While new technologies often evolve from a spark of innovation, more often they come about from discovering new ways to use existing technology. Researchers at MIT’s Media Lab have done just that. They presented their paper on the use of radio-frequency identification (RFID) tags to assist robots in locating items with precision at the 2019 USENIX Symposium on Networked Systems Design and Implementation.
The team has worked for years on the use of radio signals for tracking and identification in areas such as bottled food contamination and device communication in the human body. That research led them to develop a system, TurboTrack, which enables robots to quickly and easily locate and track moving objects.
“If you use RF signals for tasks typically done using computer vision, not only do you enable robots to do human things but you can also enable them to do superhuman things,” said Fadel Adib, MIT Media Lab assistant professor and principal investigator and founding director of the Signal Kinetics Research Group. “And you can do it in a scalable way because these RFID tags are only 3 cents each.”
While the team was well-versed in using RFID, they still faced a challenge in ensuring simultaneous speed and accuracy. They honed in on an imaging technique, super-resolution imaging, to solve this problem. This technique combines images from multiple angles to create a higher-resolution image.
“The idea was to apply these super-resolution systems to radio signals,” Adib said. “As something moves, you get more perspectives in tracking it, so you can exploit the movement for accuracy.”
Their system features an RFID reader that transmits a wideband wireless signal across multiple frequencies. These signals reflect of the RFID tag and nearby objects and bounce back to the system. Within those signals is one specific to the tag.
The team developed an algorithm, called space-time super-resolution, that searches through the signals to find the RFID tag’s response. When the tag moves, which results in an altered signal angle, the algorithm uses that angle change to track the tag’s distance as the signal moves. It then monitors distance measurement changes to determine the RFID tag’s movement. On average, the system enables a robot to locate a tagged object within 71.5 milliseconds with less than a centimeter of error.
This speed and accuracy has the potential to be beneficial in a multitude of applications. Since computer vision has the potential to miss objects or create confusion when faced with a cluttered environment, using radio frequency signals could assist robots working in manufacturing areas with numerous obstacles or drones during search-and-rescue missions to increase collaboration and control.
“You could enable a swarm of nanodrones to form in certain ways, fly into cluttered environments, and even environments hidden from sight, with great precision,” said Zhihong Luo, graduate student in the Signal Kinetics Research Group and the paper’s first author.
Interested in more robotic innovations? Check out Virtual Machine Learning Technique Could Help Robots Perform Better and Engineers Create Robot That Imagines Itself.