It’s no secret that warehouse work can be dangerous. Critics have long dogged Amazon for its warehouse practices, citing reports of inadequate bathroom breaks, long shifts, miles of walking, and injuries resulting from lifting up to 100 items an hour.
The public scrutiny has been so intense that Amazon enlisted “FC Ambassadors” to tweet positively about their experiences working in fulfillment centers — a program that has arguably backfired. On Amazon’s Prime Day this year, workers at a center in Shakopee, Minn. were able to garner significant attention by organizing a strike.
Now, researchers at the University of Washington are stepping in to solve one of the warehouse industry’s problem areas: ergonomics. The team aims to make warehouse work safer by developing a system that can warn people when they’re at risk during physical tasks.
To get started, the researchers built a dataset from videos of common warehouse activities like picking up and moving boxes. The researchers captured movements with Microsoft Kinect, a sensor built for developers to use with internet-of-things applications. Individual actions were then assigned a “risk score” based on the impact on a person’s joints. Risky behaviors included picking items from or placing them on a high or low shelf.
The idea was to identify and flag risky movements in order to prevent future injuries. The researchers plan to turn the system into an app, which would give workers the power to monitor themselves and catch their own risky behavior with a warning system.
“The Kinect is essentially able to track a 3D skeletal model,” said Ashis Banerjee, a senior author on the paper and an assistant professor at UW, in an announcement. “We can then compute how the joint angles in my body are also changing as I’m moving objects around.”
Any reduction in worker injury could have a large impact. The warehousing and storage industry added an average of more than 4,000 positions each month over the past year, according to the Bureau of Labor Statistics.
After using the app to flag risky work, Banerjee said the next step is to hand those tasks to robots. “We have a unique opportunity to split up the work so that the robots are doing the risky jobs.”
The project received funding from the State of Washington as well as a gift from Amazon Robotics. The team plans to present the findings at the IEEE International Conference on Automation Science and Engineering in Vancouver, B.C. on Aug. 23.