Everyday Robot Is Not Your Everyday Robot

A robot that can assist people with everyday, mundane tasks — it sounds obvious, right? It’s a practical solution that could improve anyone’s life in any number of ways. You need help carrying in the groceries, doing the dishes, or taking out the trash? You want to rearrange your living room furniture, pack for a vacation, or change a light bulb? The Everyday Robot project aims to create a robot that can help with all of these things and more.

The only problem is that designing a handy-dandy general purpose robot is a monumental challenge.

The problem

Right now we have to give robots a clearly defined set of rules. They require specifically programmed instructions and a predictable setting. We’re seeing an increasing number of robots move beyond research labs and safety barriers, but our most useful robots still require a controlled environment to thrive.

Things that are simple for most people to do — learning, adjusting, adapting — are really difficult to program. Robots currently struggle to operate in unstructured environments. There are too many variables; people and objects are constantly in new and unexpected places.

It’s going to take a really sophisticated robot to carry out a really simple task in the real world, or rather a lot of really simple tasks.

The solution

The Everyday Robot project from Alphabet’s X moonshot division (which used to go by Google X) aims to build a mobile learning robot that can assist everyone with everyday tasks.

Just safely navigating in an uncontrolled environment is a challenge for autonomous robots. Even picking something up in an uncontrolled environment is a difficult feat for a robot.

These everyday robots must be able to understand how to navigate and interact with our everyday settings. This requires more advanced machine learning technology than we currently have available to us. According to X, these robots will have to surpass autonomous vehicles in ability to learn as well as ability to operate in unstructured environments.

Will it work?

Ultimately, Everyday Robot hopes their robots will be able to learn from human demonstration, demonstrations by other robots, simulations, and from experiences to improve the lives of people all over the world.

Currently, these robots are sorting recycling — something that humans are surprisingly bad at. They’re slow, but they’re accurate. According to X, their robots make sorting errors with less than 5% of recycling ending up in the wrong place. People get it wrong 20% of the time, however.

The need, the research, the funding, and the desire are all there to make the everyday robot a reality. It will be a few years before we will get a sense of whether or not the dream can become a reality.

This is a lofty goal fitting for “The Moonshot Factory”. It will be exciting to monitor the progress as these robots improve. Maybe one day we will have truly useful everyday robots.

Even if that day comes, your old industrial robots will still have a place in your factory. Make sure that your system keeps running smoothly now and in the future. Call 479-422-0390, or contact us online, for any of your Indramat maintenance, support, or repair needs.