Sub-band-power-based efficient Brain Computer Interface for wheelchair control

Versatile Brain-Computer-Interface for Severely -Disabled People

“…Since this work targets severely disabled people, it is reasonable to incorporate as little constraints as possible in the proposed BCI. Hence, in contrast with previous works [7], the proposed system accommodates eye blinks and movements. In fact, these artefacts are handled in two different ways.…”

“…This section summarizes some background material about the signal processing tools used to extract the EEG features and classify the measured signals. Starting with the classifiers, most researchers resorted to Neural Networks [7,[17][18] owing to their established simplicity and fairly good performance. Other classifiers have also been used to a less extent such as Linear Discriminants (LD) [19], Bayesian [20], Hidden Markov Model (HMM) [21], and Support Vector Machine (SVM) [22].…”

“…While some researchers used a time-series prediction approach and derived their features from the power of the predicted EEG signals [23], most others resorted to the Wavelet Transform (WT) [24][25][26][27][28] wherein the coefficients of the resulting detail signals are used as features. In order to reduce the feature space dimensionality, a recent work [7] alternatively proposed to use the average powers (Mean Square) of the detail and approximation signals leading to a substantial reduction in the feature vector dimension without noticeably compromising the classification performance. In addition to efficiency, there is a clear practical justification for using the WT with such an averaging as it leads to a sub-band-power decomposition of the EEG signals.…”

“…These rely on the weak Electroencephalogram (EEG) signals that arise from the neural activities and are measured using non-invasive electrodes suitably placed on the skull's surface. Introduced by Vidal in 1973 [1], BCI is currently an active research direction with applications in diverse areas including intelligent home control [2], speech synthesis [3], spelling applications [4], readiness detection [5], Epilepsy Prognosis [6], wheelchair control [7], microsleep prevention [8], limb rehabilitation [9], mobile robots [10,11], drowsiness control [12], and assistive systems for people with severe handicaps enabling them, for example, to control electronic devices [13] or browse the internet [14] .…”

“…A versatile BCI system is used to achieve this, allowing users, for example, to control some appliances, self-adjust their laying positions, or call for emergency simply using their thoughts. It builds on a recent work [7] utilizing an efficient sub-band-powerbased BCI system in order to achieve a versatile and smart assistive system featuring flexibility, interactivity, http://dx.doi.org/10.12785/ijcds/100145 http://journals.uob.edu.bh asynchronous real-time operation, twofold-artefacttreatment, and fault-tolerance.…”