Predicting Fatigue with trained AI Voice Analysis

Providing Fatigue Risk Management Solutions to companies worldwide by predicting fatigue ahead of time

Why WOMBATT-VOZ?

The system uses advanced AI-driven voice analysis to predict fatigue up to five hours in advance. By monitoring subtle changes in speech patterns, the system detects early signs of fatigue and provides timely alerts to both the individual and management. This predictive capability enables proactive intervention helping to prevent accidents and enhance safety, helping organizations to create safer, more productive work environments by addresing fatigue before it becomes a risk.

Predictive

Knowing for experience that Prevention > Reaction, our system allows to manage fatigue before it becomes a risk.

No Hardware

It's an APP, for the users, and it can be easilly installed on any mobile device, no hardware installation/maintennance needed.

High Accuracy

Uses voice-based Pre-Trained AI to predict fatigue with up to 90% accuracy with constant use.

About Us

Founded in Adelaide, Autralia in 2000, WOMBATT Fatigue Management originated as a spinout from the University of South Autralia's Institute for Telecommunications Resarch. We introduced a breakthrough satellite communications technology that connected mining haul trucks in remote locations to central control centres.

In 2009, WOMBATT joined the European Space Agency's Business Incubation Centre to develop our patented EMDP technology, enabling "real-time" remote monitoring or driver fatigue. Our first deployment was with STRACON Ltd at the El Brocal mine in Perú.

In 2014, ESA engaged WOMBATT to help develop a voice-based fatigue predition system for astronauts on long-duration space misions. This led to a prime contract under ESA's ARTES 20 programme, with the technology succesfully passing Factory Acceptance Testing in 2020.

Today, WOMBATT provides AI-powered fatigue management solutions for industries including mining, transport, aviation, steel and civil engineering-driven by continued innovation and support from ESA.

"Managing Fatigue: It's about Sleep" - Prof. Dawson


Fatigue as a result of lack of sleep plays a major role in our society, and with the increasing pressure from social media, such as binge watching our favourite tv shows until the early hours, sleep is under pressure these days not only from our working and commuting hours, but also from sleep time having to make room for our smart device use as well.

At WOMBATT our vision is to teach people about fatigue and sleep, helping them gain insights into the dangers of fatigue in every industry. We strive for a safe work environment and safe road use in which fatigue is no longer the invisible threat to people and companies. Smart devices causing distractions from sleep can now be used to turn the tables again, helping protect us from sleep loss and fatigue and with that improving health, well being and safety!

An essential tool in managing fatigue effectively is having the right data to work with.

Carlos Ramirez

Site Manager - Monterey Steel Mill, México, TMS International

We are often unaware of our own fatigue and unable to measure it, yet it clearly impacts our health. At first, I was sceptical about how my voice could reveal my fatigue level. However, after using WOMBATT daily and adjusting my rest and family routines, I noticed significant improvements in my health, work performance, and relationships at home and at work. WOMBATT is simple to implement, your voice is all it takes. Since each person’s voice is unique and cannot be altered or faked, it offers a reliable measure of fatigue. The process is easy, you read a 15-second script at the start of your shift, after meals, and at the end of your day. In just 60 seconds, you gain valuable insight into your fatigue level. This system empowers you to make life changes that support optimal health and rest. From an operational standpoint, I observed a marked improvement in mobile equipment operators, they were more alert and focused on their tasks. Initially, safety incidents occurred daily. Over time, they dropped to weekly, then monthly, and eventually every few months, with a noticeable decrease in severity. The union appreciated that we prioritised employee well-being. Since the method is non-invasive, operators quickly embraced it. In fact, they began requesting their WOMBATT voice checks to ensure they were fit to work. In my view, WOMBATT is a simple yet powerful solution to a complex issue. Fatigue is invisible—but its consequences are not.

Mine Operations Manager

North American Gold Mine

I believe it is the only system on the market that is a predictive tool for identifying fatigue before it leads to an incident. Which then teaches the user to become the predictor and ultimately make changes to their lifestyle and create better habits to prevent fatigue events. Once this happens, you are then able to optimize time and resources to work with leading indicators vs. time and resources chasing lagging indicators and not solving the root problem. The root problem in my opinion is operators not fully understanding what is inhibiting proper rest or how much they truly need, or partially a suboptimal schedule that could be changed.

Kirk Zerkel

P.E. | President, AAP, LLC

My name is Kirk Zerkel, I am the owner of a Contract Mining company based in Alaska, USA. We operate fleets of large heavy civil equipment, including large mining trucks in surface mines in the Northwest States of the US. The nature of our work is that we have operators of large mine trucks/equipment, working long hours throughout the year. Winters in Alaska are long, dark and cold, and our operators spend up to 12 hours each shift operating in dark conditions on slick roads. Fatigue is a major factor for us. The WOMBATT system has significantly improved our ability to monitor our drivers and keep them safe. As a proactive vs reactive system, we now have the ability to identify individuals who may be starting to feel fatigued, before an incident can occur. I highly suggest the use of WOMBATT-VOZ.

Dan Wood

Managing Director, BE DHL Oman

We are particularly excited by the potential of the WOMBATT forward-looking AI fatigue prediction technology with our owner operator fleet of drivers. With over 450 drivers covering in excess of 20 million kilometres per year, often in harsh and unrelenting landscapes and geography, the detection of fatigue proactively is a necessity. We look forward to a successful conclusion of the pilot, under the guidance of our assurance and innovation lead Steven McGarry and a subsequent roll out to other areas of our business.

Frank M. Salzgeber

Head of Innovation and Ventures Office, TIA - AI Downstream Business Apllications Department, Directorate of Telecommunications and Integrated Applications, ESA

Astronaut fatigue has been found to be one of the key risk factors for long duration manned missions in Space. It has been known for some time that fatigue can be very accurately detected from computer analysis of a short voice recording, so after examining all other fatigue detection technologies the European Space Agency (ESA) funded a feasibility study funded called iVOICE. The iVOICE voice project based fatigue detection technology for use in long range manned missions to Mars and other planets. Under the Downstream Business Applications programme (Artes 20 programme) ESA contracted WOMBATT Fatigue Management Ltd (UK) to commercialise the technology for use on Earth. ESA has conducted a comprehensive Factory Acceptance Test (FAT) for WOMBATTs commercial version of the iVOICE fatigue technology. The WOMBATT iVOICE system passed the ESA FAT test and is currently being deployed commercially at mining sites and in airport air traffic control centres around the world.

Steve Dixon

Chief Executive Officer, STRACON S.

Operator fatigue is a key business risk in the mining industry, and the industry is constantly searching for more accurate, economical and unobtrusive ways to detect fatigue. We believe that the new voice-based fatigue detection technology being developed and commercialized by WOMBATT has significant potential in the mining industry to help solve the operator fatigue problem. We are interested to examine the new system when it becomes commercially available, for use at our mining projects in Peru and elsewhere. Yours faithfully STRACON S.A.

Smart Fatigue Prediction System

Always Predictive

By analysing the voice for specific fatigue elements and changes in an individual’s voice patterns, the system can identify signs of fatigue and provide the early warning to the individual and the organisation’s management team.


This predictive characteristic can help prevent accidents and improve worker and driver safety by allowing intervention time. This can be a powernap or swapping out to less dangerous jobs preventing fatigue from taking control of the situation.

Non Intrusive

As users have agency to perform their voice samples, no camera monitoring of the users or any weareable is neccesary.


  • Easy and fast setup for the User App.
  • No union problems, worker just follow the protocols.
  • Voice Verification, rejects the sample if something is wrong.
  • Does not need special equipment.
  • Requires minimal user training.
  • Encourages changes in user lifestyles.

High Accuracy

WOMBATT's system includes a learning algorithm that continuosly adapts to each user. With consistent use and adherence to protocol, it builds a personalized voice model for more accurate fatigue risk preddiction. Over time, the system updates this model using recent voice samples, ensuring it evolves alongside the user.


This adaptive apporach enables up to 90% accuracy in predicting fatigue risk during work hours. Built-in quality filters and voice verification safeguards ensure the model remains clean and reliable, free from distorded or incorrect data.

Voice Verification

Our system includes a built-in voice verification feature to ensure that each voice sample genuinely belongs to the intended user. Is a sample shows significant differences in vocal characteristics, is heavily distorted, or contains only noise, it is automatically rejected and excluded from analysis.


In such cases, supervisors are notified to maintain data accuracy and monitoring reliability

Mobile and Hardware Free

The WOMBATT user interface is delivered through a mobile application compatible with Android, iOS and laptops. AS the front-end of a larger system, the app ensures accesibility across all areas of the organization, allowing personnel in critical environments to interact with the system without the need to dedicated hardware.


This mobile apporach eliminates installation and maintenance costs, prevents downtime of vehicles or machinery, and removes the risk of hardware damage-making it a flexible and cost-effective solution for real-world operations.

Smart Sleep Data

By integrating smartwatch data, WOMBATT provides deeper insight into your actual sleep wuality-not just your sleep schedule.


Many believe they get enough rest, but sleep debt often builds unnoticed. Linking sleep data helps identify what's truly affecting your fatigue, raising awareness and supporting better recovery through informed sleep habits.

Frequently Asked Questions

How does Voice Fatigue Technology works?

  • Lack of restful sleep creates sleep debt that builds up over time, and this sleep debt can lead to future episodes of daytime sleepiness (microsleeps), which is when the brain needs to shut down to recharge. These microsleep episodes can occur at any time without any warning.
  • Before these episodes occur, the body feels the sleep debt at a muscular level, and because muscles produce the voice, this will produce physiological variations to our normal voice patterns.
  • Using Trained Artificial Intelligence, our algorithm analyses these individual voice patterns to detect those variations by comparing them against what it has learned from our previous recordings, creating a moving base model (using the last 200 recordings) and comparing each new voice recording against that model.

Why does it needs three recordings per day/shift?

  • The algorithm needs at least three voice samples per day to create a model that reflects our fatigue curve during the workday. While it is normal to feel progressively more tired as the day advances, it is not normal for this fatigue to exceed our normal values due to sleep debt as measured against the model.
  • Then it will use the created model to learn how our fatigue is reflected in the voice at different times of the day, using the three minimum required recordings, at the beginning of the day, in the middle and at the end of the day.
  • Once this daily model is created, the algorithm will create an intelligent model using the last 200 recordings and will compare the new voice sample against this model, to predict our risk level at a certain time of day.
  • Not making three recordings per day will result in an imperfect model and the AI will not be able to predict our risk levels effectively.

What happens if I make more than 3 recordings per day/shift?

  • If we make more than three recordings per day, the result will be that the algorithm will be better able to determine the risk level over a wider range of hours during the day. More information is always better.
  • But, making consecutive recordings will not result in an improvement in risk prediction, since if, let's say for the sake of compliance, I make three recordings in two hours, the base model described above will not be effective in predicting during the rest of the hours of the day.
  • So, the recommendation to get the best result is to make at least three recordings during the day, but at the beginning, middle and end of the day. Any recordings in between those three will only improve the model.

Why is the system giving me  HIGH RISK  results if I'm not feeling sleepy?

  • Because the algorithm works predictively, creating data models based on what it knows about us, it is common to think that it is not giving a good result if we do not feel sleepy at the time of making our voice recording.
  • The main objective of the system is to prevent rather than react, by building prediction models based on our previous information, it can prevent the RISK level of having a drowsiness or microsleep episode in the next 4-5 hours, with an efficiency level of up to 90%, provided it personally knows us well enough to reach those levels.
  • The level of effectiveness in prediction always depends on the user, and although it will never be 100%, up to 90% efficiency can be achieved if we make our recordings three times a day as explained above this lines.
  • In any case, the system predicts the level of RISK, which does not necessarily mean that we will fall asleep in the following hours, but it allows us to work in a proactive manner to minimize the risk of having an accident due to drowsiness.

What kind of information could I obtain from the system?

  • The system can send different types of alerts, based on the different results:  RED ,  ORANGE ,  GREEN , for each result for each user.
    • Direct APP result of voice analysis, returned as a result of the sample sent to the system for fatigue risk preddiction.
    • Email alerts, sent to the supervisors and app-users emails.
    • Failed voice analysis, sent to the supervisors emails for further intervention and user training.
    • In-app notifications, sent through Google's FCM service. Prior supervisor registration is required.
    • PUSH notifications to external systems, notification configuration and development of endpoints in the external system is required.
  •  
  • Additionally, the system has APIs that allow retrieval of user data directly from other systems.

Contact Us

We're here to support you with our extensive experience in fatigue and our cutting edge fatigue prediction technology.

Feel free to reach out to us with your questions regarding fatige risk management training and we'll get back to you!

Jean Verhardt

CEO and Founder

Email Us

info@wombatt.net

Loading
Your message has been sent. Thank you!