An expandable voice user interface as lab assistant based on an improved version of Google's speech recognition

Area of Science:

• Laboratory Automation
• Human-Computer Interaction
• Bioinformatics

Background:

• Voice assistants offer potential benefits in scientific laboratories but face challenges with specialized terminology and data security.
• Existing voice technologies often struggle with the nuanced language of scientific research.

Purpose of the Study:

• To develop Rainbow, an open-source voice user interface (VUI) tailored for scientific laboratories.
• To enhance speech recognition accuracy for laboratory-specific tasks and general computer operations.

Main Methods:

• Developed Rainbow using AutoIt scripting language to control Google Translate Site (GTS) for voice input/output.
• Integrated general Microsoft Windows tasks, lab-specific functions, and device-specific controls.
• Implemented an adaptable VUI for Windows PCs with internet access.

Main Results:

• Achieved a speech recognition accuracy of 91.3% with Rainbow, compared to 85.1% with GTS alone.
• Further improved accuracy to 98.6% in a test setup through system optimization.
• Demonstrated Rainbow's capability to handle diverse tasks, from general computing to specific lab equipment control.

Conclusions:

• Rainbow offers a user-friendly, open-source solution for implementing voice control in scientific labs.
• The system's architecture allows for continuous improvement and customization for individual laboratory needs.
• Rainbow enhances efficiency and accuracy by enabling scientists to use highly specific terminology and commands.