Skip To Content
NSF Logo Search GraphicGuide To Programs GraphicImage Library GraphicSite Map GraphicHelp GraphicPrivacy Policy Graphic
OLPA Header Graphic

News Tip


October 21, 2003

Sense and Sensor Abilities from NSF Robotics and Computer Vision Research

NSF's Robotics and Computer Vision (RCV) program provides opportunities to develop novel ideas in advanced intelligent robotic systems and visual perception systems. The following activities—from the hundreds of projects the program has supported—will be demonstrated when NSF-supported robotics and computer vision researchers convene at Bally's Hotel in Las Vegas for an NSF workshop and the IROS 2003 international robotics conference the week of October 26-31. For more information, contact David Hart at 703-292-7737 or

NSF Robotics and Computer Vision Program Director: Junku Yuh, 703-292-8930,

ROBOMOTES: ROBOTS IN SMALL PACKAGES— Tiny, two-wheeled and wireless, RoboMotes are designed to create sensor networks that move and reconfigure themselves to adapt to new situations. Developed at the University of Southern California's Robotic Embedded Systems Lab, the golf-ball-sized RoboMotes are built on the Mote platform that is becoming popular in the sensor network community. Each RoboMote includes a wireless network interface (the "Mote"), two controllable wheels with odometers; a solar cell for "always-on" power; a compass for direction; and bump and infrared sensors for obstacle avoidance. Because of their small size and low cost, RoboMotes make it possible for campus labs to experiment with larger dynamic sensor networks.

RoboMote Project:

photo of RoboMotes
The RoboMote: A platform for research in robotic sensor networks.
Credit: USC Robotic Embedded Systems Lab (PI: Prof. Gaurav S. Sukhatme)
Select image for larger version
(Size: 278KB)

FINGERNAIL SENSORS FOR A LIGHT TOUCH—Repetitive strain injuries are one of the nation's most common occupational health problems, costing businesses $2.8 billion annually according to the 2003 Liberty Mutual Workplace Safety Index. Many repetitive strain injuries stem from the use of conventional computer keyboards and mice. Researchers led by Harry Asada at MIT have developed fingernail sensors that read the subtle changes in blood flow under the nail to detect the press of a fingertip even lightly against a surface and the bending angle of the fingertip. And because the sensors mount on the fingernail, they don't interfere with the finger's sense of touch. The sensors might replace computer keyboards and mice for persons with repetitive strain injuries or with limited hand mobility.

photo of fingernail sensors
Fingernail sensors unobtrusively measure the state of the human fingertip. The sensors optically measure the pattern of blood volume in the capillaries beneath the fingernail, which depends predictably on the posture of the finger as well as the forces of the fingertip pressing against a surface (including two dimensions of shear force).
Credit: Stephen Mascaro, MIT
Select image for larger version
(Size: 48KB)

SCOUTING OUT NEW TERRITORY—A robot's design typically depends on the task it was built to accomplish, and robot teams are often composed of robots of similar form. Nikos Papanikolopoulos and colleagues at the University of Minnesota Center for Distributed Robotics are conducting research into robot teams that have the added complexity of two or more different types of robots, each with different sensing and control capabilities. The COTS Scout, for example, is a remotely controlled robot, about the size of a coffee mug, with a black-and-white video camera. The MegaScout is a 15-inch long robot, with color cameras and other sensors designed to provide command-and-control support for a Scout team. The ultimate objective is for the individual robots to combine forces and accomplish certain tasks better than any single robot.

Center for Distributed Robotics:

photo of COTSScout, MegaScout, Scout
The five-inch wide COTSScout (left), 15-inch MegaScout (center) and four-inch Scout (right) from the University of Minnesota Center for Distributed Robotics.
Credit: University of Minnesota Center for Distributed Robotics
Select image for larger version
(Size: 238KB)

DIGITAL CLAY FOR COMPUTER TOUCH AND FEEL—It often feels like people have to mold themselves to the computer rather than the other way around. Wayne Book and colleagues at Georgia Tech want to change that. They are developing a technology called "digital clay," in which computer users would interact with virtual worlds or objects much as they would shape modeling clay. In some cases, the digital-clay surface might even push back—for example, if the virtual object the user is molding resists being re-shaped. The potential applications for digital clay range from designing 3-D objects and navigating 3-D virtual worlds to providing an alternative interface for visually challenged users.

Digital Clay Project:

photo of digital clay sculpted
Digital Clay combines the intuitive sculpting of natural clay with intelligent, programmable haptic feedback to the sculptor. The photograph shows two implementation concepts: the "formable crust" and the "bed of nails." The formable crust uses a new implementation of the ball joint (foreground) that, when combined as illustrated (on the middle left), enables the shell to conform to arbitrary shapes. As implemented in stereolithography (upper left), it can be scaled to fine resolution. The bed of nails concept and its control is studied using an expandible 5x1 array of cells (upper right). With MEMS technology under development for valves and actuators, this concept will be scalable to higher resolution, similar to the surrogate bed of nails model of a camera (middle right). This stereolithography model was created for user studies of the needed resolution and optimal surface.
Credit: Georgia Tech Digital Clay Team: M. Allen, W. Book, I. Eber-Uphoff, A. Glezer, J. Goldthwaite, D. Rosen, J. Rossignac, C. Shaw and their colleagues and students
Select image for larger version
(Size: 439KB)

LONG-DISTANCE ROBOT CONTROL AND FEEDBACK—Remote operation of scientific instruments across the Internet works best if the instruments, such as telescopes or electron microscopes, have well-defined movement and behavior. However, controlling a mobile robot with a manipulator arm—perhaps on another continent—is not so straightforward when the operator can't feel the robot's movement, position or grasp, not to mention the surrounding environment. Ning Xi and colleagues at the Michigan State University Robots and Automation Lab are creating a sensory link to help humans feel connected to a robot in a remote environment. This sensory connection—in the form of visual, audio, tactile and even temperature feedback—must overcome numerous technical networking issues, such as the unpredictable delay in signals traversing the Internet.

Michigan State Robots and Automation Lab:

map of locations who receive supermedia feedback
Ning Xi at the Michigan State University Robots and Automation Lab and collaborators have designed a system so that people in different parts of the world can control mobile robots and receive "supermedia" feedback at remote locations via the Internet.
Credit: Michigan State University, Chinese University of Hong Kong, Chinese Academy of Sciences
Select image for larger version
(Size: 110KB)

IMPROVED ROBOT VISION SAVES BADLY EXPOSED PHOTOGRAPHS—Don't throw away those badly exposed digital photographs just yet. Shadow Illuminator can fix them in a snap. Shadow Illuminator technology, by Vladimir Brajovic at Carnegie Mellon University, results from Brajovic's NSF-funded computer vision research project. Unlike humans, robots are often confused when the lighting conditions change. Shadow Illuminator makes robot vision more reliable despite bad lighting by mathematically mimicking processes that take place in the human eye. Shadow Illuminator "looks" at the scene content, estimates the illumination conditions and digitally adds light to dark areas. The shadows are brightened, and details in the shadows are enhanced as if they were properly lighted. At the free Shadow Illuminator Web site, you can upload your picture and instantly see details you didn't know were there.

Shadow Illuminator:

Top of Page

STRETCHABLE CIRCUITS FOR ROBOTIC SENSOR SKIN—Human skin protects the body from the outside world, bends and stretches to let us move around and has nerve endings for our sense of touch. But robot "skins" thus far have not been able to handle all three functions at the same time because typical electrical circuits don't like to be stretched. Sigurd Wagner and Stephanie Lacour at Princeton University have now changed that. They have created the first stretchable metal film interconnects for elastic integrated circuits. Consisting of 25-nanometer-thick stripes of gold on a silicone membrane, these interconnects can be stretched by at least 15 percent and bounce back, conducting electricity all the while. Some samples have been stretched to twice their original length and remained conductive. But so far, they don't know why it works. The elastic circuits represent a new material for robotic skin that can stretch, wrinkle or shrink while transmitting data to embedded sensors.

Princeton Macroelectronics Group:

samples of robotic sensor skin
Samples of robotic sensor "skin" with stretchable gold electrodes deposited onto compliant silicone elastomer membranes.
Credit: S.P. Lacour, S. Wagner, Princeton University
Select image for larger version
(Size: 155KB)

Larger versions (Total Size: 1,268KB) of all images from this document

  Note About Images



National Science Foundation
Office of Legislative and Public Affairs
4201 Wilson Boulevard
Arlington, Virginia 22230, USA
Tel: 703-292-8070
FIRS: 800-877-8339 | TDD: 703-292-5090

NSF Logo Graphic