Human Activity Recognition Using Portable EEG Sensor and Support Vector Machine

Applications of human activity recognition in health care and monitoring of Activities of Daily Livings (ADLs) can assist in the detection of abnormalities that may result in neurological disorders such as Alzheimer's, stroke and epileptic seizures, etc. Wearable sensors such as accelerometers embedded in smartphones, EEG headsets, etc. can be used for continuous monitoring of ADLs. This research aims to detect ADLs using wearable EEG headset Neurosky MindWave. The data files are recorded in an un-constraint environment from volunteers performing different activities. Collected data is preprocessed using Discrete Wavelet Transform and average power. The selected features are further used for the classification of activities into stationary, light ambulatory, and intense ambulatory using SVM. The results show that SVM classifier can detect ADLs with reasonable accuracy of 95.75% in an un-constraint environment using EEG signals from a wearable headset. Therefore, EEG can be used for long-term activity monitoring and assisted living applications.