Abstract
Movements provide the only means we have to interact with both the world and other people [1] and mediate communication, including sign language, gesture and writing. Following the concept of robotic re-embodiment, we aim at providing users with limited mobility and manipulation capabilities the possibility to interact again with their surroundings by means of a robot avatar. This robot avatar is solely controlled through a selective-attention EEG-based brain-computer interface (BCI), which bypasses the natural muscular and neural pathways [2]. For non-invasive BCIs based on P300 or steady-state visual-evoked potentials (SSVEP), users attend to one stimulus (target) out of many in order to encode their intentions. In our implementation, intentions are centered around goal-oriented commands like navigation- and pick-, and place-commands or incremental commands to realize small movements of either robotic platform or arms.
In first experiments we asked users to navigate the robot avatar to a predefined place in space, to pick up an object of interest and to place it at another known position. Such goal-oriented commands were mediated through a P300-based interface, whereas incremental commands, such as the forward or backward movement of the robotic platform, were mediated through a SSVEP-based BCI interface. Stereoscopic video streams were additionally captured at the remote site from the ego- perspective of the robot avatar and were presented to the user through a head-mounted display. This way, we aimed at increasing the feeling of embodiment to gain a more intuitive way of interacting with the remote environment.
Results showed that most users were able to use the BCI system to complete the given task, which can be considered a first success in realizing systems for robotic re-embodiment. In future, we aim at building embodiment systems that increase the user’s feeling of embodiment by intelligently adapting the presented controls as well as the way of executing commands to the user’s mental and environment state.