Misty II — A Powerful and Refined Robot

It feels like every day is an exciting day at Misty, however, today is one of those few exceptionally exciting days! Today we launch a 30 day crowdfunding campaign for our new robot, the Misty II. The Misty II incorporates all of the learnings we’ve been collecting from the Misty I Developer Edition that started shipping earlier this year and puts them into a highly refined, fully injection-molded and manufactured body — on top of more great features.

So, what have we learned from Misty I and how? Most of what we’ve learned has come from people like you who’ve joined our community, attended our Robothons, and spent time developing for the Misty I. These people range from professional software engineers, to artists, to mobile developers, to high school students, and everyone in between. Here are a few of the top things we learned and have iterated on:

• Robustness — The robot needs to be solid and well built. Misty II is entirely injection molded with most parts cast in polycarbonate.
• Voice Input — We learned that people really want high quality voice interaction with their robot. With a three MEMS microphone far-field array that’s capable of beam forming, we’re putting a lot more energy into making sure that voice interaction works great and is easy to use.
• High Fidelity Audio System + Bass Port — Great sound is really important. We hired an audio engineer and went through many iterations to get to a design that is loud and clear.
• Automatic Charging — It’s a pain to have to always plug and unplug your robot. Misty II can now run skills at any time and charge herself when her battery gets low. For instance, if you want to write a skill that can check on your pet while you’re away from home, Misty will be ready to go.

Misty II automatically charges herself

• Edge Detection — People want to know Misty is safe. If left unattended, Misty II won’t fall down your stairs.
• Carrying Handle — Moving your robot can be difficult. The new carrying handle on the back of Misty’s head makes this easy.
• Touch — More interaction with Misty is better. We added six capacitive touch panels to Misty’s head to make this possible.
• Track Drive — The robot shouldn’t get stuck in the hostile environment of your home. This meant using a track-based drive system
• Expandability — We learned that people wanted more options for making additional hardware for Misty. Misty II has a magnetic panel on her head where you can add things like a projector. Remove one screw, and her arm pops off if you want to create your own gripper arm. A trailer hitch allows you to tow a custom delivery trailer. And a backpack with serial and USB ports give you power and data to your add-ons.
• Arduino Backpack — Make existing projects robot enabled. With the Arduino backpack you can simply plug your existing Arduino shields into Misty and load in the same code you wrote before.
• Manipulation — More than just pushing objects, Misty can use her arms to point, or you can attach a pen to allow her to draw.

• Expressiveness — In addition to her eyes, Misty II has a three degree of freedom neck, so she can look up, down, side to side, and tilt her head if she’s curious.
• Personality Engine — Misty II has a lot of personality built in, and we make it easy for you to carry this personality into your skills. If you want, you can even load in your own eyes and sounds to create your own custom personality for Misty.

And these are all on top of the features already built into the hardware of Misty I, like:

• Depth Sensors — For detailed room mapping and navigation.

Floor plan of part of our office generated by Misty

3D Version4K Camera — For things like face detection and recognition or creating a telepresense video chat skill.

• 4.3" LCD Display — To display Misty’s face or whatever data you want to show from your skill.
• Bump and Time of Flight Sensors — So Misty doesn’t get stuck or run into objects.
• Two Powerful Snapdragon Processors — One running Windows 10 IoT Core, as the main processor, and the second running Android 7, which handles the navigation and perception systems (but you don’t need to know how any of that works, we make it super easy).

But why Misty II?

To set the context, our vision is to rapidly progress robotics and eventually put a robot in every home and office — taking all of the mundane tasks out of our lives so we can spend our time doing the things we really want to be doing. We spent a long time doing research on how we could accomplish this and came to the conclusion that we’d have to satisfy four requirements in our products:

1. Has to be friendly, simple to use, and easy to program.
2. Can be iterated on in future versions to allow for even more capabilities to do even more useful things.
3. Has the capability to begin to do real useful stuff for us.
4. Must be able to do many different things. Not single purpose.

To expand on items 3 and 4 above — to do useful things you need a mobile platform (otherwise it’s just a really expensive Amazon Echo). You also need navigation so the robot can find its way around from room to room. You need to be able to interact with the robot using voice. Computer vision is important for identifying things and following you around. And finally you need to be able to easily take advantage of all of these features to program skills for the robots (and share them with others to use and build upon). We can’t keep spending weeks just getting the thing to move like we currently do in robotics today.

When we looked at existing solutions we found quite a few platforms but none even got close to satisfying our four criteria. Here’s what we found:

Many are STEM/educational robots that are geared towards learning programming with sometimes some flexibility on the mechanical side. On the mid to higher end this could be Lego Mindstorms, MakeBlock, Meccano, or the Alpha 1S as examples. They teach coding but you generally can’t do useful things with them. And if you did find something useful it’s very difficult and time consuming for others to duplicate. Creations are also fragile and single purpose.

If you want more advanced capabilities you can buy a Roomba that does a task well but you can’t program a Roomba to do anything else. You can also buy other robots that can start to do useful tasks but the price points are well out of reach of consumers in the tens of thousands of dollars for the PR2, Baxter, Pepper, etc. They can also take weeks to learn how to use — even if you’re an expert.

Then you have DIY robots where you’re on your own. You have to buy a mobile base, select your processor, sensors, etc and then integrate it all. The closest to a complete solution is the Turtlebot 3 at $1400-$1800 depending on the version but it’s very difficult to use if you aren’t an expert in Linux and ROS. Integrating components is INCREDIBLY difficult and just getting something that can drive around and avoid all of the crazy obstacles you’ll find in a home without getting stuck would take many months. And again, it’s a one off creation that can’t be shared with others. Not to mention, you haven’t even gotten to useful tasks because you’re still trying to figure out how to get it to not get stuck on your socks.

Misty is a huge leap forward! We hope you find her friendly and incredibly easy to program. We will iterate on Misty in the future, but she has the capabilities to start to do useful tasks for us today.

Today is an exciting day!

We have 50% off for the next 30 days to reward our early customers: Click here for the crowdfunding website and an additional $100 off

And even if you aren’t ready to buy a robot, please join our community. We would love to hear from you! https://community.mistyrobotics.com

~Ian