Fractal-Structured, Wearable Soft Sensors for Control of a Robotic Wheelchair via Electrooculograms

Abstract

Neurodegenerative diseases create a significant issue by affecting the mobility of individuals. One such disease is Parkinson's disorder, which is a chronic problem for more than 1 million citizens in the United States. Typical symptoms, including tremors, imbalance, and decreased mobility, are a resultant of the death of nerve cells in the brain. As the disease continues to further deteriorate the health of a person, the individual is in need of a wheelchair. The problem is that tremors make the use of a conventional joystick difficult. Here, we propose a soft wearable electrode system for control of a robotic wheelchair without the use of a joystick. A set of skin-like electrodes enables users to control a wheelchair via electrooculograms (EOG) from eye movements. Advances in data acquisition and classification algorithms allow a wireless human-machine interface. We use a set of statistical measures with machine learning techniques. The linear discriminant analysis (LDA) classifier yields the classification accuracy of ~87% and ~92% for rigid gel electrode and soft fractal electrode, respectively. Collectively, the wearable, soft electronics with the electronic wheelchair interface provide a non-invasive, persistent human-machine interface.