WOMBATT commercialises UCL’s iVOICE fatigue prediction and detection system

WOMBATT has commercialised iVOICE, a voice based driver fatigue prediction and detection system invented by the Centre for Space Medicine at University College London. A combination of voice analysis and machine learning originally developed to track the health of astronauts on long term space missions has been shown to have acquired this very human capability.

“We have all had the experience of hearing tiredness in a person’s voice. Somehow, they sound different to normal and less alert. Even a single sentence spoken over the phone can be enough for you to say ‘you sound tired, are you OK?’” The voice analysis system, iVOICE, has been developed at the Centre for Space Medicine at University College London (UCL) to gauge a person’s level of tiredness simply by listening to his or her voice for some seconds.

iVOICE was developed under contract to the European Space Agency as part of research into systems to monitor the physical and mental health of future space flight crew. However, its applications here on Earth have not gone unnoticed. “The iVOICE system works by first getting to know the voice characteristics of each individual driver so that it can identify small changes to their voice occurring over the working day which indicate increasing fatigue. This is similar to the way humans detect fatigue in the voice – we notice tiredness most easily in people well-known to us than with strangers whom we have just met for the first time.”

On published tests, iVOICE was able to detect significant fatigue with an accuracy of 90% once tuned to individual voices. Being just a software algorithm iVOICE can be installed wherever there is a capability to make voice recordings onto a computer, either in bespoke systems installed in vehicles, or via existing telecommunication channels to a system in the cloud.

For mining haul truck drivers an important feature of the new technology, apart from its accuracy, is complete unobtrusiveness. There is no need for specialist equipment such as cameras or wearable sensors, and drivers would only need to speak as they might anyway as part of their job. iVOICE will automatically recognise drivers from their voices when they speak in the cabin, perhaps in response to a simple question such as “How are you feeling? What is your current activity? Which shovel are you working from?.” From the driver’s response iVOICE may say “John, you sound very tired. You may not have received sufficient rest to drive this truck tonight. The shift supervisor will call you over the radio in a few moments to discuss.”

The technology has been tested by fatigue management company WOMBATT with mining haul truck drivers at a mine in Peru to demonstrate that the system can work in a haul truck cabin as well as it can aboard a spaceship. WOMBATT are commercialising the technology in partnership with University College London – Centre for Space Medicine and Datasat Communications Ltd.

WOMBATT states: “With fatigue implicated in up to 65% of all fatalities in the mining industry, and with rapidly increasing regulatory oversight of this problem, we expect that within a very short time, all mining trucks will be fitted with iVOICE low cost, fully unobtrusive fatigue prediction and detection as standard equipment. Indeed, when fitted to VHF radios and mobile phones, every person in the mine working in safety-critical jobs could be monitored for dangerous levels of tiredness.”