The first automated teller system was installed in 1986 at a Kroger grocery store just outside of Atlanta. It’s taken several decades, but the technology is finally proliferating in the U.S. Given the automated direction grocery stores are headed, it seems robotic packaging can’t be far behind.
MIT’s CSAIL department this week presents RoboGrocery. It combines computer vision with a soft robotic gripper to package a wide range of different items. To test the system, the researchers placed 10 objects unknown to the robot on a food conveyor belt.
Products ranged from delicate items such as grapes, bread, cabbage, muffins and crackers to much more substantial items such as soup cans, lunch boxes and ice cream containers. The vision system goes in first, detecting the objects before determining their size and orientation in the zone.
As the handle touches the grapes, the pressure sensors on the fingers determine that they are, in fact, sensitive and therefore should not go to the bottom of the bag – something many of us no doubt learned the hard way. He then notes that the soup can is a more rigid structure and glues it to the bottom of the bag.
“This is an important first step toward creating robots to pack groceries and other items in real-world conditions,” said Annan Zhang, one of the study’s lead authors. “Although we are not quite ready for commercial development, our research demonstrates the power of integrating multiple sensing modes into soft robotic systems.”
The team notes that there is still a lot of room for improvement, including upgrades to the grasping and imaging system to better determine how and in what order to pack things. As the system becomes more robust, it can also be scaled out of the grocery store to more industrial sites such as recycling plants.
