Robotic Re-Embodiment means the physical embodiment of a person on to a robot avatar using physiological signals as control inputs. This can enable people with limited mobility, such as para-plegics or tetra-plegics, to enhance their participation and interaction in day-to-day life situations.

The main goal of this research is to increase the feeling of embodiment, i.e. that a task is actually performed by the avatar as intended by the user. Several research fields arise from this challenge including the following which are adressed at this institute:

1. Semi-autonomous control
   The anticipation and suggestions of upcomming actions as well as the adjustment accordingly in a closed-loop design between the user and the avatar require new semi-autonomous control strategies.
2. Incorporation of physiological and mental state into motion
   Advances in autonomous robotics provide a considerable range of tasks a robot is capable of. The next step towards increasing the feeling of emodiment, is to alternate a specific motion according to the state of the user. We believe that adopting the movement the same way it would vary if it was performed by the real body will strongly contribute to this end.

One way of abstracting these states is by using emotions. These are extracted from EEG-Signals, i.e. brain waves. We investigate an apropriate mapping of the state-variables on to the motion parameters. Herein, the following questions are asked:

• How are different emotions are displayed in daily activities and motions?
• How can we model the emotions and identify parameters associated with them?