Speech‑Driven Execution of Windows Operating System Commands for Users with Motor Impairments: A Sphinx‑4–Prototype

Abstract

This paper describes a practical speech-driven framework that lets users with motor impairments run Windows operating system commands by speaking them. The system uses the open-source Sphinx-4 engine and Java Speech Grammar Format (JSGF) to match spoken phrases to actions that have already been set up in a command dictionary. The prototype does basic things like opening folders, starting programs, and changing settings. Tests in different levels of background noise showed that the response accuracy was about 90% in quiet settings and 82% in moderate noise. All of the functions that were tested showed that the system could respond to commands in less than one second. These results show that grammar-based speech control is a lightweight and easy-to-use alternative to traditional input devices.