Robots have pertained our lives in various forms (as robots used in domestic settings such as vacuum cleaners and robot lawn mowers, as smart toys, in therapies, in stores as guides etc.) While we see these robots as servants or entertainers in some situations, we also can see them as care partners in other situations - for instance, when we provisory lose our abilities due to an accident, a physical hazard at work, a stroke, or in other situations. Feeding is a necessary activity for us to sustain our lives, like breathing. When we lose this ability due to a specific life circumstance, this can become a challenge.
This master thesis aims at investigating how humans may perceive various robots (UR5, TIAGo) in a feeding scenario. The student taking on this project can focus either on the development of an algorithm for using UR5 in a feeding situation, or applying the algorithm to both UR5 and TIAGo and compare the efficiency results in terms of safety, accuracy. When the implementation is ready, the student will conduct experiments with real users on the task of feeding, where their perceptions and acceptance of a such robot executing this task are investigated. Here, the student can investigate: perceived safety, perceived privacy, perceived dignity and other similar values that are appropriate in a such situation.
Goal of the project: Investigate the task of feeding with one or two physical robots (UR5, TIAGo), to be used during the rehabilitation process, for care receivers who cannot use their arms or have reduced mobility.
Research questions:
RQ1. To what extent can a robotic arm be programmed to efficiently and safely perform a feeding task for users, balancing precision, adaptability, and safety in the process?
RQ2. How do humans in vulnerable situations perceive the activity of feeding with a robot as a care partner? What is their perception of safety, privacy? Does the robot act intuitively in this according to the user? Would the users accept a such robot to feed them? if yes, in which situations? if not, why not? Which of the robots would the users prefer: UR5 or a more humanoid robot, such as TIAGo, and why?
Project tasks:
- Get an overview of earlier work on robots used in activity of feeding
- Decide how to perform the feeding task with the robot depending on how much the student wants to be involved in the technical aspects. The task can be fully autonomous, semi-autonomous, or fully controlled by an operator (Wizard of Oz). It could range from a simple teleoperated control task using a joystick to a fully autonomous process, which would require the development of appropriate motion planners or suitable sensing frameworks for monitoring the user during the feeding process.
- Develope relevant algorithms for the robot depending on the desired level of autonomy for the task, from basic operator control to fully automated execution by the robot.
- Eventually develop a physical add-on (physical prototype) for the robot to be able to execute better the activity of feeding (not mandatory)
- Conduct user studies (either a video-based survey, in-lab experiment, or an experiment outside of the lab with real users) - details to be decided in agreement with the student.
- Compare the results to previous work available.
- Write the master thesis
Starting point references:
[1] T. Schulz, J. Torresen, and J. Herstad, “Animation Techniques in Human-Robot Interaction User Studies: a Systematic Literature Review,” ACM Trans. Hum.-Robot Interact., vol. 8, no. 2, Jun. 2019, doi: 10.1145/3317325.
[2] Xu, T., Zhao, T., Cruz-Garza, J. G., Bhattacharjee, T., & Kalantari, S. (2023). Evaluating Human-in-the-Loop Assistive Feeding Robots Under Different Levels of Autonomy with VR Simulation and Physiological Sensors, Proceedings of the 14th International Conference on Social Robotics (ICSR 2022) (https://link.springer.com/chapter/10.1007/978-3-031-24670-8_28)
[3] Park, D., Kim, Y. K., Erickson, Z. M., & Kemp, C. C. (2016). Towards Assistive Feeding with a General-Purpose Mobile Manipulator. In Proceedings of the ICRA 2016 Workshop on Human-Robot Interfaces for Enhanced Physical Interactions. arXiv:1605.07996.
Relevant courses
- IN4140 Introduction to robot technology or good familiarity with ROS will be necessary to work with the robots.
- IN5400 Machine learning for image analysis can be useful if the student wants to develop AI methods for real-time user monitoring during feeding by the robot.
- IN4050 Introduction to artificial intelligence and machine learning can be useful for implementing or understanding algorithms
- IN5590 Rapid Prototyping of Robotic Systems may be useful if you want to add a physical add-on to a robot (a glove to the robot gripper)
- IN5490 Advanced Topics in Artificial Intelligence for Intelligent Systems can be helpful for understanding how to read articles and write your masters.
Contact:
Diana Saplacan Lindblom dianasa@ifi.uio.no & Adel Baselizadeh adelb@ifi.uio.no