Research On Reader Anti-collision Algorithms For Dense RFID System

Radio Frequency Identification Technology(RFID), a solid IoT(Internet of things) technology, has been widely used in recent years. Quantities of tags and readers are deployed in a big RFID system. In order to avoid the situation that some tags cannot be read, the readers should be deployed closely and then the reading region will overlap, which will cause reader collision problems. With RFID technology used more and more widely, the research of anti-collision readers becomes increasingly significant. Meanwhile, it will be much easier to realize the reader anti-collision algorithms, with the data processing and communication capabilities growing of readers.Firstly, this paper described the operating principle of the RFID system, then introduced the existing reader anti-collision algorithms, the Centralized Algorithms NFRA is introduced detailedly. The NFRA algorithm is high efficient and practicable, but the throughput and fair properties can still be improved. GDRA is an improvement algorithm of NFRA, which improves the throughput by exploiting the Sift geometric probability distribution function. However, the fair property of GDRA is still poor.And then the paper proposed the EGDRA(Enhanced GDRA) algorithm, an improvement algorithm of GDRA, which is fairer than GDRA. The EGDRA algorithm contains two improved mechanisms, the classification and Beacon collision improving. The former mechanism improves the fairness of the algorithm and reduces the throughput(number of readers work in the system per second); the second mechanism make up the shortage. The classification mechanism grades readers into different priorities and the readers with a higher level will have a greater probability of getting a smaller slot value and competing for the channels. It can improve the fairness of the algorithm via increasing the level of the readers, which falls in failure in last round of the competition. The Beacon collision improving mechanism contains two methods. In the first method: each reader will give a number to itself and the ones with an even number will get a supplementary chance to compete for the channels after the Beacon collision. During such a supplementary competition, the probability of reaching the channels have an increase of two times, thereby increasing the throughput rate of the system. In the second method: the server will send the(K+1)-th OC signal again after sending K OC signals, and then the readers without completing for the channel and receiving signals in the system. During this OC cycle, the system check whether the detection channel is busy or not, and if not busy, the readers will get access to the channels.Finally, this paper gave a comparison of simulation results of the EGDRA, and GDRA?NFRA algorithms. Simulation results showed that GDRA algorithm has an higher throughput than NFRA algorithm; EGDRA algorithm is more fair than NFRA and GDRA algorithms.