Optimized tongue driven system using artificial intelligence
Mansour H. Assaf, Rahul Kumar, Krishneel K. Sharma, Bibhya Nand Sharma · Computer Methods in Biomechanics and Biomedical Engineering Imaging & Visualization · 2022
This paper presents a cost-effective design of a wearable wireless tongue drive system (TDS) for disabled individuals, particularly with spinal cord injuries. We propose a basic TDS whose language is specifically designed for issuing movement-based commands, be it a gadget or a self-governing transport chair. While the overall industrial penetration of TDS is low, we offer a cheap and straightforward design for a TDS that consists of items which amount to as low as US$30. A data-driven-based approach has been used to classify the tongue-gestures with a response time of less than 0.25 s. Compared with other machine learning algorithms, the proposed TDS is equipped with shallow neural networks that govern all the decisions when it comes to classifying the tongue gesture. This has been achieved after a thorough data analysis and rigorous comparison with other classification techniques along with hardware deployment. With its open architecture, the newly designed TDS can be an ideal tool to control other movement-based peripherals. The final test accuracy amounts to 93.4% when the TDS is tested on human subjects.