Collision prediction for a mining collision avoidance system

Abstract

Accidents caused by wheeled mining machines contribute to approximately 30% of injuries and fatalities in the global mining industry. Wheeled mining machines have limited driver assist features when compared to the passenger vehicle market and are typically limited to collision avoidance by braking. These products are often subject to false positive interventions leading to production losses, increased wear, and resistance to adopt the technology by end users. This study proposes a sampling-based method to expand the collision avoidance by braking approach to include steering. The sampling method is based on the vehicle’s kinematics and the application of a Gaussian distribution to the steering rate to determine the probability of a collision occurring. Initial results indicate that the inclusion of steering rate on the collision prediction model may increase the operator’s situational awareness, leading to fewer false positives.