Abstract
Brain Computer Interface (BCI) offers an effective medium of communication for physically disabled people who are partially or completely locked in their body. It also improves the performance level of healthy people by providing them a new input modality. Recently, BCI systems based on motor imagery electroencephalogram (EEG) signals have been focussed intensively. Extraction and classification of motor imagery features is the most crucial part of motor imagery based BCI system. The biggest challenge in processing motor imagery EEG signals is that these signals are non stationary. In the present investigation, left hand and foot motor imagery tasks have been classified. Time-frequency features have been extracted using discrete wavelet transform (DWT) to deal with the non stationarity of motor imagery EEG signals. Linear discriminant analysis (LDA) is used to reduce the dimensionality of data. The efficacy of DWT and LDA has been tested using state vector machine (SVM), k-nearest neighbour (k-NN) and artificial neural network (ANN) classifiers. Overall, combining DWT with LDA and ANN achieved a promising average accuracy of 90% in 4.453 secs. The present study may prove to be helpful in developing intelligent devices to help people in locked-in state for controlling some application by imagination of the movement of their body parts.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call