Human-in-the-Loop Approach for Enhanced Mobile Robot Navigation

Human-robot interaction (HRI) studies realistic and applicable solutions for assisting both healthy and disabled people in their Activities Of Daily Live (ADL), supported by a non invasive Electroencephalography (EEG) signals recording system used to provide inputs to some devices or algorithms to execute some control action or correct possible faults and errors during the robotic task. The aim of this work is to design a BrainComputer Interface (BCI) protocol that exploits EEG signals to utilize specific brain cognitive responses as Error-related Potential (ErrP). Data acquired in real experimentation are processed for testing different machine learning and classification algorithms implemented in open source platforms, in order to choose the best one for the real-time approach of involving humans in the mobile robot controller. The results from this study have shown that the proposed methods and the designed protocol are applicable to be used as a real-time approach to involve human in control-loops of a mobile-robot to enhance the obstacle detection.