Hand Gesture Wheelchair Control Based on the Wrist Rotation with Agglomerative Hierarchical Clustering Method

Abstract

A person with disabilities is in a state of self-limitation, which can be physical, cognitive, mental, sensory, emotional, and other, so they need assistive devices. A wheelchair is a tool used to help people with walking difficulties. The use of conventional wheelchairs and electric wheelchairs with a joystick is uncomfortable for a person who has a problem with fingers and toes. This study provides an alternative method to control an electric wheelchair using wrist rotation to solve the limitations of conventional wheelchairs and enhances mobility for individuals without fingers. Two flex sensors and one gyro sensor is attached to the hand to detect five wrist rotation for forward, backward, turn right, turn left, and normal (stop). An Agglomerative Hierarchical Clustering (AHC) is implemented to recognize the wrist rotation. There are 450 training data for each wrist rotation with 100% accuracy. The developed algorithms is tested for real implementation with six adults operating the wheelchair along the trajectory. The result shows that this method is available to control the electric wheelchair. Finally, this intelligent electric wheelchair system controlled by wrist rotation based on the AGHC method can be used as an alternative to help persons with a disability in moving, especially those with problems with their fingers.