On the computer screen, each user sees a three-dimensional room. Within that room are a black box and two tiny square pointers that show the users where they are in the room. They then use the robotic arms to collaboratively lift a box.

As a user at MIT moves the arm—and therefore the pointer—to touch the box, he can "feel" the box, which has the texture of hard rubber. The user in London does the same thing. Together they attempt to pick up the box—one applying force from the left, the other from the right—and hold it as long as possible. Each user can feel the other’s manipulations of the box.

The feeling of touch comes through a pencil-like device that sends small impulses at very high frequencies—up to 1000Hz—through the Internet. "You cannot only feel the resulting force, but you can also get a sense of the quality of the object you’re feeling—whether it’s soft or hard, wood-like or fleshy," says Mel Slater, a professor of computer science at University College.

"Touch is the most difficult aspect of virtual environments to simulate," explains Slater. "The impulse frequencies need to be very high to imitate convincingly the sense of touch. In much the same way that the brain reinterprets still images into moving pictures, the frequencies received are similarly integrated to produce the sense of a continuous sensation."

The new technology may spawn numerous engineering applications, such as remote assembly, test and inspection, in addition to simultaneous product design. "The applications of this technique, if it succeeds, are vast," claims Slater. "There are possible applications in telemedicine and training for designers, artists and architects. Tasks requiring manual dexterity could be rehearsed in advance of executing them. It enhances the sense of being together even though the physical distances involved are vast."

Mandayam Srinivasan, director of MIT’s Touch Lab, envisions a group of artists from around the world collaborating on a virtual sculpture. "They could create different forms, colors, sounds and textures accessible over the Internet," he points out.

"We really don’t know all of the potential applications," adds Srinivasan. "Just like Bell didn’t anticipate all of the applications for the telephone."

However, there are still some technical problems that must be solved before everyday applications will become available. For instance, the researchers need to reduce the element of delay—or latency—in sending large chunks of data through the Internet and receiving it promptly at the other end. There is a long time delay, due to Internet traffic, between when one user "touches" the on-screen box and when the second user feels the resulting force.

"Each user must do the task very slowly or the synchronization is lost," warns Srinivasan. "In that circumstance, the box vibrates both visually and to the touch, making the task much more difficult."

A one-way trip from hand to brain takes about 30 milliseconds. That same trip from Boston to London takes 150 to 200 milliseconds, depending on network traffic. "If the Internet time delays are reduced to values less than the time delay between the brain and hand, the Internet task would feel very natural," claims Srinivasan.