A robot for brain-controlled grasping

In this paper we present a system that allows patients with motor disabilities to grasp everyday objects. A closed-loop Brain-Machine Interface (BMI) instantiates a connection between the human central nervous system and an industry robot designed to execute the grasping task. Two central problems were addressed in this paper, namely I) the implementation of a BMI to translate a subject's voluntary modulation of brain activation patterns into commands for target selection and grasp execution and II) the development of a robot that autonomously grasps natural objects. Our approach in the work was to implement as much intelligence as possible into an industry standard robotic system to claim a minimum of information flow from the patient to the system. The final system demonstrates the feasibility of brain controlled grasping of natural objects with a robotic arm which is an important step towards the development of intelligent prostheses for paralyzed patients.