Empowering Learning: A Multimodal LMS for Inclusive Education of Disabled Students

Abstract

his thesis introduces a Student and Faculty Management System with AI support and accessibility as the priority that automates and simplifies the academic processes of dis abled students. The platform combines all-time voice input, Blind Mode navigation, AI-controlled assistant support, and student/teacher messaging. It addresses the limitations of traditional LMS systems, including reliance on visual interfaces, manual data input, and delayed feedback, which hinder self-directed learning for visually impaired students. The system leverages offline speech recognition (Vosk) and Web Speech APIs to allow users to navigate, dictate, and confirm actions hands-free; utilizes audio emotion-aware prompts and blind-mode safety controls to provide confidence-level feedback; employs an AI chatbot (Gemini API) to generate context-sensitive help, interview-like practice questions, and onboarding instructions; and offers predictive analytics dashboards for faculty to make data-driven decisions. The AI voice stack is highly reliable: uninterrupted listening achieves over 98% command recognition accuracy in Chrome with automatic recovery, while audio announcements and notifications ensure real-time information accessibility. By integrating machine learning, natural language interfaces, and multimodal feedback, this LMS advances AI accessibility by automating the end-to-end academic process, empowering students and faculty with disabilities to work independently and inclusively