The system could enable greater collaboration and precision from robots working in packaging and assembly, and from swarms of drones performing search and rescue missions.
In a paper presented at the USENIX Symposium on Networked Systems Design and Implementation, researchers show that robots using the system can locate tagged objects in an average of 7.5 milliseconds, with an error of less than one centimeter.
In the system, called TurboTrack, an RFID (radio-frequency identification) tag can be applied to any object. A reader sends out a wireless signal that reflects off the RFID tag and other nearby objects, and bounces back to the reader. An algorithm reviews all the reflected signals to find the response from the RFID tag. The final calculations then take advantage of the RFID tag's movement, although this generally decreases accuracy, to improve localization precision.
The researchers say the system could replace computer vision for some robotic tasks. Just like its human counterpart, computer vision is limited by what it can see, and it can fail to notice objects in cluttered environments. Radio frequency signals have no such limitations: They can identify targets without visual input, within clutter, and through walls.
To validate the system, researchers attached an RFID tag to a bottle cap and another to a bottle. A robotic arm located the cap and placed it on the bottle, which was held by another robotic arm. In another demonstration, the researchers tracked RFID-equipped nanodrones during docking, maneuvering, and flight. In both tasks, the system was as accurate and fast as traditional computer vision systems, while working in scenarios where computer vision fails, the researchers report.
“If you use RF signals for tasks that are typically done with computer vision, you not only enable robots to do human things, but you can also enable them to do superhuman things,” says Fadel Adib, assistant professor and principal investigator at the MIT Media Lab, and founding director of the Signal Kinetics Research Group. "And it can be done in a scalable way, because these RFID tags only cost 3 cents each."
In manufacturing, the system could allow robotic arms to be more precise and versatile in, say, picking, assembling, and packing items along an assembly line. Another promising application is the use of handheld "nanodrones" for search and rescue missions. Nanodrones currently use computer vision and image stitching methods for location purposes. These drones often get confused in chaotic areas, lose each other behind walls, and cannot be uniquely identified. All of this limits their ability to, say, spread out over an area and collaborate to search for a missing person. Using the researchers' system, swarms of nanodrones could better locate each other, for greater control and collaboration.
"A swarm of nanodrones can be allowed to form in certain ways, fly in crowded environments, and even in environments hidden from view, with great precision," says first author Zhihong Luo, a graduate student in the Signal Kinetics Research Group.
The other co-authors from the Media Lab in the paper are student Qiping Zhang, postdoctoral fellow Yunfei Ma, and research assistant Manish Singh.

Super Resolution:
Adib's group has been working for years on using radio signals for tracking and identification purposes, such as detecting contamination in bottled food, communicating with internal body devices, and managing warehouse inventory.
Similar systems have attempted to use RFID tags for location tasks. But these come with trade-offs of accuracy or speed. To be accurate, it can take several seconds to find a moving object; as speed increases, accuracy decreases.
The challenge was to achieve both speed and accuracy simultaneously. To do this, the researchers drew inspiration from an imaging technique called "super-resolution imaging." These systems stitch together images from multiple angles to achieve a finer-resolution image.
"The idea was to apply these super-resolution systems to radio signals," Adib says. "As something moves, you get more perspectives from which to track it, so you can exploit the movement for accuracy."
The system combines a standard RFID reader with an "auxiliary" component used to locate the radio frequency signals. The assistant fires a broadband signal comprising multiple frequencies, based on a modulation scheme used in wireless communication, called orthogonal frequency division multiplexing.

The system captures all signals that bounce off objects in the environment, including the RFID tag. One of those signals carries a signal specific to the RFID tag because RFID signals reflect and absorb an incoming signal in a particular pattern, corresponding to bits of 0 and 1, which the system can recognize.
Because these signals travel at the speed of light, the system can calculate a "time of flight"—a measure of distance—to determine the location of the tag, as well as other objects in the environment. However, this only provides a rough estimate of the location, not sub-centimeter accuracy.

Leveraging Movement:
To narrow down the tag's location, the researchers developed what they call a "space-time super-resolution" algorithm.
The algorithm combines location estimates for all the bouncing signals, including the RFID signal, which it determined using time-of-flight. Using some probability calculations, it narrows that group down to a few potential locations for the RFID tag.
As the tag moves, its signal angle changes slightly, a change that also corresponds to a specific location. The algorithm can use that angle change to track the tag's distance as it moves. By constantly comparing the change in distance measurement with all the other distance measurements from other signals, it can find the tag in three-dimensional space. All of this happens in a fraction of a second.
"The high-level idea is that by combining these measurements in time and space, you get a better reconstruction of the tag's position," says Adib.
The work was sponsored, in part, by the National Science Foundation.

Author: Rob Matheson, MIT News Office