Research On Anti-collision Algorithm In RFID System

As a widely used non-contact automatic identification technology, RFID technology will get a huge development space under the guidance of the Internet of things. Compared to other automatic identification technology, RFID technology have the advantages of non-contact, strong ability to adapt to the environment, non-visual range reader, which is widely used in industry, military technology, transportation,health care and daily life and so on, has a very broad application prospects.However, due to the extensive use of RFID technology, has also had some problems in multiple target recognition process: when multiple RFID tags request identification, the reader can not automatically sort clear distinction, it causes information collision and the reader miscalculation or failure, which affecting the normal work of the RFID system. Therefore, information collision issue must be solved carefully in the process of the development of RFID technology. On the premise of controlling tags cost and complexity of the algorithm, the design of anti-collision algorithm focus is reducing recognition time, improving RFID system throughput and recognition efficiency.The paper first introduces the composition and the working principle of RFID system, and then it focuses on the tag anti-collision strategies of RFID system,including two aspects, one is probabilistic anti-collision algorithms based on ALOHA and the other is determined anti-collision algorithms based on binary tree. The anti-collision algorithms based on ALOHA mainly includes pure ALOHA, slotted ALOHA, frame slotted ALOHA and dynamic frame slotted ALOHA algorithm. The latter algorithm made better than the former one. In view of the dynamic frame slotted ALOHA algorithm needs to accurately estimate the the tag number, several common tag estimation method was introduced in detail. The anti-collision algorithms based on binary tree mainly covers binary search algorithm, dynamic search algorithm, capturing backward search algorithm and binary bound search algorithm. The implementation procedures on above divers algorithms are detailed analyzed and the performances are full compared.At the end of this thesis, in order to solve the problem that the binary tree algorithm search time is big and takes long time to recognize the tags, an improvedalgorithm is proposed which is based on the binary bound search algorithm. According to the characteristics of the sequence number, the electronic tags are divided into two groups, and only one group can be recognized each time, so the search time can be reduced. The reader send command parameters with the highest collision only, which can reduced the amount of communication. The simulations indicate that the improved algorithm can reduced search time and the amount of communication. As the number of tags increases, the more obvious advantages of the improved algorithm.