Misty Robotics Builds on Developer Platform With New Personal Robot

The Misty II personal robot is designed to do whatever you can program it to do, and more

7 min read

The Misty II personal robot is designed to do whatever you can program it to do, and more
Image: Misty Robotics

Misty Robotics announced their developer platform, Misty I, just a few months ago. Misty I was a hand-built prototype, with similar essential functionality in hardware and software to a more refined production robot that the company planned to release later in the year. Later in the year has now arrived, and today Misty Robotics is launching a crowdfunding campaign for a much more polished personal robot, Misty II, which can be yours to program starting at $1,500.

In theory, I can definitely appreciate this idea of a platform being what’s holding useful personal robotics back. It’s very appealing to think about, and it would be wonderful if true, with legions of independent developers who can’t wait to work with something capable and accessible enough for them to program to do what they want it to do without any special training. Whether or not it’s true in practice remains to be seen, though— I’d love to be more optimistic, but one thing robotics has definitely taught me is that being optimistic is usually just leads to disappointment.

Misty Robotics, fortunately, seems to have a compelling combination of optimism, experience, and patience. They’re not expecting Misty to take off right away, and they have a timeline that’s about a decade long. From the sound of things, it’s fully expected that despite its consumer look, Misty won’t actually become "useful" to most consumers for many, many years. And Misty Robotics (and their investors, presumably) are totally fine with that. So really, we’re not going to be able to judge how successful the Misty Robotics approach is for quite a while. We’ll watch with interest, and hope that they’re on to something.

Misty 2

At 14 inches high and six pounds, the Misty II personal robot has an advanced 3D Occipital sensor for mapping, a 4K Sony camera for facial and object recognition, eight time-of-flight sensors for obstacle avoidance, and is equipped with two smartphone-level QualcommⓇ SnapDragon™ processors running Windows IoT Core and Android 7 operating systems. The robot is fully capable of running Tensorflow, Caffe, and WindowsML within hardware optimized runtimes. Owners can physically expand Misty II to add functionality and fun with several hardware accessories, created by Misty Robotics and the community of makers and inventors at large.

One of the things that should set Misty apart is that it’s been designed specifically to be able to perform advanced behaviors without requiring advanced knowledge of robotics. Out of the box, Misty II can:

  • Move autonomously as well as dynamically respond to her environment
  • Recognize faces
  • Create a 3D map of her surroundings
  • Perform seeing, hearing, and speaking capabilities 
  • Receive and respond to commands
  • Locate her charger to charge herself
  • Display emotive eyes and other emotional characteristics

All of this stuff can be accessed and leveraged if you know how to code, even a little bit. Or even not at all, since Misty can be programmed through Blockly:

The Misty II personal robot is easy for non-technical owners to program using the Misty Blockly client, a visual block-based programming interface, to create new skills for the robot that can make it move, talk, roar, and more. Pre-set blocks, or skills, will come installed on Misty II to quickly get started. Those owners with programming experience, can utilize JavaScript APIs to create more sophisticated skills and modifications, like integrating with third party services such as Alexa, Microsoft Cognitive Services, Google Assistant, and Cloud APIs.

If you’re quick, you can snag a "Founder Edition" Misty II for $1,500, and after that, the crowdfunding price is $1,600. The press release says that this price is "a 50% discount," making the robots quite expensive. If they’re as fun and as useful as Misty Robotics hopes they’ll be, maybe the price won’t be too much of an obstacle. Again, we’ll be following closely, and hoping that this strategy pays off both for Misty Robotics itself and the larger personal robotics industry as well.

For more details on Misty II, we spoke with Misty Robotics Head of Company Tim Enwall.

IEEE Spectrum: Misty I has been out for several months now. Can you give us some examples of useful applications that Misty I developers have come up with? What kind of feedback have you gotten from those folks on the overall experience of developing applications on Misty, and did that result in any changes in design or philosophy between Misty I and Misty II?

Tim Enwall: Our goal of developing Misty I is to get dozens of lessons every week from early customers and community, and they have shared what features they want, and precisely how they expect to use them. Through our focus on making programming Misty easy, we’ve had the chance to meaningfully engage with people who had never thought about where a robot might be useful to them. Here are few examples: 

  1. The HVAC repair specialist who wants to use Misty to ensure his repairs were effective by checking different locations and temperatures over time.
  2. The Sound Engineer who has to manually roam a venue to calibrate the speaker system wants to utilize Misty to take those readings for him so he can stay in the booth to configure the system. 
  3. The two students at Galvanize who used in Misty I in their JavaScript class. They integrated Alexa and had Misty read text messages to its owners.
  4. The pair of developers at our most recent Robothon who, within a few hours, trained Misty to recognize each of them individually and say things like ‘Ian, you’re a great boss’ or ‘Shane, would you like a snack?’
  5. One of our earliest users who is interested in using Misty as an informational tool at conferences.
  6. The users who have also integrated Misty’s output to externals services like Slack, Twilio, and Google Translate and Dominos to notify them of when their pizza will arrive. 

By having Misty I, we’ve been able to focus on the building the tools to suit our users instead of guessing such as: on-board SDKs are vital to really extending the power of the robot; dust covers are important for time-of-flight sensors; battery power indicators in the companion mobile app are valuable; continue working on personality, because it’s a pleasant delight; Python may be even more appealing than Javascript; access to machine learning engines and microphones are a higher priority than we thought, and dozens more.

The $1,500 price for a crowdfunded Misty II may be a significant investment for many developers. How much of an obstacle do you believe that cost will be for adoption of Misty II?

We feel that $1,500 is the most affordable price for such an advanced robot; several others in the market are more expensive with less capability. We believe, by including a sufficient number of developers, entrepreneurs, students, and makers in our plans early, the result will be better personal robots in the future because they will do hundreds of useful things— not one or two— which increases the value regular consumers are willing to pay. There has already been a lot of excitement from people who are interested in AI, machine learning, or robots, and developers who are dedicated will embrace the opportunity to participate in creating skills for the Misty II. We’re confident that for software developers with business problems to solve, students who want an AI and robotic future, and hackers/makers who love building novel creations, $1,599 is a floor price, not a max price.

We’ve learned from other platforms, like cell phones and computers, that as the uses for the device go up, consumers’ willingness to pay increases as well. What value will a consumer ascribe to a robot that can act as security guard, nanny, remote eyes, and entertainment pal not to mention other skills all in the same package? Laptops and TVs teach us that such general purpose tools are often quite valuable.

By crowdfunding, rather than pre-sale, we are able to launch Misty on her journey and build a community around Misty. Crowdfunding aligns with our early mission of building a community of developers and makers. It’s about inviting people to support our community in this early stage as we embark on this 10 year road to putting a robot in every home and office. We also believe the crowdfunding campaign will serve as a channel for getting feedback and will help to reinforce our assumption that the target market is ready for an easily programmable robot at a consumer price point.  

The press release says “owners can physically expand Misty II to add functionality and fun with several hardware accessories, created by Misty Robotics and the community of makers and inventors at large.” Can you describe some of these accessories?

Misty can be expanded with hardware accessories that are serial or USB enabled (backpack), attached to the head (headpiece), non-powered attachments to the arm motors (arms), and non-powered items carried behind (via “trailer hitch”) to make her more useful for daily tasks. During the crowdfunding campaign and after, we’ve made an Arduino backpack capable of plugging in most Arduino shields. Some of the ideas that have circulated are: 

  • Backpack: Carry temperature and humidity sensors for the HVAC technician to leave while the robot monitors the environment after installation/repair
  • Backpack: Smart home (Z Wave or Zigbee) addition for robot to interact with homes
  • Arms: Projector arms (powered via backpack) to project an image on a wall/screen
  • Arms: Laser pointer (powered by backpack) to play with cats
  • Earpiece: night-vision camera (powered by backpack)
  • Trailer hitch: pull a beer cart at a party

What was the thinking behind making Misty explicitly female?

When originally deciding what to name the robot, we thought ‘Misty’ was a friendly name that’s easy to remember, spell, and say. We want to encourage people to feel comfortable bringing a robot into their home, and we believed making Misty female would give her an approachable, friendly connotation. In practice, many of our customers assign a male gender to “their robot” and many, a female one.

Misty II is described as “the second robot in a series of robots... over the course of the next 10 years.” What is the general trajectory that you expect the Misty series to take over the next several years? 

The Misty II is primarily for software developers, STEM students, and makers. This community will invent thousands of uses, which, we predict, will take about three to five years. During this three to five year timeframe, there will be advances in processing speed to enable general object recognition, advances in motor life, manipulation, as well as finding ways to make the battery last longer. After truly homing in on improving the extensibility and depth of APIs, it will then be early adopters who will utilize robots in the home and office. Five years following that, we feel that Misty will be ready for mass adoption (i.e. greater than 20% penetration) consumers. A robot will not appeal to a regular consumer unless it is comfortable to interact with and multi-purpose (capable of hundreds of tasks), which is where developers come in to create these skills. Nevertheless, the robotics market could accelerate based on market shifts and our goal is to be ready when that happens. Our product release cycles will be somewhere between a highly-competitive smartphone market (typically one year) and an industrial upgrade market like thermostats (usually 7-10 years) with a weighting towards the former rather than the latter.

[ Misty Robotics ]

The Conversation (0)