| While RFID technology is widely used in people’s life, in addition to facing the issues of high system cost, diverse technical standards, immature application environment and etc, it also has many urgent problems to be solved in antenna design, secure communications, privacy protection, anti-collision and many other key technologies. When a reader communicates with multiple tags around, data collision between different tags is often generated by channel competition. So how to effectively avoid such collisions from happening will be of great significance to ensure the normal communication between the reader and tags in RFID system.This thesis focuses on researching the tag anti-collision technology in RFID system. Firstly, the existing tag anti-collision algorithms are described in detail by dividing them into three categories. Then, after enumerating and analyzing the working principle and pros and cons of some typical algorithms in each type of method, the advantages and disadvantages of three kinds of algorithms are compared.Secondly, this thesis mainly studies the dynamic frame slotted aloha(DFSA) algorithm in aloha-uncertainty method. The importance of tag estimation in DFSA is pointed out through detailed analysis of system model of DFSA algorithm. In order to estimate for the number of identification tags more accurately and efficiently, the relationship between the average number of non-empty slots and the number of tags to be identified is found on the basis of researching the existing tag estimation method, then a new tag estimation algorithm based on the number of non-empty slots(NESB) is proposed. The simulation and comparison analysis show that when the frame is 128 long, identification tag number is 300, NESB algorithm is 87.74% lower than Cratio estimation method and is 83.13% lower than Adaptive Slotted ALOHA Protocol(ASAP) algorithm in terms of estimation error rate. What’s more, the estimated time cost by NESB algorithm is decreased by 25.49% and 14.03% respectively compared with Maximum a posteriori probability(MAP) and ASAP estimation method. Furthermore, in order to verify the practical feasibility of NESB estimation method, it is applied into DFSA algorithm so that an improved DFSA algorithm is proposed. Next, an RFID communication system model is built to simulate the improved DFSA algorithm in OPNET. The simulation results show that the performance of improved DFSA algorithm is better than before; its throughput can always stay near the theoretical maximum.Finally, this thesis analyzes the query tree(QT) algorithm and its improved algorithm like the quaternary query tree algorithm and the adaptive multi-tree algorithm. A new algorithm called adaptive two- quad hybrid query tree is proposed for improving the shortage of above mentioned algorithms. As the empty slots generated in the process of tag recognition are completely eliminated, the system’s query times and throughput in the case of smaller number of tags are both optimized by the new algorithm. Besides, the amount of information transmitted between the reader and the tag is also reduced by the proposed algorithm. Simulation results show that, under the condition of which identification tag number is less than 60% of the maximum number of tags, the performance of the proposed algorithm is superior to the QT algorithm and the quaternary query tree algorithm. When the tag ID number is 8 bits occupied, the number of identification tag is 50, by comparing the proposed algorithm with the other two algorithms, the number of queries is decreased by 24.35% and 25.64% respectively while the throughput is increased by 32.18% and 34.66% respectively, and the amount of information transmission were reduced by 53.9% and 50.3% respectively. |
