Brain-actuated Pick-Up and Delivery Service for Personal Care Robots: Implementation and Case Study

In this paper, we propose a smart architecture able to provide an automated pick-up and delivery service for personal care assistance. The presented architecture consists of a human-robot interface that connects the user intentions, at the cortical level, with the functionalities of a personal care robot (PCR). This interface must, firstly, acquire and interpret the user’s electroencephalographic (EEG) signals. Then, it must uniquely formalize these EEG-driven requests, and continuously communicating with the environment to provide an online-updated list of available services. The users’ intentions recognition is entrusted to a nested 2-choice asynchronous Brain-Computer Interface (BCI). It bases the feature extraction and discrimination steps on an end-to-end binary technique: the local binary patterning (LBP). The experimental results demonstrated that the LBP-based BCI, here proposed, can decode EEG and drive the actuator in ~883ms with an accuracy of 84.22%. Also, the tests proved that the 79.2% of the requests have been successfully satisfied by the PCR.