MIT's New System Utilizes UAVs and RFID Tags to Track Inventory [Full Text]

Trends in science and technology At present, logistics companies have launched a full-scale competition in the area of ​​“no human logistics”, and the competition between Ali and Jingdong is particularly fierce. In early August, JD.com announced the opening of an unmanned sorting center in Kunshan, Jiangsu Province. Its biggest feature was the unmanned operation of the entire process, from the supply of bags to loading. At the same time, Ali's rookie network also announced that its largest robot warehouse in China has been put into use in Huiyang, Guangdong.

Even the best companies, tracking inventory in modern distribution warehouses is a huge challenge. According to reports, U.S. retailers have lost more than 45 billion U.S. dollars each year, so developing technologies that make it easier to track inventory can have a huge impact. Now a new system developed by MIT can use small drones and RFID tags to track inventory.

Data Map: Drone

The MIT research team stated that the biggest challenge in developing this drone is to find a safe method. The only drone that can fly close to where people work is a small, lightweight, plastic rotor. Unfortunately, these drones cannot carry RFID readers that carry high radio frequencies.

Instead of relying on drones to carry readers, researchers have developed a method of using drones as relays, allowing drones to be used without the development of new RFID tags, tag readers, or software. But this method is not without challenges.

The research team found that determining the precise location of a product by passing radio frequency identification signals can create problems in signal processing. The RFID tag is wirelessly powered by the reader, which means that the transmitter and tag operate at the same frequency. Relays and RFID tags also need to maintain mutual transmission at the same frequency. The simultaneous transmission requirement causes mutual interference between the four channels.

However, drones are constantly moving, so the time of this process makes it impossible to accurately record the position of an object. To avoid this problem, researchers have developed an analog filter that separates the signal from the tag.

MIT researchers stated that the problem is made worse by the need to accurately locate RFID tags in their systems. Positioning detection systems traditionally use antenna arrays to calculate the source of RFID signals, but drones are too small to carry. The solution to this problem is simple: because it is constantly moving, readings taken at different times will be associated with different locations, basically simulating the effects of traditional antenna arrays.

MIT researchers installed RFID tags for each drone. The drone alternates between the signal of the item to be marked and its positioning signal, and sends its own signal to the reader so that the system can calculate how much the movement of the drone has affected the result, and Correct it.

In one set of experiments, the average positioning error of this system was about 19 cm. Researchers will work with a retailer in Massachusetts to conduct the second phase of the trial.

The original title drone with RFID relay ensures more accurate inventory tracking