Unearthed Sydney 201812 Oct - 14 Oct
Bringing in fresh & exciting ideas to industry for real rewards
This competition has finished.
We used machine learning to train a realistic model of the rimpull curve based on real world data. It works on the assumption that the drivers know the machinery and the roads the best. We train on this data. Using this model we calculate time (t) to travel over a segment of length l and grade g while carrying mass m. We can calculate time as a function of mass that is tailored for the drivers, the road, the truck. As a pit is a series of segments -> we infer using the machine learning model then find the payload that maximises the throughput of the system.