In 1986, the inaugural self-service checkout was introduced at a Kroger store near Atlanta, marking the beginning of a slow yet steady adoption of the technology throughout the United States. As grocery shopping evolves with more automation, the advent of robotic bagging assistants appears to be on the horizon.
This week, MIT’s CSAIL department is unveiling RoboGrocery, a project that marries computer vision with a delicate robotic arm to effectively bag a myriad of items. In a trial, the robot was presented with 10 items on a conveyer belt, all unknown to it beforehand.
The assortment of items included fragile goods such as grapes, bread, kale, muffins, and crackers, alongside sturdier items like cans of soup, boxed meals, and ice cream containers. Initially, the computer vision system identifies and assesses the objects, figuring out their size and where they lie on the belt.
Upon gently grabbing the grapes, the robot’s pressure-sensitive fingers recognize their fragility, wisely deciding against placing them at the bottom of the shopping bag—a lesson many shoppers learn through less technological means. It then discerns the soup can’s solidity, deeming it suitable for the bag’s base.
“Robots packing groceries and similar tasks in everyday environments mark a notable advancement,” commented Annan Zhang, a leading researcher of the study. “While the technology isn’t ready for market just yet, our findings highlight the capability enhancements possible through the integration of sophisticated sensing technologies in soft robotics,” Zhang explains.
The researchers acknowledge that there is significant potential for refining the technology, citing possible improvements to both the gripping mechanism and the vision system to optimize packing strategy. Moreover, as the technology matures, its application could extend beyond supermarkets, potentially benefiting sectors such as recycling management.
Compiled by Techarena.au.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
