Train Depot Safety with iVoice: The Future of Radio Communication Audits

In the fast-paced environment of train depot operations, precision in communication is crucial to ensuring safety, especially during manual train driving. Radio Communication Voice Procedure plays a vital role in controlling train movements within depots, preventing accidents, and maintaining operational safety. However, traditional methods of transcribing these communications are time-consuming and prone to human error.

This is where iVoice, an innovative, AI-enabled system is transforming how SMRT is auditing and managing these critical communications.

iVoice is a state-of-the-art solution designed to automate the transcribing of Train Captain (TC) Radio Communication Voice Procedures. Unlike manual transcription methods that can be inefficient and error-prone, iVoice provides a more accurate and streamlined approach. It allows for 100% audits of Train Captains' (TCs) communications, flagging readily any non-compliance of the TC, ensuring that they will receive additional training, thereby ensuring adherence to standards for safe train driving operations.

Technologies behind iVoice

At the core of iVoice are two key technologies:

• Automatic Speech Recognition (ASR): This feature transcribes spoken communication into text with high precision, eliminating the delays and errors associated with manual transcription.
• Natural Language Processing (NLP): After transcription, NLP analyses the text's syntax and semantics to ensure it adheres to standard voice procedures. Any deviations are immediately flagged in a report for supervisors to review and address.

Key Benefits of iVoice

iVoice offers significant advantages over traditional auditing methods, including:

• Enhanced Productivity: iVoice increases productivity by over 60% with the automation of the transcription process, allowing 100% of the audits to be completed in just 50% of the usual time.
• Comprehensive Audits: The system ensures that 100% of all Train Captains' communications are audited, providing a complete picture of compliance with safety standards.
• Improved Compliance: With faster detection of non-compliance, i iVoice enables quicker interventions, reducing the risk of errors in train operations and enhancing overall safety.

iVoice was conceptualised by SMRT and the Singapore Institute of Technology (SIT), leveraging SMRT's problem statement and SIT's operational technology know-how.

By combining with the multi-domain Automatic Speech Recognition (ASR) and Natural Language Processing (NLP) expertise of I²R, this collaborative effort led to the development of iVoice, aimed at enhancing productivity and improving safety compliance in the rail transport sector.

If you are interested in collaborating with us on this or delve deeper into the research of iVoice, contact us here.