IIT Guwahati develops model to turn air vibrations to voice commands

The Indian Institute of Technology Guwahati (IIT Guwahati) has developed an underwater sensor which can convert exhaled air vibrations into voice commands. This research is set to aid people with voice disabilities for using the voice recognition technology with mobile phones and home appliances.

The new underwater voice technology uses a chemically reactive porous sponge, placed directly below the air-water interface. The exhaled air from the mouth produced subtle waves on the water, which are then converted to measurable electrical signals. Convolutional Neural Networks (CNN) is a deep-learning model which “accurately” recognises the signal patterns, allowing the users to communicate without generating sound.

The lab-scale working prototype of the innovation costs approximately Rs 3,000. This cost is expected to be reduced as the product becomes available for real use through industry partner collaborations. The prototype was developed in partnership with the Ohio State University, USA.

Also read JMI gets Rs 1.38 crore research grant for skill development of Assam tea tribes

IIT Guwahati clinical research

The research team at IIT Guwahati plans to get clinical validation by analysing more individuals for further air-articulation research. This will help the model to recognise day-to-day words and phrases for people with voice disabilities.

The IIT Guwahati research has potential beyond voice recognition uses. It can be used for sensing exercise-tracking and body movement-detection. Furthermore, its stability during long-term underwater use suggests future uses in underwater communication practices.

Also read NIT Rourkela, UAE University research improves Mars weather prediction using 20 years of data

The research findings have been published in the Advanced Functional Materials journal. The paper is co-authored by Uttam Manna, department of chemistry, and Roy P Paily, department of electronics and electrical engineering, along with research scholars Debasmita Sarkar, Rajan Singh, Anirban Phukan, and Priyam Mondal. Xiaoguang Wang and Ufuoma I Kara from the Ohio State University, USA, are also contributors.

Speaking about the model, Manna said, “It is one of rare design of material allowing to recognise voice based on monitoring the water wave formed at air or water interface because of exhaling air from mouth.”