Inertial sensors and muscle electrical signals in human-computer interaction

Abstract

Assistive technology, such as interactive computer applications, has a major role in providing independence to many individuals, but computer interaction using traditional input devices can be challenging for people with disabilities. In this study, a bimodal computer control device is proposed uniting muscle electrical signals and inertial sensor data to provide efficient manual target selection in addition to existing inertial sensor-based solutions for head position tracking and computer cursor control. An embedded system consisting of 9-axis inertial measurement unit and electromyography sensors was proposed and a wireless headband prototype was developed in order to measure system performance and compare it with similar studies. Results show that manual target selection using facial muscle electrical signals instead of automatic dwell time increases the speed of human-computer interaction.