Alessandra Rossi

Investigating human perceptions of trust in robots for safe HRI in home environments

Principle Supervisor:
Prof. Dr. Kerstin Dautenhahn
University of Hertfordshire

Collaboration partners:

  • Universität Hamburg
  • Ecole Polytechnique Federale de Lausanne

Competence Area: Interaction


To investigate and develop socially inspired techniques whereby through interaction with a human a robot learns patterns and limits for its own and the human’s safety. This may be achieved through techniques of observing scenarios within a domestic environment, or directly from teaching signals generated by the human partner. For example the robot could learn, via observation of human behaviour in the house, that it is important for a human to always be present if the cooker is on a high setting and warn the person accordingly if this is not the case. In direct interaction with the robot teaching signals could be used to indicate to the robot as to what is a safe and preferred interaction distance or for example areas in the house which would be unsafe for the robot to enter.

Expected Results

Over time and though interaction the robot would learn the safety procedures operating within a typical household environment. These procedures may in fact be obvious to the human, and therefore not explicitly recognized, however we would expect the robot to infer these procedures from observing the human and house environments via interaction or from polling its own and environmental sensors.