AI-Driven Digital Twins for Robots

Robots are playing key roles in many social and industrial applications such as in hospitals, automation in automotive domains, etc. Digital twins, i.e, live and digital representations of a physical twin such as robots, promise to significantly improve their reliability, helping them in making optimized decisions in real-time and predict future activities such as predictive maintenance. While the concept of digital twins is currently attracting a lot of attention, there is little research done on the principles to design and implement them.
Master

This research topic focuses on building digital twins of rovers from various objectives either through machine learning techniques, model-based systems engineering, or a combination of both. In the context of digital twins for rovers, we aim to explore the following scenarios:

  • Digital-twin enabled techniques to determine whether the behavior of a rover deviates significantly from the expected one.
  • Building approaches that can be used for automated refactoring of digital twin models, e.g., for supporting predictive refactoring, using genetic algorithms or any predictive models such as ML.
  • Considering how to learn and to transfer such a learning, possibly with uncertainty, to evolve models within the digital twins.
  • Utilizing path planners inside the digital twins to determine the right action plan in real-time.
  • Building generic digital twins for various types of rovers that can be easily configured for various.

Goal

The candidate will address the following activities:

  • Establish the state of the art on digital twins, specifically for rovers.
  • Define the relevant properties related to the implementation of such digital twins, e.g., reusability, modularity, evolution, uncertainty and transfer learning.
  • Define a reference software architecture for designing and implementing such digital twins, maximizing the identified properties, and offering the aforementioned scenarios.
  • Explore a generative approach to automate the implementation of digital twins according to the proposed reference architecture.

Learning outcome

  • Learning innovative digital twin technology
  • Apply various AI techniques
  • Work with real rovers

Qualifications

  • Programming skills
  • Familiarity with machine learning frameworks is a plus

Supervisors

  • Shaukat Ali
  • The research will be conducted in the context of INRIA-Simula Associated Team RESIST. More details about the team can be found at gemoc.org/resist/

The research will be jointly supervised by Benoit Combemale (people.irisa.fr/Benoit.Combemale/) and Djamel Khelladi (people.irisa.fr/Djamel-Eddine.Khelladi/) from Inria, and Shaukat Ali (www.simula.no/people/shaukat). As part of the research, the following physical rovers can be used: 1) Leo Rovers (www.leorover.tech/) at Simula; 2) the Aion Robotics R6 (www.aionrobotics.com/r6) at Inria.

Collaboration partners

INRIA, France

References

  • Qi, Q., Tao, F., Hu, T., Anwer, N., Liu, A., Wei, Y., and others & Nee, A. Y. C. (2021). Enabling technologies and tools for digital twin. Journal of Manufacturing Systems, 58, 3-21
  • Kritzinger, W., Karner, M., Traar, G., Henjes, J., & Sihn, W. (2018). Digital Twin in manufacturing: A categorical literature review and classification. IFAC-PapersOnLine, 51(11), 1016-1022
  • Romina Eramo, Francis Bordeleau, Benoît Combemale, Mark van den Brand, Manuel Wimmer, Andreas Wortmann: Conceptualizing Digital Twins. IEEE Softw. 39(2): 39-46 (2022)
  • Uncertainty-Aware Transfer Learning to Evolve Digital Twins for Industrial Elevators. Xu Qinghua, Shaukat Ali, Tao Yue, Maite Arratibel

Contact person