The December 2022 issue of IEEE Spectrum is here!

Close bar

MoonBot Kit Does Sophisticated Vision That's Both Easy and Cheap

This robotics kit comes with a vision sensor that promises to be simple to program with autonomous behaviors

3 min read
Moonbot robot kit
Photo: Morpx

For those of us who aren’t already robotics experts, it’s been encouraging to see the recent variety of educational robotics kits that offer interesting and extendable capabilities at an affordable cost. One such kit on Kickstarter this week comes from Morpx, a company founded by Tianli Yu, who spent several years at Google working on machine learning and image recognition. The kit is called MoonBot, and it manages to pack some nice hardware and impressive AI and autonomy into a cute (and affordable) customizable robot.

MoonBot (it’s one kit that you can use to build three robots, if that wasn’t clear) is currently US $149 on Kickstarter, and the campaign is already funded. This is certainly not the cheapest robot kit, but it looks like a good mix of ease of use and capability—you can start programming it visually with Scratch, and since it’s based around an Arduino-ish board, it’s easy to expand from there. 

Despite the relatively low cost, there seem to be a bunch of fairly polished, custom components, including a vision sensor, which makes MoonBot a bit of a standout relative to other, similarly priced kits. The sensor includes a deep learning pipeline along with algorithms that allow it to detect things including colors, lines, people, faces, and specific images and patterns, and putting those things together means that you can teach MoonBot to autonomously play basketball:

Implementing these algorithms can be done in Scratch, so programming the robot to use them is as easy as saying “if you see a person, do this thing,” and the vision board takes care of all of the “am I seeing a human” part. It’s features like this that are really making robotics exciting right now—clever software that enables all kinds of new capabilities without having to buy all kinds of fancy hardware.

For a bit more detail, we spoke with Tianli Yu, CEO an co-founder at Morpx.

Why did you decide to make MoonBot?

We want MoonBot to create an easy way for learning vision AI in robotics. With the dramatic advance of AI today, vision is moving into many more industries and products. We want to create a kit that people (especially children) can access early in their learning.

There are a lot of educational robotics kits available. What makes MoonBot unique?

What really makes MoonBot unique is the use of vision as its primary sensor. This creates a very unique experience—your robots suddenly can sense a lot more things. Adding all of these sensing capabilities together can deliver a whole new experience. Our basketball playing MoonBot is an example of a very integrated experience powered almost entirely by computer vision.

Sometimes robotics kids for people who are “just getting started” are still very complicated. What kind of skills or experience do you need to learn about robotics with MoonBot?

On the programming side we will need some basic Scratch programming skills and on the construction side, you should have some experiences building Lego sets. These are not that difficult as many kids already know both. And of course, you need to know how to use a screwdriver. The screwdriver is necessary because the real world is not always as simple as Lego, especially when you want precise and sturdy construction.

How will MoonBot stay interesting over the long term for people who buy one?

Good that you ask! We actually are planning a new service that delivers new robot building ideas and instructions on weekly basis, so you will have something fresh to try for at least half a year. And we are building some classic games into MoonBot such as basketball competition and tic-tac-toe. These games will have long term entertainment value and will stay interesting for the owners of the MoonBot.

What are you most excited about in robotics right now?

I am excited about the fact that we are starting to see AI powered robots now, with vision and voice recognition starting to make their way into robots. Of course, AI still has limitations, and will need a more cooperative relation between people and robots. One of the missions that MoonBot wants to achieve is to teach such cooperative relationships.

[ Kickstarter ]

The Conversation (0)

The Bionic-Hand Arms Race

The prosthetics industry is too focused on high-tech limbs that are complicated, costly, and often impractical

12 min read
Horizontal
A photograph of a young woman with brown eyes and neck length hair dyed rose gold sits at a white table. In one hand she holds a carbon fiber robotic arm and hand. Her other arm ends near her elbow. Her short sleeve shirt has a pattern on it of illustrated hands.

The author, Britt Young, holding her Ottobock bebionic bionic arm.

Gabriela Hasbun. Makeup: Maria Nguyen for MAC cosmetics; Hair: Joan Laqui for Living Proof
DarkGray

In Jules Verne’s 1865 novel From the Earth to the Moon, members of the fictitious Baltimore Gun Club, all disabled Civil War veterans, restlessly search for a new enemy to conquer. They had spent the war innovating new, deadlier weaponry. By the war’s end, with “not quite one arm between four persons, and exactly two legs between six,” these self-taught amputee-weaponsmiths decide to repurpose their skills toward a new projectile: a rocket ship.

The story of the Baltimore Gun Club propelling themselves to the moon is about the extraordinary masculine power of the veteran, who doesn’t simply “overcome” his disability; he derives power and ambition from it. Their “crutches, wooden legs, artificial arms, steel hooks, caoutchouc [rubber] jaws, silver craniums [and] platinum noses” don’t play leading roles in their personalities—they are merely tools on their bodies. These piecemeal men are unlikely crusaders of invention with an even more unlikely mission. And yet who better to design the next great leap in technology than men remade by technology themselves?

Keep Reading ↓Show less