International Journal of Advancements in Electronics and Electrical Engineering
Author(s) : GENCI CAPI, MARSEL MANO, MITSUKI KITANI, ZULKIFLI MOHAMED
Recent research has shown that Brain Machine Interface (BMI) can be used to assist disable people in navigating a robotic wheelchair by using voluntary mental intentions. BMI based navigation is a very challenging task. In this paper we present a novel adaptive method to improve BMI based robotic wheelchair navigation. The robot is controlled by an adaptive navigation platform that provides the user with scalable navigation assistance. The platform is able to detect and avoid collisions by using a laser range finder sensor. Furthermore, by using computer vision it can read assistive information for visually impaired people (tactile paving) on the floor and autonomously navigate the robot following tactile paving directions. Based on user intentions and environment context, the robot navigation adaptively changes between assisted and unassisted mode. Experimental results show that with the assistance of the adaptive navigation platform the robot navigation improves significantly. Furthermore, the user’s mental focus is reduced and BMI classification accuracy is improved as a consequence.