Abstract:
In this paper, we discussed the use of brainwaves to make an object move in one
of two directions namely right or left. Our aim was to find the type of brain waves
that was most suitable to interpret the intention of a volunteer when using his/her
vision, hearing and pure thinking. We proposed a methodology in which the brain
was excited using six stimulants which were Vision Right, Vision Left, Voice
Right, Voice Left, Thinking Right and Thinking Left. Recordings of four
volunteers were obtained using a Neurosky headset. Each recording contained 60
readings for one of eight types of brainwaves and a class representing the
direction of motion intended by the volunteer. The average of the recordings for
the right class was higher than that for the left one. Moreover, the standard
deviation of the recordings for the left class was higher than that for the right.
The average and standard deviation parameters meant that the identification of
the recordings for the left class was higher than that for the right class. Moreover,
the K-nearest neighbour was used to classify each recording to either right or left
class. The Alpha family of brain waves showed values of relative standard
deviation in the range of 80 to 120% for 5 stimulants out of the 6 stimulants
tested. This meant that recordings from the Alpha family contained enough
information to classify them correctly and identify the type of direction intended
by the volunteer. In addition, they were not highly contaminated with noise. The
Alpha2 brain wave type showed the highest accuracy in our classification
problem. 82% was the percentage achieved for the Precision, Recall and F-
measure parameters used to describe the accuracy of success in classification.
