MIT’s RFly system uses autonomous drones to relay signals emitted by a standard RFID reader to track inventory. (Image credit MIT)


Implementing RFID in supply chain management was supposed to make tracking inventory a whole lot easier; however, in 2013 Walmart reported a $3-billion loss due to product mismanagement. Even the US Army suffered warehouse inventory losses to the tune of $5.8-billion between 2003 and 2011. A 2016 DoD Audit also found the Army lost $1-billion worth of weapons and equipment in Iraq and still have no idea where it went due to poor tracking.


MIT may have just solved that costly inventory bleed by utilizing drones to take over the tracking process using a novel approach to onboard RFID they’ve codenamed RFly. Their new system allows small, safe drones to fly around and read RFID tags and their locations from tens of meters away with an average error in recognition of around 19-centimeters.


The research team encountered several notable issues during the development of their RFly system- most notably drone size. Most drones that can be safely operated among humans are on the small side so that they won’t inflict any damage; this makes them too small to carry a RFID reader. To overcome this issue, they used the drones themselves to relay signals emitted by a standard RFID reader to track the inventory.


Not only would this fix the safety problem posed by using large drones, but it also means the drones could be deployed with existing RFID systems already in place without the need for new tags, readers or even software, two bird with one stone.


This fix however, created additional obstacles to overcome- considering RFID tags are powered wirelessly by the reader, both transmit the same frequency simultaneously. Throwing a relay system in the mix compounds the problem- you now have two other frequencies fighting to be king of the hill, making it a foursome in a system battle royal.


Now add to that the issue of finding or localizing the RFID tags and the problem grows bigger as the platform uses an antenna array to do so. If those antennas are clustered together, a signal broadcasted at an angle will result in different arrival times, meaning the signals transmitted to the array will be slightly out of phase. It’s from those phase differences that the software can locate where the transmission originated from, which is key for the drones.


Since the drones are constantly moving and taking readings at different time increments from different locations, it simulates that multi-antenna array, providing the ability to effectively grab signal location. To separate the signals (those emitted from the reader and tags), the researchers outfitted the drones with an analog filter. The low frequency emitted from the tag is then coupled with the base frequency resulting in location identification.


It’s the researchers hope that their new RFly system will be deployed in large warehouses for continuous monitoring of product inventory to prevent inventory mismatches and loss while allowing employees to focus more on customer demands.   


Have a story tip? Message me at: cabe(at)element14(dot)com