Multi-Modal Assistive Head Gesture Recognition-Based Smart Wheelchair for a Paralyzed Person

Abstract

Assistive mobility technology significantly enhance the quality of life for those with severe motor limitations. This research describes a multimodal smart wheelchair for paralysed people that is operated by head gesture recognition. To record head motions in real time, the suggested system combines vision-based tracking and Inertial Measurement Unit (IMU) sensors. The system recognises basic movements such as backward, left, right, and forward. A sensor fusion strategy reduces mistakes caused by noise or changes in the environment and increases robustness. The system is built to function with low latency in real time. Several users participated in the experimental validation under various circumstances. The overall accuracy of the suggested model's gesture recognition was 96.2%. The response time was found to be within reasonable bounds for secure operation. In both indoor and semi-outdoor settings, the system performs dependably. There are safety features like neutral detection and automatic stop. Cost-effectiveness and usability are prioritised in the design. The suggested approach offers hands-free control in contrast to traditional joystick or voice-controlled devices. Users with significant upper limb limitations will benefit most from the solution. IOT integration and obstacle detection are two ways to expand the system. The recommended smart wheelchair offers an assistive mobility solution that is scalable, efficient, and easy to use.