Speech-Driven Hybrid Assistive Application for Motor-Impaired Students: Embedding Agile-DBR Framework

Abstract

Despite the rapid evolution of digital pedagogy, Students with Upper Limb Motor Impairments (USULMI) continue to encounter difficulties accessing Educational Support Tools (ESTs), particularly WhatsApp and Google Meet. This study introduces Ewoxa Meet, a Windows-based, speech-controlled hybrid application designed to enable independent, multilingual interaction with these tools under variable connectivity conditions. In this study, we adopted an enhancement-driven methodological paradigm that embeds a Design-Based Research (DBR) approach with Agile methodology under a consolidated framework. We conducted performance tests by executing 120 speech commands across three internet scenarios, alongside usability testing with 17 university students via the System Usability Scale (SUS). The developers’ assessment confirmed technical stability across diverse usage conditions, with no crashes or forced terminations observed, achieving a near-optimal successful execution rate of 98% for speech commands in target ESTs. The results indicated an overall mean reaction time of 1.92 ± 0.69 s, with stable internet achieving an average of 1.39 ± 0.27 s, weak connectivity extending to 2.74 ± 0.39 s, and offline mode yielding 1.64 ± 0.41s. Field usability testing produced a mean SUS score of 79.41 with a Cronbach’s Alpha (α = 0.821), indicating high reliability and strong user satisfaction. Ewoxa Meet consequently illustrates resilient accessibility through multilingual speech control, adaptive connectivity, and a cognitively transparent interface. Moreover, these integrated methodologies scale the development of inclusive Assistive Technologies (ATs). Future work will endeavor to integrate more ESTs and investigate long-term implementation in diverse educational contexts.