Augmenting robot intelligence via EEG signals to avoid trajectory planning mistakes of a smart wheelchair

Abstract Assistive robots operate in complex environments and presence of human beings, but the interaction between them can be affected by several factors, which may lead to undesired outcomes: wrong sensor readings, unexpected environmental conditions, or algorithmic errors represent just a few examples possible scenarios. When safety user is not only an option must guaranteed, feasible solution rely on human-in-the-loop approach, e.g., monitor if robot performs action during task execution conditions affect human-robot interaction, provide feedback accordingly. The present paper proposes framework enable safe autonomous navigation electric powered sensorized (smart) wheelchair. During wheelchair towards desired destination indoor scenario, problems (e.g. obstacles) along trajectory cause generation electroencephalography (EEG) potentials when noticed user. These used as additional inputs algorithm order modify planning preserve safety. has been preliminarily tested using simulator implemented ROS Gazebo environments: EEG signals from benchmark known literature were classified, passed custom simulation node, made available stack perform obstacle avoidance.