Development of a Multilingual Voice-Controlled Smart Wheelchair with Advanced Features
Locomotion is a basic necessity for every individual. The weak, the elderly, and the specially abled encounter numerous challenges while commuting short distances, let alone long distances. Thus, they require mobility aids like wheelchairs. Since most of the wheelchairs are mechanical, it puts great toil on them to roll the wheels by themselves. Over the past few years, some electrical wheelchairs were either built from scratch or modified from the conventional mechanical ones to make the movement effortless which were either remote-controlled, which burdened the user with the need to learn how to use it, or voice-controlled, which lacked security features like the user-specific voice command input. Moreover, the commercially produced voice-controlled wheelchairs are for only English-speaking persons, expensive and lack functionalities and safety features like obstacle avoidance, emergency braking, and distress calls. The resolve of this research is to modify a conventional mechanical wheelchair into an electrically powered and user-specific voice-controlled wheelchair with the necessary safety features, which will be a more cost-effective and technologically sound alternative to the prevailing solutions making its users independent. Besides being user-specific, the developed voice-control system is language modular and accent-independent. Neglecting the computationally expensive deep learning models, a minimal and cost-effective voice recognition module which uses MFCC based feature extraction and GMM classifier is chosen. As the mother tongue of the developers of this product is Bengali, so it was decided to program to recognize Bengali voice commands. The proposed system can recognize Bengali voice commands with an overall average accuracy of 91.2% in a closed room over different commands in Bengali and function accordingly. The users of this wheelchair can be monolingual and do not necessarily have to learn another language to operate it.
Authors
- Shaikh Anowarul Fattah (ORCID: https://orcid.org/0000-0001-8090-2327)
- Kamel M. Hassan
- Sadad Hasan
- Shourov Joarder
- Al Nayem
Institutions
- Bangladesh University of Engineering and Technology (BD)
Publication Details
- Published
- 2025-02-13
- DOI
- https://doi.org/10.1109/ecce64574.2025.11013935
- Citations
- 1
- Primary Topic
- Mobile and Web Applications
- Type
- article
- Field-Weighted Citation Impact
- 3.25