AN EEG-BASED TECHNIQUE FOR OPERATOR FATIGUE DETERMINATION

Operator fatigue is responsible for significant injuries and losses to the Australian mining industry each year. In the absence of appropriate fatigue strategies and technologies, this problem will only be exacerbated by the combination of an aging workforce and ever-increasing demands on operators to achieve production quota. A number of commercial products have been developed over the past decade to tackle the operator fatigue problem. The majority of these uses either eye/head behaviour or operator response time to determine the level of fatigue. A limitation common to these technologies is the inability to cope with driver-to-driver variations, which results from the fact that they infer fatigue levels based on individuals ’ response/behaviour rather than the direct, physiological measurement of fatigue. This paper describes an EEG-based technique for the detection of physiological fatigue in haul truck drivers, which utilises a neural network classifier trained with data collected from multiple operators. Comparisons are made with more popular techniques for fatigue detection to highlight the universal nature of this approach, using data acquired from both laboratory and in-pit trials. A discussion addressing potential countermeasures and fatigue management strategies is also included