VANS — Voice-Activated Navigation System
Hands-free Chrome extension translating speech into browser actions — empirically benchmarked against keyboard/mouse.
Overview
Hands-free Chrome extension translating speech into browser actions — scroll, zoom, search. Empirically benchmarked against keyboard/mouse in a 10-task within-subjects study measuring completion time, error rate, and satisfaction.
The Problem
Users with motor impairments or situational disabilities (driving, cooking) cannot efficiently navigate the web. Existing voice solutions are limited to dictation — VANS maps speech commands to browser actions.
Questions Addressed
- 01
Can voice commands match keyboard/mouse efficiency for common browser tasks?
- 02
What is the error rate of Web Speech API recognition in real-world conditions?
Methodology
Chrome Extension Development
Built a Chrome extension using the Web Speech API to recognize voice commands and translate them into browser actions: scroll up/down, zoom in/out, search, navigate back/forward, and click links by number.
Empirical User Study
Designed a 10-task within-subjects study comparing VANS voice commands against traditional keyboard/mouse. Measured task completion time, error rate, and user satisfaction (SUS scale). Recruited 20+ participants.
Analysis & Findings
Analyzed results using paired t-tests. Found voice commands were 15% slower on average but reduced physical effort significantly. Error rate was comparable for simple tasks but higher for complex navigation.
Key Results
Key Findings
Voice commands were 15% slower than keyboard/mouse on average but rated significantly higher on satisfaction for accessibility use cases.
The Web Speech API achieved 92% recognition accuracy in quiet environments but dropped to 78% with background noise.
Participants preferred voice for simple commands (scroll, search) but keyboard for precise actions (small link clicks).
Conclusion
VANS demonstrates that voice-based browser navigation is viable as an accessibility tool today. While not yet faster than traditional input, it fills a critical gap for users who cannot use a keyboard or mouse.