PhD Position in Safe Energy-aware Robotic Teleoperation

at  Syddansk Universitet

SDU Robotics , part of the Maersk Mc-Kinney Moller Institute , University of Southern Denmark (SDU), invites applications for one or more 3-year PhD position(s), where the candidates will work on the project SENSIBLE (Symbiotic tEleoperatioN for Safe and cost-efficient wInd turBine bLade maintEnance). The starting date for the positions is as soon as possible in 2024 and will be agreed upon with the successful candidates.

PROJECT DESCRIPTION

The aim of the project SENSIBLE is to teleoperate a robot by a human operator to perform challenging maintenance tasks for wind turbine blades, e.g., leading-edge repair tasks.
As part of SENSIBLE, t he overall objective of the PhD project is to use energy-aware approaches to build a stable, safe, reliable teleoperation system which considers time-varying human and environment dynamics and variable large time delays (e.g., around 100 ms).
To this end, the research activities include 1) Employing methods from the energy-aware robotics framework, including passivity theory and port-based modeling, to develop energy-aware methods that guarantee robust stability of the teleoperation system under varying environments and in the presence of large time delays.; 2) Building on recent developments in energy-aware control approaches, including control barrier functions, to create a safety layer capable of enforcing safety constraints. These include the energy content of the teleoperation system, applied forces, and energy flow between the system, human operator, and environment.; 3) Developing (reinforcement and/or imitation) learning based methods, producing policies that can assist the human operator in actively varying the remote robot’s impedance in real-time, while maintaining to satisfy the robust stability and safety constraints through the energy-based safety layer; 4) (Optional) Researching how to use large language models to dynamically allocate the authority between the operator and the teleoperation system in shared control/autonomy for improved stability and safety.

PROFILE AND REQUIREMENTS

The candidates will participate in SDU Robotics’ comprehensive research and teaching activities within the areas of robotics, mathematical modeling and physical human robot interaction. The position requires a master’s degree within robotics, electrical engineering or a similar/relevant area, and the candidates are expected to have a profound knowledge on the majority of the following topics:

• Robotics (e.g., modelling, kinematics, dynamics and control)
• Familiar with machine learning/deep learning frameworks and tools, e.g., TensorFlow, PyTorch, OpenAI GYM, CUDA.
• Programming skills (particularly in C++, Python and ROS)
• Knowledge of energy-based modelling and energy-aware approaches would be considered a plus.

Furthermore, the successful candidate is expected to:

• Have excellent knowledge of English (written and spoken).
• Be able to work well and communicate expert knowledge in an interdisciplinary team.
• Have high self-motivation to learn.
• Have the ability of working independently and have a strong sense of responsibility.

Please refer the Job description for details