Mohammad Thabet

Exploiting social cues for safe human-robot interaction in home environments

Principle Supervisor:
Prof. Dr. Angelo Cangelosi
University of Manchester

Collaboration partners:

  • University of Hertfordshire
  • Universität Hamburg

Competence Area: Interaction


To investigate the dynamic alignment between robots and humans, specifically for the role of non- verbal behaviours in supporting safe interaction and linguistic communication. Joint action and cooperation between humans and robots are a key issue for the design of safe service robots. Cooperation is based on both linguistic interactions (e.g. a human giving instructions, and the robot asking for clarifications) and non-verbal behaviour (e.g. gestures, facial feedback).

Expected Results

The design of the cognitive architecture and the human-robot interaction experiments will result in novel design principles and mechanisms for the robot’s use and understanding of gestures (e.g. pointing) and facial expressions (gaze direction) to support the learning and understanding of demonstrative pronouns such as “this” “that” during joint tasks between the robot and human participant.

C.V. Highlights

  • Double M.Sc. in space science and technology, Luleå University of Technology, Sweden, and Aalto University, Finland. Major in space robotics and automation
  • B.Sc. in Electrical Engineering, Ain Shams University, Egypt. Major in computer and systems engineering.
  • Former researcher at  the Intelligent Hydraulics and Automation department, Tampere University of Technology, Finland.
  • Former control systems engineer at PGESCo, Egypt.


  • Mohammad Thabet, Massimiliano Patacchiola, Angelo Cangelosi. “Sample-efficient Deep Reinforcement Learning with Imaginary Rollouts for Human-robot Interaction.” IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2019).
  • Mohammad Thabet, Massimiliano Patacchiola, and Angelo Cangelosi. "Toward Imagination-assisted Deep Reinforcement Learning for Human-robot Interaction." International PhD Conference on Safe and Social Robotics (SSR-2018), 2018
  • Mohammad Thabet, Alberto Montebelli, and Ville Kyrki. "Learning movement synchronization in multi-component robotic systems." Robotics and Automation (ICRA), 2016 IEEE International Conference on. IEEE, 2016.