Researchers Are Using YouTube To Teach Robots How To Cook

Image source: University of Maryland

Humans aren't the only ones who outsource to YouTube when learning how to cook up meals. With funding from DARPA's Mathematics of Sensing, Exploitation and Execution (MSEE) program, researchers at the University of Maryland have programmed a robot to learn the most basic cooking skills just by watching YouTube tutorial videos.

The system enables the robot to process visual data from a series of YouTube cooking videos. The robot is then able to identify, pick up, and use the kitchen object to perform the task demonstrated in the videos. All of the robot's actions are with extreme accuracy and don't require extra human input.

Baxter, shown above, is the factory and assembly line brainchild of Rethink Robotics. Founded by iRobot co-founder Rodney Brooks, Rethink Robotics has made their Baxter robot one the simplest assembly robots available on the market. With minimal human input, Baxter is able to learn, memorize, and perform tasks.

In addition to learning what's shown in the YouTube videos, Baxter is able to store learned information and build upon it.

“This system allows robots to continuously build on previous learning—such as types of objects and grasps associated with them—which could have a huge impact on teaching and training."
-Reza Ghanadan, program manager with DARPA’s Defense Sciences Offices

Ghanadan explains that this technology could allow for faster, less expensive robot programming. Robots could also share what they've learned to help automate the hassle that is robot programming.



Source: DARPA Via: GigaOm

Q&A With Drone Fleet Management Startup PixiePath


PixiePath is one of the few drone management platforms that exists. It's a young startup, less than 6 months old in fact, that provides a system for controlling entire fleets of drones. From drone delivery to disaster relief, the cloud-based PixiePath is here to make drones work.

A few weeks ago, we got in touch with one of their PR reps and they we're nice enough to forward our questions to PixiePath's founder Bryan Field-Elliot.

When was PixiePath founded?

PixiePath was founded in August 2014, less than six months ago.

How was PixiePath initially funded? Angels, VC, or will a crowdfunding campaign further fund the company?

PixiePath is funded to-date by early seed money. Shortly, we will open up a formal seed round offering after releasing the first version of our product and enabling a few beta customers.

What could PixiePath provide for a fleet of drones such as the DJI Phantom?

PixiePath’s value proposition is in two parts: 1) Connecting individual drones to the cloud, and 2) Directing the activities of fleets of drones once they are cloud connected. In the case of the DJI Phantom, PixiePath will offer a simple means to connect the telemetry, battery strength and camera output to the cloud. PixiePath also provides a simple way of processing that information in the cloud, as well as sending positioning information back down to each Phantom. In this way, for example, a Search & Rescue operation over a wide area might become quite easy to survey using a fleet of Phantoms. Simply draw a map over the area to be surveyed and watch as the map is divided up into regions, one for each Phantom. The Phantoms (which might, for example, number around five) would take off and aerially sweep different regions. All the while, their video will be sent up to the cloud where a rescue team can view it all, live in real-time from a single console.

Could PixiePath work with services similar to Amazon’s Prime Air?

One of PixiePath’s original goals is to provide a software platform upon which a delivery service could realistically be built. If Amazon were to focus on hardware and on-board software, and then find their fleet management layer to be lacking, then I would anticipate having conversations about using the PixiePath SaaS platform for fleet management.

Due to FAA regulations with drones and the increasing industry, when could we expect to see PixiePath in action?

As soon as PixiePath delivers it’s beta (expected by February), PixiePath will immediately be legal to use in the U.S. by hobbyists for recreational purposes. It will also be legal to use in the U.S. for commercial purposes by companies who have obtained an FAA exemption permit, the number of which is growing every week. Lastly, PixiePath has global interest from prospects in the Far East and Canada. We will be flying in many places very soon.

Ultimately, what does PixiePath provide for drones and how could it advance the technology and use of drones in a positive way?

By controlling movements of individual drones or fleets of drones in a programmatic way, PixiePath ultimately offers the positive benefits of safety, predictability and reliability for drone operators; as well as the people and structures nearby. It does this by giving a framework for describing movements, routes and tasks; and then executing them automatically thereby removing the need for humans to pilot each maneuver. That said, there are still plenty of other uses for PixiePath even for scenarios in which a human pilot still needs to guide a drone.

Nixie Answers Your Questions, Says Wearable Drone Will Cost "Slightly More Than A GoPro"


The Nixie wearable drone is on the verge of hitting the market. With early adopters looking forward to the first wearable drone, the startup behind the drone has answered some very important questions regarding Nixie. Here are the answers from a source provided to SimpleBotics.

  • Battery life will last a full day of typical use (50 selfies)
  • An HD camera (similar to those in smartphones) will be used
  • The drone will cost slightly more than a GoPro ($300+)

Boston-Based JIBO Company Secures $25.3M Series A Funding


The personal, artificial intelligent JIBO robot was a whopping success on Indiegogo last year. The brainchild of roboticist Cynthia Breazeal raised over $2M in crowdfunding. Forward less than a year, and the company has reported that is has secured $25.3M in Series A funding. The funding round is led by RRE Ventures and includes Charles River Ventures, Fairhaven, Osage University Partners, Flybridge Capital Partners, Formation 8, Two Sigma Ventures, and Samsung Ventures.

According to VentureBeat and other sources, the company plans to use some of the funding to fulfill some of the 4,800 orders they received from the Indiegogo campaign. Expanding staff to around 50, developing robot prototypes, and getting developers to pump out new apps are also in mind. The company states on their site that early adopter developer editions of JIBO will ship late Fall 2015. The home editions (regular consumers) will ship out Winter 2015. Any JIBOs pre-ordered after the campaign are set to ship in late Spring of 2016. That's pretty decent keeping in mind that they're putting together 4,800 half-sentient robots.





DARPA's ATLAS Robot Gets An Upgrade, Can Walk Untethered


The Google-owned Boston Dynamics company has taken another big step forward with their humanoid ATLAS robot. In a video uploaded by DARPA, ATLAS is shown in new body “armor”, is completely untethered, and is walking all on its own. This is a big improvement from the ATLAS robot we saw last time.


Google acquired Boston Dynamics last year to add to its robotic/future technology agenda. However, it’s DARPA that's interested in using the ATLAS robot for disaster relief and aid.

The upgraded ATLAS has been redesigned for DARPA by Boston Dynamics, and the main focus is on improving power efficiency and usage. Sources say that approximately “75 percent of the robot was rebuilt.” The lower legs and feet are some of the few pieces left over from the previous robot. But the big upgrade from the previous model is that wires are no longer required. Thanks to the internal 3.7-kilowatt-hour lithium-ion battery pack, ATLAS has a potential one hour to walk, open doors, and use tools. No more limitations that would have been by using a tether.

Image source: DARPA

Other upgrades include the following.
Repositioned shoulders and arms allow for increased workspace in front of the robot and let the robot view its hands in motion, thus providing additional sensor feedback to the operator.
New electrically actuated lower arms will increase strength and dexterity and improve force sensing.

The addition of an extra degree of freedom in the wrist means the robot will be able to turn a door handle simply by rotating its wrist as opposed to moving its entire arm.

Three onboard perception computers are used for perception and task planning, and a wireless router in the head enables untethered communication.

Re-sized actuators in the hip, knee, and back give the robot greater strength.

A wireless emergency stop allows for safe operation.

As a result of the new pump, Atlas is much, much quieter than before!

ATLAS is also set to compete against twenty teams in the DARPA Robotics Challenge Finals. This challenge will be from June 5-6, 2015 in Pomona, California. More information on the challenge can be found on DARPA's site.



Source: DARPA

Robotbase Raises $80,000 On Kickstarter To Put A Robot In Every Home

Image source: Robotbase

Update: You can test out Robotbase's object recognition at see.robotbase.com

One of the greatest ambitions of today's robotics companies is to fulfill what science fiction has promised us for decades. A robot in every home is an incredible (and somewhat exaggerated) feat, but one startup has turned to crowdfunding to accomplish just that.

The New York-based startup Robotbase has turned to crowdfunding to build the world's most advanced and affordable personal robot. With their robot's promising features and connectivity to a large amount of apps, the campaign has already raised over $80,000 since its inception.

Maya, unlike acting as another Roomba, is a smart robotic assistant. She stands at around 4 feet and roams around on a small, wheeled robot base. Maya is compatible with a large collection of apps. From Tweeting updates to telling you if you'll need an umbrella, Maya is always connected. A front-facing screen displays a friendly face while acting as the robot-to-human interface. Don't worry, Maya doesn't fall into the freighting abyss of the Uncanny Valley. Take a look.


Along with giving weather updates and ordering food, Maya can be used as a robotic security guard around the house. Maya can track and avoid objects as it roams different rooms. She can notify users if doors are locked, lights are off, and what the current temperature is. This does require smart home technologies like Lockitron and Nest of course. But there's more to Maya than just a few built-in apps.

Image source: Robotbase

SimpleBotics got in touch with the founder and CEO of Robotbase Duy Huynh. Huynh explains that Maya will be able to expand its library of apps by downloading apps from the "Ability Store." In regards to the ability store and its apps, "It will cover a variety of categories, but the main ones are Home Automation, Personal Assistant, Entertainment, Games, Home Security, etc." Changing the robot's personality based on the user's preferences will also be a feature. "You will be able to personalize voice, face, and name."

$50,000, when looking at similar campaigns, doesn't seem like much, but Duy Huynh says that "the goal of our project is to raise awareness and collect feedback from our backers." So far Robotbase has acquired funds from eight angel investors. Some notable investors include Dave McClure from 500 Startups and Dave Nemetz founder of Bleacher Report.

Maya was pitched at TechCrunch's Hardware Battlefield at CES 2015.


Backers who found the project early on were able to snag the robot for $995. The regular price is currently $1500 (though there's still an early bird deal available for $1200). If Robotbase is capable of making the final product as useful as its shown in the video, I'm excited and welcome our new robot friends.



Source: Kickstarter Via: TechCrunch

Amazon Japan Is Now Selling Giant, Piloted Mech Suits


Have you ever dreamed of driving around a giant, badass mech suit? If so, the Japanese Kuratasu Starter Kit can be all yours for just 120,000,000 yen on Amazon Japan. That's about $1,000,000 US dollars.

See also - How to pilot a Kuratas (mech shown above)

The company does specify in the Amazon description that the mech weighs about 5 tons and stands at 3.8 meters tall. A little warning to the pilot explains, "Unlike normal riding toy, safety and comfort is not guaranteed." Keep in mind that the starter kit doesn't come with the mech's arms. Sadly, they're sold separately.





Via: Technabob

Review: Dash And Dot Robots Teach Kids The Basics Of Code Through Fun, Connected Apps

"A pricey tech-toy that teaches kids to code through fun, connected apps."



Meet Dash and Dot (or you may know them as Bo and Yana). They're two little robots that teach kids how to code through fun, connected apps. But they're much more than just that.

See Also: We take a look at Sphero's Ollie

Wonder Workshop (originally Play-i) is a young startup that successfully raised over $8M through crowdfunding to back the idea that robots could be used to teach kids code. Less than a year later and their product has already made it into the homes of the original backers.

As we love to cover consumer robots, we took the time to do hands-on, in-depth review of Dash and Dot. Be sure to stay tuned for a video review.

Both robots lay inside their own aesthetically-pleasing box. Dash, the bigger of the two, has its own ramp that opens up from the packaging. Each bot also sports a nice name tag. Instructions, a USB cord, a Lego adapter, and a quick-start guide are also included.

The robots look and feel unique right at first glance. The material, design, and overall feel is quite nice and isn't found in many robots on the same side of the market. A simple power button on the side of each robot powers the robot on. One click and each robot starts up with a friendly tone and sound. LEDs mounted around each robots' eye illuminate and add an extra touch of personality. This sort of robotic personification can get kids really attached to Dash and Dot.

So the robots are nice and all, but what do they do?

There are four apps developed by Wonder Workshop that allow kids to either program or manually control Dash and Dot. The main app is Blockly, and it's an intuitive, block-based programming app. With Blockly's slick and cartoonish GUI, kids can program Dash to move around, avoid obstacles, and solve mazes. The other three apps aren't as code-oriented, but they are still pretty fun to use. Path lets kids draw a path for Dash to follow, Xylo lets Dash play the xylophone (attachment required), and Go lets kids control Dash manually. Take note that Dot doesn't move but can still be used to communicate with Dash. Blockly, Path, and Xylo do require either an iPad (3, Air, or Mini) or Android tablet. Phones are currently not supported for those three apps, but Wonder Workshop does have plans.

The only app compatible with iPhone and Android smartphones is Go. The Go app allows kids to control Dash, change the LED colors, play some fun sounds (dog barking, siren, trumpet, etc.), and activate preset moods. Dot's LEDs can be changed and it can also play sounds.

Multiple accessories are also available for purchase. A phone dock for Dash, bunny ears, and Lego-compatible bricks are some of the simpler attachments. The best attachments include a small bumper and xylophone that Dash can use. You can view an example of Dash playing the xylophone through the "Xylo" app right here.


So how well do Dash and Dot work in terms of teaching kids the basics of programming? I personally don't know, but others have said that Dash and Dot have done well. You can read a review on Business Insider or reviews from parents on Amazon.

With the complete Dash, Dot, and accessory bundle being $365, these robots make for some pricey tech-toys. Dash alone is $199 and spending more for an immobile Dot seems pretty pointless. I personally recommend just getting Dash and the accessories separately. In summary, Dash and Dot make a great introduction to computer education for young kids. Sure beats learning through text editors.




CNN And The FAA Partner Up To Make Drone Reporting A Reality


CNN has been trying for months to integrate drones into news and research gathering. The news organization confirmed earlier today that they've made an agreement with the FAA (Federal Aviation Administration) to do just that. As stated in the agreement, CNN has "entered into a Cooperative Research and Development Agreement (CRDA) with the Federal Aviation Administration to advance efforts to integrate Unmanned Aerial Vehicles (UAVs) into news gathering and reporting." Part of this plan involves setting up a framework in which certain drones (beyond hobby-level) can be used in a safe manner to gather high-quality news. FAA administrator Michael Huerta explains “We hope this agreement with CNN and the work we are doing with other news organizations and associations will help safely integrate unmanned news gathering technology and operating procedures into the National Airspace System." My guess is that it won't be long before we see news organizations like Fox and CNN using drone swarms to report the news. I predict less than five years.

What's your view on drones being used for media and news purposes? We'd love to hear your comments!



Source: CNN Via: Engadget

SimpleBotics CES 2015 Readers' Choice: Robots And Drones


It's been one crazy CES this year. Many companies from around the world have revealed their latest robotic and drone consumer products. From micro drones that take selfies to robotic grippers that beat humans in beer pong, CES 2015 has not let us down. It's your turn to choose your favorite robot in the SimpleBotics CES 2015 Readers' Choice poll. The voting session will be open tonight and conclude tomorrow at 11:59 CST. May the best bot win.