Brain Controlled Switching Mode for Paralytic Patients Using Wavelet Transform in EEG Signal Processing

摘要

When it comes to the disability of a human body the EEG has been approved as the most reliable alternative to perform simple day to day tasks. The study of BCI has expanded since long mainly due to the increased interaction of people with machines for applications involving monitoring and detection of rare events and weak signals. The control of switching is one of the most simple task performed number of times in a day by us, but the same task is seemingly impossible for a paralytic patient. Processing of the EEG signals is able to overcome these disabilities. Here active EEG is used to determine concentration of a person; this in turn is used for switching mode operation. Wavelet Transform is used here in EEG signal processing. The proposed framework explores the use of wavelet transform in EEG signal processing. Among the various time frequency methods used in analysis of non stationary signal like EEG, wavelet analysis is the most efficient method in the frequency domain. For asynchronous tasks which take a long duration of continuous time or even real-time experiments, like switching operations, mental concentration may change smoothly and continuously, and the classification of different mental states, i.e. concentration levels, is more difficult. In this case, traditional continuous or discrete wavelet-based feature extracting methods are not competent enough to capture the minute and detailed change of rhythmic activity in the frequency domain. Here we mainly focus on the power spectrum. The electroencephalographic activities in the δ(0-4 Hz), θ(4-8 Hz), α(8-13 Hz) and β(13-35Hz) bands, reflect the change of the physiological concentration. The main aim here is to establish a method for processing EEG signals and distinguishing between concentration levels, thus paving the way for a real-time system which can accurately monitor the concentration levels of a person.