The fuzzy neural network based haul truck driver fatigue detection in surface mining

Abstract

Haul trucks are one of the most important transportations in the surface mining. The traffic accidents involved haul trucks result in serious damages of the vehicles, even the loss of the drivers' lives. Drive fatigue is one of the key features that cause accidents. This paper analyzed the possible driver fatigue reasons based on the special features of the haul trucks. Then, it proposed the fuzzy neural network based fatigue detection for haul truck drivers. The CCD camera was amounted in the cabin of the haul truck to capture the key features of the drivers. The eyes of the driver could show most important information of fatigue. The head nod frequency, yawn frequency could also reflect the driver's fatigue degree. Each of these methods has its own limitation and detection errors. The results shows the fuzzy neural network detection method has more accurate detection by combining with PERCLOS (Percentage of eyelid closure over the pupil over time), AECS (Average eye closure speed), NodFreq (Nod frequency) and YawnFreq (Yawn frequency). It has great significance in reducing the accidents rate caused by the drive fatigue.