Abstract

As autonomous service robots become more affordable and thus available alsofor the general public, there is a growing need for user friendly interfaces tocontrol the robotic system. Currently available control modalities typicallyexpect users to be able to express their desire through either touch, speech orgesture commands. While this requirement is fulfilled for the majority ofusers, paralyzed users may not be able to use such systems. In this paper, wepresent a novel framework, that allows these users to interact with a roboticservice assistant in a closed-loop fashion, using only thoughts. Thebrain-computer interface (BCI) system is composed of several interactingcomponents, i.e., non-invasive neuronal signal recording and decoding,high-level task planning, motion and manipulation planning as well asenvironment perception. In various experiments, we demonstrate itsapplicability and robustness in real world scenarios, consideringfetch-and-carry tasks and tasks involving human-robot interaction. As ourresults demonstrate, our system is capable of adapting to frequent changes inthe environment and reliably completing given tasks within a reasonable amountof time. Combined with high-level planning and autonomous robotic systems,interesting new perspectives open up for non-invasive BCI-based human-robotinteractions.

 

Introduction (beta)

None

 

Conclusion (beta)

None