Title: Intention Detection for Physical Human-Robot Interaction  Using Machine/Deep Learning

Faculty: Prof. Dr. Cagatay Basdogan (cbasdogan@ku.edu.tr)


Cobots (collaborative robots) are designed with the aim to integrate human cognitive abilities and quick adaptation skills with the speed, strength, and repeatability of robots. As cobots penetrate into different domains such as manufacturing, logistics, medicine, home use, etc, it is becoming increasingly clear that the skills expected from them are significantly different from what is expected from industrial robots. In fact, many of these skills center on the challenge of interacting with humans. In particular, understanding human intention during a collaborative task and then adjusting the interaction controller of the cobot accordingly to comply with human intention is a major challenge in physical human-robot interaction (pHRI). The post-doctoral fellow is expected to design human-human and human-robot experiments, collect data from various sensors, and develop models based on machine/deep learning techniques to estimate human intention during a collaborative manipulation task. She/he is expected to have research experience in robotics, AI, and machine/deep learning with a background in CS, EE, or ME.

One of our applications is collaborative drilling with a robot. Here, learning algorithms are used to understand the sub-task executed by human operator first and then the admittance controller of the cobot is adjusted accordingly to help with that sub-task. The major components of our setup are a power drill, two force sensors (Mini40, ATI Inc.) and a handle attached to the end effector of a LBR iiwa 7 R800 robot (KUKA Inc.), and an augmented reality interface (Hololens, Microsoft Inc.). One of the force sensors is used to measure the interaction force between the drill bit and the workpiece, while the other one measures the force applied by human operator alone. The augmented reality interface informs the operator about the steps of the task, distance of the drill bit to the workpiece, and the current drilling depth. 

Some Related References:

  1. Sirintuna, D., Ozdamar, I., Aydin, Y., Basdogan, C., 2020, “Detecting Human Motion Intention during pHRI Using Artificial Neural Networks Trained by EMG Signals”, Proceedings of IEEE International Conference on Robot and Human Interactive Communication (Ro-Man), Aug 31-Sept 04, Naples.
  2. Sirintuna, D., Aydin, Y., Caldiran, O., Tokatli, O., Patoglu, V., Basdogan, C., 2020, “A Variable Fractional-Order Admittance Controller for pHRI”, Proceedings of IEEE International Conference on Robotics and Automation (ICRA), pp. 10162–10168, Paris, Jun 1-5.
  3. Aydin, Y., Tokatli, O., Patoglu, V., Basdogan, C., 2021, “A Computational Multi-Criteria Optimization Approach to Controller Design for Physical Human-Robot Interaction”, IEEE Transactions on Robotics, to appear.
  4. Sirintuna, D., Aydin, Y., Basdogan, C., “Towards Collaborative Drilling with a Cobot Using Admittance Controller”, Transactions of the Institute of Measurement and Control, to appear.
  5. Aydin, Y., Tokatli, O., Patoglu, V., Basdogan, C., 2018, “Stable Physical Human-Robot Interaction Using Fractional Order Admittance Control”, IEEE Transactions on Haptics, Vol. 11, No.3, pp 464-475.
  6. Kucukyilmaz, A., Sezgin, T.M., Basdogan, C., 2013, “Intention Recognition for Dynamic Role Exchange in Haptic Collaboration”, IEEE Transactions on Haptics, Vol. 6, No. 1, pp. 58-68.
  7. Mörtl, A., Lawitzky, M., Kucukyilmaz, A., Sezgin, T.M., Basdogan, C., Hirche, S., 2013, “The Role of Roles: Physical Cooperation between Humans and Robots”, International Journal of Robotics Research, Vol. 31, No. 13, pp. 1656-1674.