (UK and EU students only).

How to apply

We’d encourage you to contact Dr Jiacheng Tan (Jiacheng.tan@port.ac.uk) to discuss your interest before you apply, quoting the project code CCTS4520219.