Implementation of Wheelchair Motion Control Based on Electrooculography Using Simulation and Experimental Performance Testing

Abstract

The aim of this study is to perform the experimental verification on the fuzzy-based control designed for wheelchair motion. This motion control based on the eye movement signals using electrooculograhphy (EOG) technique. The EOG is a technique to acquire the eye movement data from a person, i.e tetraplegia, which the data obtained, can be used as a main communication tool. This study is about the implementation of the designed controller using PD-type fuzzy controller and tested on the hardware of the wheelchair system using the eye movement signal obtained through EOG technique as the motion input references. The results obtained show that the PD-type fuzzy logic controller designed has successfully managed to track the input reference for linear motion set (forward and backward direction) by the EOG signal.

Human–Machine Interface: Multiclass Classification by Machine Learning on 1D EOG Signals for the Control of an Omnidirectional Robot

“…Other systems classifying EOG signals using fuzzy logic and a database that store waveform information from different users have been developed [ 14 , 15 , 16 , 17 , 18 ] in order for the interface to compare the parameters of each user with previously established commands. One of the most representative works is presented in [ 19 ]; Fuzzy PD control is applied to the horizontal EOG channel that generates a wheelchair’s rotation to the right or left and the vertical EOG indicates forward or reverse. In [ 20 ], a writing system for people with disabilities is designed; the similarity of the trajectories generated by the movement of the eye and the shape of the letters is determined by fuzzy Gaussian membership functions.…”

“…The authors in [ 22 ] also developed a portable EOG acquisition system, which generates position control commands for an industrial robot using a nine-state machine, concerning which it was tested whether making the end effector followed five points; the obtained response time was of 220 s with trained users. In [ 23 ] a review of EOG based human–computer interface systems is presented; the work of 41 authors is explained, where the interfaces used to move a device always generate points in coordinates X-Y, as is the case for control of wheelchairs, Mohd et al [ 19 ]. In this paper, research did not generate a three-dimensional workspace; unlike the one presented in [ 24 ] where the EOG signals activate a robot with three degrees of freedom in 3D Cartesian space, the Cartesian coordinates X, Y, Z are generated by a fuzzy classifier that is automatically calibrated using optimization algorithms.…”