Researchers at Stanford and Google's DeepMind have presented an open-source housekeeping robot and trained it to fry shrimp, rinse pans, place pots in kitchen cupboards, and clean up spilled wine relatively quickly. But this robot's ambitions are bigger.
The Mobile ALOHA platform is far from a sexy-looking machine. It has a flat, wheeled base with a 12-hour battery pack that supports the brutally ugly scaffolding, a laptop, a top camera, and a pair of clawed robotic arms each with a wrist camera and 14 degrees of freedom.
This aesthetic disaster can be fitted with a detachable training setup, allowing operators to train the bot by pushing it and manually operating its arms and claws.
This 75 kg (165 lb) robot is completely open source. The project team will provide you with a parts list and guidance on how to build the parts yourself. It's a relatively inexpensive build considering its features, but that doesn't mean it's cheap. All you need is a hot glue gun and a 3D printer, and you can get the parts for a little over $30,000.
The ALOHA system can then train you by reinforcing the skills you already know from previous training. The team said that executing a task 50 times increases the likelihood of the robot performing the task autonomously by up to 90%, and released a video showing some of the manual skills the robot has acquired.
Mobile ALOHA – Smart home robot – autonomous technology intensive
But as you can see in this second video, the platform can do much more. None of these videos are autonomous. The bot is operated remotely. But it does give us some insight into what jobs we can expect household robots to take on in the not-too-distant future.
Mobile ALOHA: Your Lyrics Robot
And for added realism, here's a third video that shows why Mobile ALOHA is still far from a commercial product. You don't want to handle fine china just yet because you'll still make a lot of dumb mistakes.
Mobile ALOHA Funny Failure
This may just be a student-led research project, but it is part of a growing wave of mobile robots that are accelerating their ability to see, learn, and autonomously repeat tasks in diverse and dynamic real-world situations.
Suddenly it seems like the end of household chores may not be far away. I can't say I'll miss it.
Source: Mobile ALOHA