Context
A postdoctoral position is opened within the Inria-KAIST partnership. It is part of the SHAARE associate team, initiated between the Auctus team at Inria and the IRiS lab at KAIST that focuses on haptics and shared teleoperation. Together, we aim at developing shared-control approaches that, either, better guide the human through adaptive haptic guidance, or adjust the robot behavior according to the human gestures.
The postdoc contract will have a duration of 12 to 24 months. The start date will be between November 1st, 2024, and not later than January 1st, 2025, depending on the candidate availability. The postdoctoral fellow will be recruited by the Inria center of the University of Bordeaux (Auctus team) in France, where he/she will start the research contract. The project duration will be half-time divided between France and South Korea, with exchange periods at KAIST (IRiS lab) to share the works and progresses.
Research activities
Humans and robots can jointly perform a task in haptic teleoperation. The human operator remotely controls the robot while receiving feedback on the task environment. Such remote interaction is particularly beneficial in confined, unsafe, or sensitive environments such as hazardous sites, underwater or space. It naturally combines human high-level intelligence and robot physical capabilities while maintaining the safety and comfort required for the Human. Unfortunately, conventional teleoperation methods do not leverage the robot assistance and collaborative ability to its fullest, since the operator fully controls the remote task, with a high mental workload and poor performances.
Recent shared-autonomy concepts have been proposed in the literature to transfer part of the task from the human to the robot. These approaches range from complementary and predefined subtask allocations to adaptive shared-control methods. Focusing on this second and more flexible paradigm, the postdoctoral project aims at improving shared control in haptic teleoperation.
To act as an effective collaborator, the robot should adapt its assistive behavior with respect to the human intent. The human inputs are first analyzed to infer the operator goal (such as the target object in pick-and-place) and consequently planned the robot assistive behavior. The authority level, that gives the impact of each agent on the action, is then fixed depending on some task-oriented criteria. The shared-autonomy approach combines the human motions and the robot assistive motions into a joint action, based on the authority distribution. A Model Predictive Shared Controller is developed at Auctus team [1] to compute the robot motion on a time horizon, given both the human and robot assistive trajectories. It generates a unified action, formulated as an optimization problem under robot and human limits and environmental constraints.
Two key aspects must be studied in our shared-autonomy approach to improve human-robot coordination and better share the task. According to their background and preferences, the postdoctoral scholar will work on these challenges:
- Developing a method that dynamically adapts the authority level to better distribute the task between the human-robot agents. In conventional approaches the authority level is computed at each task state, as a function of task-oriented criteria (proximity to target [2]) or human-based indices (expertise [3], human activity [4]). The IriS lab has proposed a dynamic authority distribution method [4] that allocates the control authority in real-time from the energy produced in the human-robot interaction. This approach should be extended to a global optimization problem that capture task, human, environment factors to online shift the authority level.
- Improving human intent detection in our shared controller to better adapt the robot assistance to the human need. Up to now, we use simple task descriptions to detect the human task goal and plan the robot assistive motions. We want to go through the extensive literature about human intent prediction [5] to implement more advanced methods that decode the user commands and infer his/her intention. Such intent prediction will be based on Hidden Markov Models or Gaussian Mixture Regressions. It will predict the most likely elementary action, encoded as a force-motion manipulation pattern, that the Human wants to do.
Skills
The candidate should have graduated with a PhD in robotics. He/she should have solid skills in robotic control, programming (C++, Python), and kinematic/dynamic modeling. Any additional experience in haptics, telerobotics, planning, or machine learning would be appreciated. We would value past balanced researches that had combined fundamental works to experimental studies.
Instruction to apply
Applications for this Inria-KAIST DRI postdoc are submitted online and must include:
- A detailed CV with a description of the PhD and the list of publications.
- A motivation letter showing the candidate interests in the research project.
- 2 letters of recommendations.
- A passport copy.
Lab contacts:
References
[1] E. Jabbour, M. Vulliez, J-P. Gazeau, V. Padois, C. Préault, “Haptic shared control in human-robot collaboration”, Poster at JJCR 2023 (Journée des Jeunes Chercheurs en Robotique), Oct 2023, Moliets et Maâ, France
[2] V. K. Narayanan, A. Spalanzani, and M. Babel, “A semi-autonomous framework for human-aware and user intention driven wheelchair mobility assistance,” in 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS 2016, Daejeon, South Korea, October 9-14, 2016, pp. 4700–4707, IEEE, 2016
[3] C. E. Mower, J. Moura, and S. Vijayakumar, “Skill-based shared control,” in Robotics: Science and Systems XVII, Virtual Event, July 12-16, 2021 (D. A. Shell, M. Toussaint, and M. A. Hsieh, eds.), 2021
[4] USMANI, Naveed Ahmed, KIM, Tae-Hwan, et RYU, Jee-Hwan. Dynamic authority distribution for cooperative teleoperation. In : 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2015. p. 5222-5227.
[5] LOSEY, Dylan P., MCDONALD, Craig G., BATTAGLIA, Edoardo, et al. A review of intent detection, arbitration, and communication aspects of shared control for physical human–robot interaction. Applied Mechanics Reviews, 2018, vol. 70, no 1, p. 010804.