Automaton iconAutomaton

MonoCopter Modeled After Maple Leaf

Researchers at the University of Maryland studied the aero-dynamics of a maple leaf and used their findings to create and optimize a high-performance Micro Air Vehicle. Early results show that this 2 DoF MAV can potentially outperform more complicated helicopter and ornithopter style MAVs.

The following video shows the most recent work, showing its high performance and controllability.

From the creators:

The culmination of 3.5 years of research has led to controllable monocopter that can autorotate like a maple seed (Acer diabolicum Blume) and fly like a helicopter (hover and forward flight). The vehicle, invented at the University of Maryland, Aerospace Engineering Autonomous Vehicle Laboratory and Alfred Gessow Rotorcraft Center, is the smallest and most capable to date as it meets most of the challenges set forth by DARPA's nano-air-vehicle program.

The extended video goes into much more detail, prototype history, and even a demonstration of autonomous control.

The Robowranglers From Texas (Audio Slide Show)

At the FIRST Robotics Competition, the defending champion Robowranglers team learns engineering -- and life -- lessons.

This slide show is part of our special report "Robots for Real."

Embed this audio slide show on your site: 

Power Line Inspection Robot Balances Like Tightrope Walker

First time I see a robot balancing on a high-voltage power line like a tightrope walker!

This robot, dubbed the Expliner, was created by Japanese start-up HiBot, which spun off from the laboratory of famed Tokyo Tech roboticist Shigeo Hirose (known for his incredible snakebots).

Check out the video below showing an operator remote controlling the robot and performing some acrobatic maneuvers on a high-voltage line. Talk about unstructured environment! Read the full Spectrum article here.

Video: HiBot

MIT's Shape-Shifting Robots (Audio Slide Show)

MIT researcher Daniela Rus wants to create self-reconfiguring robots that can change their shape -- and become anything.

This slide show is part of our special report "Robots for Real."

Embed this audio slide show on your site: 

Lessons From a Mechanical Child (Audio Slide Show)

A child humanoid robot called iCub is helping Swiss scientists at EPFL's Learning Algorithms and Systems Laboratory study cognition, learning, and mobility.

This slide show is part of our special report "Robots for Real."

Embed this audio slide show on your site: 

Driving a Car With an iPhone


DIY enthusiasts have used their iPhones to remote control all sorts of contraptions -- LEGO robots, RC cars and helicopters, even an R2-D2 replica. But this is the first time I see a car -- a real car. A group of engineers at National Instruments -- they call themselves Waterloo Labs -- have transformed an Oldsmobile Delta 88 into a remote controlled car that they can drive using an iPhone app. In the video they give an overview of how they did their hack in just a few weeks. The control system uses a few motors, potentiometers, a NI Compact RIO embedded controller, and LabVIEW. The iPhone talks with the controller via Wi-Fi. Neat. Check out their site for more videos and technical details.

Thanks, Trisha!

How To Make a Humanoid Robot Open a Door

Roboticists know that some things are still hard for humanoid robots to do. Like walk, or run, or... open doors.

So presentations with titles like "Whole-Body Motion of a Humanoid Robot for Passing Through a Door," given at the IEEE/RSJ International Conference on Intelligent Robots and Systems last month, should perhaps not raise any eyebrows. But as a non-roboticist, I scratched my head. Can it really be that hard to open a door?

I showed up at Hitoshi Arisumi's talk to find out. Arisumi is a researcher at Japan's National Institute of Advanced Industrial Science and Technology (AIST), one of the world's top centers for humanoid research. And as it turns out, opening a door -- in this case, a swinging door -- is actually pretty darn hard, though we humans don't even have to think about it.

First, what do you need? Swinging doors are heavy, and to open them, you need a large force at the beginning of the motion, then you have to follow through with a continuous force to keep the door open as you walk through it. (If you're picturing a robot in hat, chaps, and holster strutting into an Old West saloon, like I was, Arisumi’s doors are more like the kitchen doors you might find in restaurants.)

One way a humanoid robot could open a swinging door would be to punch it, Arisumi said. But that doesn't give the door enough momentum to stay open as the robot passes through. What about a kick? That would leave the bot off balance -- not to mention the door swinging back in its face.

What you need, argued Arisumi, is to open the door to a desired angle and support the door through the entire opening action. It’s not as easy as it sounds.

 

 

The factors to consider are numerous: the robot's upper body position and foot position, the opening angle of the door, the angular velocity of the door just after opening, and the impact velocity of the robot, among other things. And keep in mind that when the robot is handling another object like a shopping bag or a tray of food, it's hard to open the door with just one arm. (By now I was impressed that we do it so well, after all.)

Arisumi went through equations and diagrams and graphs to demonstrate the proposed technique of using the robot's whole body to impact with the door, just like you would open it with your hip and side. Their experiments with an HRP-2 humanoid robot produced successful, stable opening of a swinging door with a mass of more than 80 percent of the robot's mass.

Now we just need a gun-slinging robot to complete the Old West picture.

Images: AIST

AWE Robotic Wall Reconfigures Itself Into Workspace, Lounge

 

The video session at the IEEE/RSJ International Conference on Intelligent Robots and Systems last month opened with a bit of chaos. The tables and chairs in the room were set up in board-meeting style, taking too much space and making it hard for attendees to face the screen. Eventually the tables were moved and the mass of people crowded in the doorway was able to come in and bring more chairs.

Which made me wonder, Where's the robotic furniture?!

It may be here soon. One of the videos presented at the session featured a robotic wall called AWE (animated work environment), developed at Clemson University's school of architecture and department of electrical and computer engineering.

Presenting the AWE Project was Keith Evan Green, the director of the school’s program of intelligent materials and systems for architecture. His opener: "We're in the business of reconfiguring rooms," he said, "which we just did here manually." 

He explained that the segmented wall can be rearranged to make a single space that is useful for working at a desk, giving presentations, even watching football. "The idea is for multiple people to convene in one place" with different objects and uses. 

The eight-degree-of-freedom robowall, powered by electric motors, has multiple touch-screen displays and mobile desk units. Users can select six different configurations and fine-tune them by gesturing at proximity sensors. 

Their original idea for the contraption's design came from studying the movement of an elephant’s trunk, though for their current prototype they went to a "hyper redundant" system more like the links in a watch's wristband. "Out of cowardice," Green said when asked, because they wanted to be sure it worked. But Green and collaborator Ian Walker are currently working on "continuum robotics," more like that trunk.

Maybe at IROS next year Green will wave his hand and the hotel conference room will morph into a cool movie theater?

Video: AWE Project

iRobot Launches Healthcare Robotics Division

During the TEDMED conference taking place in San Diego this week, iRobot announced the creation of a new product unit: healthcare robots. CEO Colin Angle said the overall goal is to add “one million hours of independent living” to seniors’ lives.

This is not at all a surprising move. Eldercare robots development has lagged in the US. Japan and South Korea have dedicated many more resources – including government dollars – to that development.  So iRobot is well-positioned to take advantage of the relatively empty US market. But this is also a logical outgrowth of the short-lived ConnectR project they worked on last year. ConnectR was designed to facilitate telepresence, such as between a parent and child in different geographical locations. Telepresence has applications in remote physician visits, or for adult children to check up on their elderly parents. Combine iRobot’s experience with telepresence with their innovative ways to automate common household chores in a user-friendly way, and you have a compelling case for a useful, assistive robot that may allow elderly people to live independently for a greater length of time.

Since the Internet is a great place to advertise brilliant ideas here’s one for the iRobot folks: I’d love to see a little robot that carries one of those day-of-the-week, time-of-day array medicine containers, and a bottle of water, that rolls over to my grandparents when it’s time for their next dose. There. You’re welcome.

But seriously, this is a great place for iRobot to be and a good thing for the US in the global robotics market. What I do find particularly interesting is that they’ve established this group under a division separate from Home Robots and Government and Industrial Robots. Since Angle was clear that they would focus on home-based healthcare robots, as opposed to surgical robots or the delivery robots used in many hospitals, I admit I’m a little surprised that this isn’t falling under Home Robots. But I have a few theories. In particular I wonder if a new distribution model may be in place: could these robots be subsidized by insurance companies and sold through special distributors like power wheelchairs? Will they need FDA approval for any of these products, like the iBot robotic wheelchair DEKA developed? If either of these is true, while the iRobot Healthcare Robots philosophy may be the same as Home Robots, such a unique business infrastructure does require its own division.

Volkswagen Demonstrates Autonomous Valet Parking System

VAIL autonomous car

At the official public introduction to the Volkswagen Automotive Innovation Laboratory and Stanford’s new robotic car on Saturday, the VAIL showed off their latest foray into intelligent vehicle technology: an autonomous valet parking system. The system does exactly what you’re probably hoping it does: you get out if the car, tell it to park itself, and off it goes to do just that:

And of course, when you’re ready to go home, you just call the car back and it’ll be waiting for you. VAIL researchers are envisioning something like an iPhone app to control all this, but it doesn’t need to be any more complicated than a “park” button and a “return” button. The car does the rest, all by itself.

VAIL autonomous car

The most awesome part of all this is that the Volkswagen Passat used in this demo, Junior 3, is almost (almost) entirely stock, as far as the hardware goes. There are three different primary sensors in use: a front radar (which is available as part of Volkswagen’s adaptive cruise control system), a camera mounted in front of the rear view mirror (also available for night driving assistance), and a couple little off the shelf LIDAR units mounted on the sides. The LIDAR isn’t currently part of any option package, of course, but Volkswagen does offer other side looking sensors, like lane assist and blind spot detection. The only other major difference is the giant rack of computers in the trunk, but it turns out that the computer system is a standard package for Stanford’s autonomous vehicles, and the computer that comes with the car is actually capable of running everything. Bottom line is, we’re technologically more or less ready for autonomous parking already, with in production vehicles.

So what’s the hold-up? Well, the car currently can’t detect obstacles. Like, you know, people. There’s no reason it couldn’t do that with the current hardware, and that’s in fact the next step, but it’s still a huge liability issue that Volkswagen doesn’t want to tackle. The car also needs a detailed map of the parking lot, so you wouldn’t be able to just drop it off anywhere (yet). So when we first see this technology, VAIL envisions parking garages that are specifically designed for self parking vehicles. There would be a drop off area, and no pedestrians would be allowed in the garage, which wouldn’t need elevators, stairs, walkways, or even lights. This infrastructure isn’t ready yet, but the vehicle technology is here now.

VAIL autonomous car

It’s a little bit frustrating to robotics proponents like myself that in cases like these, it seems as if technology is advancing faster than society is prepared for. This is part of what VAIL is here for, though… To figure out not just the technology, but also tackle all of the related issues. There are a bunch of really smart people working on this stuff, and I have to say, if they continue making strides like this, we’re all gonna be worrying a lot less about driving in the near future.

Most Commented Posts

Automaton

IEEE Spectrum's award-winning robotics blog, featuring news, articles, and videos on robots, humanoids, automation, artificial intelligence, and more.
Contact us:  e.guizzo@ieee.org

Editor
Erico Guizzo
New York, N.Y.
Senior Writer
Evan Ackerman
Berkeley, Calif.
 
Contributor
Jason Falconer
Canada
Contributor
Angelica Lim
Tokyo, Japan
 

Newsletter Sign Up

Sign up for the Automaton newsletter and get biweekly updates about robotics, automation, and AI, all delivered directly to your inbox.

Advertisement
Advertisement
Load More