Robot Control through BMI

Abstract

Robot navigation is a critical component of mobile robots. In non-autonomous modes, robots are typically controlled using direct commands through keyboard, joystick, or mouse. However, in certain cases, such as robotic wheelchairs, a disabled patient might not be able to use these methods. This paper presents an indirect method of robot control used for navigation. Our method uses analyzing EEG signals of brain though a deep learning network which is used to control a robot. The system can detect four states of the person with good precision. These states can be used for controlling a mobile robot. Results with actual devices show that the method can accurately predict user’s motives. The robot control is demonstrated through a simulation software.