We all know how fast a room can go from tidy to messy, especially if children toys are involved. What if there was a robot smart enough to sort things up, so you never step on a LEGO or come back to a messy room again?

While not available now, the CEATEC 2018 demo from Preferred Networks shows that this may not be far off. Using an existing development HSR Robot by Toyota, Preferred Network has written all the A.I software that powers the tidying robot (A.I = artificial intelligence). The machine is capable of moving around, looking for things that are out of place, pick them up and putting them in a specific bin. In theory, it could sort various objects and put them in different places.

We’ve included some official video footage below to show the robot in action. This is essentially the same demo that we saw at CEATEC this week. The robot uses a proprietary image recognition technique, powered by Preferred Networks’ deep learning research. Few people have heard of this company, but they recently were ranked #2 out of 450+ contestants in the Google AI Open Images competition.

You will also notice that the Robot’s speed is currently a bit underwhelming, and it is one of the things that makes the whole idea still a bit remote from commercial availability, even though the concept is relatively proven now.

It turns out that this is the maximum speed that the robot can physically perform at, and not a reflection of the software’s performance. These robots are the fruit of Toyota’s research and are currently only available to developers. Once specific use cases have been identified, many aspects can be optimized. Preferred Network chose this robot because it could move its body and its arm at the same time, resulting in a net saving.

In any case, we hope to see rapid progress because there is certainly a demand for this kind of products. I would certainly pay (a reasonable price) for such a functionality!

Filed in Robots. Read more about , , and .

Discover more from Ubergizmo

Subscribe now to keep reading and get access to the full archive.

Continue reading