A new system developed in MIT uses RFID tags to help home-based robots move around with unprecedented speed and precision. The system could provide greater collaboration and precision in the work of the packaging and assembly robots, and the drones that carry out search and rescue missions.
In a paper presented next week at the USENIX Symposium on Design and Implementation of Network Systems Researchers have shown that works using this system can find tags with a mean of 7.5 milliseconds and an error of less than a centimeter .
In a system called TurboTrack, the RFID tag (radio frequency identification) can be applied to any object. The reader sends a wireless signal that reflects the RFID tag and other nearby objects and restores the reader. The algorithm sift all the reflected signals to find the RFID tag response. Final calculations then affect the movement of RFID tags ̵
Researchers say the system can replace the computer vision for some robotics tasks. Like its human counterpart, the computer vision is limited to what it can see and it may not notice objects in intricate environments. Radio Frequency Signals do not have the following limitations: they can identify goals without visualization, in disarray and through walls.
To test the system, the researchers attached one RFID tag to the lid and the other to the bottle. The robotic hand placed the cap and placed it on a bottle held by another robotic hand. In another demonstration, researchers tracked RFID-equipped nanodrons during docking, maneuvering, and flying. In both cases, the system was as precise and fast as traditional computer vision systems, working in scenarios where the computer vision fails, reporters say.
"If you use RF signals for tasks that are usually performed using a computer view, not only do you allow human beings to do the work, but you can also allow them to do superhuman things," says Fadel Adib , associate professor and chief research officer at MIT Media, as well as the founding director of the Signal Kinetics Research Group. "And you can do this in a scalable way, because these RFID tags are only 3 cents each."
In the production system, the system could allow robotics be more precise and versatile, ska The other promising program is the use of portable "nanodromes" for search and rescue missions. Nanodrons are currently using computer vision and methods for linking image images for localization purposes. These drones are often confused in chaotic areas, they lose each other behind the walls and can not uniquely identify each other. All this limits their ability, say, to spread across the territory and cooperate to search for the missing person. Using a researcher system, nanodrons can better position each other for greater control and collaboration. with great precision, "says Zhihong Luo, a postgraduate student for kinetic signal studies.
Other collaborators at Media Lab visit the student Qiping Zhang, postdoc Yunfei Ma, as well as researcher Manish Singh.
Adiba Group has been working on the use of radio signals for tracking and identification purposes for many years, such as detecting contamination in food bottles, communicating with devices inside the body and managing stockpiles
Similar systems tried to use RFID tags for localization tasks. But they come with compromises in accuracy or speed. To be accurate, it may take a few seconds to locate a moving item; To increase speed, they lose accuracy.
The task was to simultaneously achieve speed and accuracy. To do this, the researchers drew inspiration from the technique of visualization, which is called "super department image". These systems cross-image images from multiple angles to achieve a clearer resolution of the image.
"The idea was to apply these super-resolution systems to radio signals," says Adib. "When things are moving, you get more prospects in tracking it, so you can use the movement for accuracy."
The system combines a standard RFID reader with a "helper" component, which is used to localize radio frequency signals. An assistant shoots a broadband multiband signal based on the modulation scheme used in the wireless communication, called orthogonal frequency division multiplexing. One of these signals carries a signal that is specific to a specific RFID tag, since RFID signals reflect and absorb the input signal in a particular pattern corresponding to the bits 0s and 1s that the system can recognize.
At light speed, the system can calculate "flight time" – measuring distance by calculating the time required for the signal movement between the transmitter and the receiver – to determine the location of the label, as well as other objects in the environment. But this gives only the number of localization localization, not the accuracy of the subconscious.
Use of motion
The algorithm combines placement estimates for all rebound signals, including the RFID signal, which it determines using flight time. Using some probabilistic calculations, it limits this group to several potential locations for the RFID tag. The algorithm can use this angle change to track the distance of the tag as it moves. Constantly comparing that by changing the distance measurement to all other measurements of the distance from other signals, he can find a label in a three-dimensional space. All this happens for a fraction of a second.
"The idea of a high level is that by combining these measurements over time and over space, you get a better reconstruction of the label's position," says Adib. The work was sponsored, in particular, by the National Science Foundation