Self-Learning Robots Teach Each Other in the Cloud

A new startup is receiving $8 million in funding from investors including Andy Rubin, co-founder of Android and the former head of Google’s robotics division.

It’s called RightHand Robotics out of Somerville, Massachusetts.  Their new robot with it’s head in the cloud is called RightPick. They say it can pick up and sort small objects, under five pounds, 500 to 1,000 times an hour. The bot uses machine learning to figure out how to handle various items on the fly.

The robotic arm has a fingered hand with a suction cup in the middle. An embedded camera helps the hand figure out which appendage  to use and how to grasp the item.

Once it’s learned to handle objects such as food, shaving cream, toothbrushes etc., it can share it’s newfound knowledge to it’s hive-mind in the cloud.

If it can’t figure it out an engineer logs into the system remotely and solves the problem, or to help train the robot to pick a new object.

The collaborative nature of this AI embeded robot has  given rise to a new breed of robot called a co-bot.

It may seem simple for a human to do this sort of work, but for a bot, this is really hard.

Where this robot will have the biggest impact is in warehouses, picking and sorting inexpensive items for shipping. It will lower labour costs, and perform faster than it’s human counterpart.

Leif Jentoft, co-founder RightHand, says they already have customers, and are getting traction in apparel, consumer packaged goods, cosmetics, electronics and grocery.

“It’s a nice combination of a new technology that has some transformative capabilities and a big need in the market,” he said.

Currently focused on selling robots to ecommerce and brick-and-mortar retailers, they also plan to expand to manufacturing, military and home automation.

The company currently has 20 employees, but the new financing round will help expand product development, hiring and marketing.

Advances in robotic software and hardware, combined with computer vision, and teleoperation with the ability to learn and collaborate over the cloud, will transform warehouse fulfillment forever.

Best of all, the bots are operating with humans in the loop, so we don’t have to worry about them building a time machine to go back to 1984, looking for Sarah Connor.

Thank you for reading.

Subscribe to iTunes Podcast, Stitcher, or Google Play to get future episodes.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.