Hand Gesture Recognition-Based Control of Motorized Wheelchair using Electromyography Sensors and Recurrent Neural Network

Mobility has been identified to be a major characteristic of living things. Humans who are deprived of efficient mobility either by natural or man-made factors loose significant relationship with their environment. The growing demand to produce effective rehabilitation devices for the aged population and disabled individuals, have spurred us to develop a reliable and easy to use biosignal based auto control wheelchair. This is to ensure independent mobility of persons with disabilities and the aged. In this paper, a Recurrent Neural Network (RNN) architecture called Long Short Term Memory (LSTM) is engaged for the classification EMG signals to the corresponding hand-gesture category. The LSTM model in this study yielded a validation accuracy that provides a basis for an improved solution towards real-time deployment.