Researchers have slowly been introducing robots into kitchens, aiding humans in tasks like frying chips and peeling bananas. Up until now, however, such robots were better at large-scale operations than delicate tasks like peeling bananas without crushing the fruit inside – however with some clever programming this dual-armed robot managed 57% success! This promising start could one day allow robots to self-assemble bananas!
Researchers at the University of Tokyo’s ISI Laboratory developed a machine learning system that taught a robotic arm how to gently peel bananas without squashing them – no easy feat! Researchers employed “deep imitation learning,” wherein hundreds of demonstrations of this task were shown repeatedly – this enabled it to pick up and peel one in three minutes, according to Reuters reports.
A telescopic channel fitted with two equal rollers and pins can be used to hold bananas while they’re being peeled, and its distance between pins and rollers can be automatically adjusted according to different sizes of bananas in order to minimize damage during peeling.
Two elastic springs are employed to support and absorb shocks caused by handling a banana, and increase durability and reliability of the device. They are composed of mild steel spring grade material.
Two robotic arms resembling hands rest over an unpeeled banana on a table. The left arm gently lifts it, while its right counterpart takes pieces of its peel. As they peel, turning around as they go, this robot finishes stripping it entirely.
After peeling a banana, it’s fed into the center of the machine where it will be sliced and packed for sale in bags or containers. Scientists behind the robot hope to improve its speed and accuracy so it can eventually handle other fine tasks like picking apples or packing fresh produce as well as solving labor shortages in certain industries.
While robots may seem distant now, their introduction into restaurants and cafeterias could soon be on the horizon – perhaps even serving George Michael Bluth’s Banana Stand one day! This prototype shows promise.