UPS took a data-driven approach to introduce machine vision solutions to the problem of jams in its automated facilities, its enterprise director of advanced analytics explained in a presentation this week.
Mohammed Chaara told the audience at the AI Summit in London that the company had generated a solution for the recurrent problem through its “analytics innovation theatre” in which anyone in the organisation can contribute ideas.
Conveyor jams, where a number of parcels build up and are unable to flow as normal down the belt, usually require human intervention to clear the affected area and can cause damage to a conveyor if the belt is not stopped quickly enough. With UPS having revealed plans for 80% of packages to go through automated facilities this year, the impact on productivity is meaningful.
With the goal of predicting jams earlier in order to minimise disruption, the analytics innovation theatre came up with three possible ways to use video to detect potential jams.
One solution was automatically detecting boxes as they moved along the belt, which ran into problems when a high concentration of parcels in one place obscured the others.
Another was analysing the amount of free space on a conveyor, called belt mask detection, with the assumption that a minimal amount of free space on a belt meant a high likelihood of a jam. However, this was unhelpful when there was a continuous flow of parcels.
The third concept was analysing the movement of belts to see whether it was in motion or not, titled package movement flow indicator.
After extensive testing of each option, UPS settled on a combination of belt mask detection and package movement flow, combining this into a minimum viable product.
Chaara confirmed to eDelivery after the talk that the impact of the technology had been “significant” though he could not share specific figures on jam reduction.
He also noted that the primary role of the technology is in freeing up the staff who would have had to clear the conveyor, making better use of existing capacity rather than adding new capacity.
Chaara explained the company’s attitude to innovation, saying that it encouraged employees to question everything.
“In God we trust – everyone else needs to bring data.”
Another principle was that employees should “dream big but in small steps”.
“Creative or sci-fi thinking is allowed, we can take on any challenge but eventually you have to break it down into realistic steps.”