Research on Anti-collision Algorithm Optimization of RFID Tag Based on Binary Search

Research on Anti-collision Algorithm Optimization of RFID Tag Based on Binary Search Jinyan Liu, Quanyuan Feng School of Information Science and Technology, Southwest Jiaotong University, Chengdu 610031, China Abstract This paper explores the optimization path of anti-collision algorithm for RFID system based on binary search. The structure and working principle of the RFID system are firstly analyzed, and then the binary anti-collision algorithms are explored, pertinent to tag collision in the RFID system. Based on the exploration of the operation principle of the basic binary search algorithm, the dynamic binary search algorithm and the bit-locking backward binary search algorithm, it can be learnt that the basic binary search algorithm has the disadvantages of data redundancy and relatively large number of queries and that the latter two can effectively make up the deficiencies of the basic binary search method. Finally, the jumping binary search algorithm is proposed based on the summarization and analysis of the two search algorithms with its realization process discussed and its related characteristics compared and analyzed. Keywords: Binary search, Rfid, Tag Anti-Collision, Radio frequency. 1. RESEARCH STATUS OF RFID TECHNOLOGY 1.1 The application status of RFID technology Radio frequency identification technology (RFID technology) originated in the research field of military radar of the British Air Force in the 1940s. In the 1960s, the theory of RFID technology has rapidly developed and gradually adjusted to the application fields. Britain, the United States, Japan and Switzerland are at the worldleading level in research and application of RFID technology, especially in aspects of chip technology development and electronic label technology. China s RFID technology research starts relatively late but develops rapidly, and the RFID technology has been gradually applied to various fields of society, attracting great concern of the economic market. The application of RFID technology in payment for bus and city subway by swiping cell phone greatly facilitates the traveling of the citizens. And in Shanghai World Expo, people can enter the Expo Park and buy food by swiping cell phone. RFID technology plays an important role in user identification, asset management and goods tracking, contributing to its great application value in the fields of economy, military, medicine, logistics transportation, objects storage and management. (Figure 1) For paper book management in libraries, the RFID technology can be applied to accurately locate every book so that the misplaced books can be immediately discovered and the reasonable location can be accurately determined. (Figure 2) RFID technology can be used to automatically scan the goods in the shipping cart and quickly generate the bill on the cell phone of the consumers for them to confirm and settle. Figure 1. RFID warehouse logistics management system technology 15

1.2 Research status of anti-collision algorithm Figure 2. Intelligent library RFID technology plays an increasingly important role in the various fields of society and in people's daily life, and the promotion of its market value in turn contributes to the development of the research on RFID technology. The key direction of RFID technology research is anti-collision algorithm, and the crucial issue of RFID technology research is to solve the long-standing communication collision problem in the application of RFID technology and thus to ensure the integrity and correctness of communication data. The communication collision problems in the applications of RFID technology are mainly caused by the simultaneous multi-label and multi-reader occupation of the communication channels and the complex external interference factors. The multi-access mode is the major way to realize the anti-collision algorithm of RFID technology both at home and abroad, which includes SDMA (space division multiple access), TDMA (time division multiple access) and CDMA (code division multiple access). The current common RFID anti-collision algorithm is based on the TDMA binary tree anti-collision algorithm and its improved algorithm. 2. SYSTEM STRUCTURE OF RFID The RFID system mainly consists of electronic tags, computer communication networks and readers, and the structure of the RFID system differs according to the specific application fields. The electronic tags, computer communication networks and readers are the core components of the RFID system, which play an indispensable core role in any RFID system architecture. By optimizing the coordination relationship of the three, the RFID system can operate more efficiently. 2.1 The classification and structure functions of electronic tags The RFID tag with a specific electronic code is the carrier of data in the system. The association of the electronic tag with the identified object determines the unique and specific identity of the latter, enabling the identified object to be systematically managed. The electronic tags store data through the chip and realize data transmission though wireless antenna. The information reception and feedback function enables electronic tags to be applied in more areas. Electronic tags can be divided into active tags and passive tags, among which the former with power supply can play an active role and exert a powerful application effect in a particular environment and the latter without built-in power supply is unable to transmit data actively. Though incapable of active data transmission, the passive tag has more extensive scopes of application because of its flexibility. The electronic tags can also be divided into high-frequency electronic tags, low-frequency electronic tags and microwave electronic tags according to the operating frequency, among which the microwave electronic tags are the widely-used type of electronic tags. The passive electronic tags will present four varying states during the operation of the RFID system, i.e. the ready state, selected state, off-power state and passive state. (Figure 3) The electronic tags have the following functional characteristics: First, the electronic tags carry information in the RFID system. Both the active electronic tags and the passive electronic tags must be equipped with the function of information storage. The storage of the specific information of objects forms the electronic tags with independent and unique information characteristic, which enables the objects to be equipped with specific identity characteristics. 16

Second, the article information contained in the electronic tags is of great application significance, which requires that the electronic tags should not only store the article information safely and completely but also modify or delete the information if necessary. Third, the chip as a key part of the electronic tag is responsible for the acceptance and decoding of signals and decoding of information as well as encoding on the return signals. 2.2 The structure and function of readers Figure 3. Electronic label status switching Readers are mainly responsible for the two-way data transmission with the electronic tags in the RFID system, and the management, storage, control and other instructions can be transmitted to the reader through the computer communication equipment. In the operating process of the RFID system, the readers can realize the read-in and read-out of the electronic tag data through the data transmission with the tags. The RFID system realizes effective operation due to the data transmission and identification functions of readers. The different operating frequency of the readers determines their different effective recognition distances. The reader is mainly composed of the RF interface, the antenna and the logic control unit. The RF unit of the reader is the transmitting terminal of energy, emitting energy of different frequencies during its operation and carrying out logic modulation on the received signals. The logic unit of the reader mainly functions to read and write the data information and to establish data connections with the application system. The main features of the reader are: First, there are two separate signal channels inside the reader. One is the channel for signal transmission from the transmitter to the electronic tags, and the other is the data channel for receiving the electronic tag data information. The non-interference signal transmission of the two channels enables information exchange to be achieved between the reader and the electronic tags. Second, the reader can receive the instructions sent from the computer end, according to the data contents of which certain operations will be conducted. In this process, the reader can also operate on the electronic tags through the instructions of the computer end. Third, the antenna of the reader can not only convert electromagnetic waves into a specific current signal but also transform the electric current into electromagnetic waves which will be transmitted to the passive electronic tags. The passive electronic tags can only be recognized in the RFID system by accepting signals from the reader antenna. Fourth, the control module of the reader is a key part to realize the information interaction between the electronic tag and the reader, and it can also decode the data information. The reader in the RFID system mainly functions to identify and transmit data and signals, which process must be completed by the control module. The core task of the control module is to control the information interaction between the electronic tag and the reader, thus realizing the anti-collision algorithm of the reader. 17

3. THE BINARY SEARCH-BASED ALGORITHMS The binary tree algorithm is deterministic. The core of the anti-collision principle of binary tree algorithm is to divide the collided tags into two subsets of 1 and 0. If tag collision occurs during the query of the subset 0, the tag will be further divided into two subsets of 01 and 00, on which the query will be re-conducted. Whenever collision occurs during the query, the set will be divided for further query. When the subset 0 is recognized without any collision, the query will be conducted on the subset 1 until all the tags are recognized without collision. (Figure 4) 3.1 Binary search algorithm Figure 4.Binary tree collision avoidance algorithm model The binary search algorithm compensates for the low channel utilization of the ALOHA algorithm. Binary search algorithm has significant superiority, mainly manifested by the non-necessity for the tag to memorize the previous queries, which greatly reduces the operation links and enhances the operational efficiency. The binary search algorithm only reads and queries a bit prefix; when the query command is the same as the tag serial number, the tag will respond accordingly and send the serial number for comparison with the query demand. During the comparison, the response of a tag indicates the result of correct recognition; in case of the response of several tags, the reader will determine the location of the collision, according to which the query demand will be changed. The reader will recognize all tags during the continuous correction of the query demand. The specific algorithm can be divided into three steps: First, the reader will send query commands based on the data bits, all of which area sequences of 1. Provided that the data bit is 11111111 and that the reader sends the query command of 11111111, the query command will lead to the reaction of the tag receiving the command, which feeds back the serial number to the reader through the information channel. Second, read and compare the REQUEST with the serial number on the tag and examine whether there exists collision bit (Figure 5). Since the serial numbers of tags are variant, a unique tag will be recognized if there is no collision; otherwise, there will be collision of multiple tags, in which case the reader will correct the REQUEST to obtain a new one according to the collision bit by maintaining bits from the highest order to the highest collision bit as 0 and setting the bits from the highest collision bit to the lowest order to be 1. Figure 5. Manchester detects the collision bits 18

Figure 6. Binary search process Third, re-send REQUEST and execute the second step until the tag is correctly recognized. It enters the passive state after the reader reads its data. The algorithm again performs the search from the first step until all the tags are identified. The algorithm flow is shown in Figure 6. 3.2 Dynamic binary search algorithm Binary search algorithm possesses high computational efficiency in practical application. But in cases of long serial number during the binary search recognition process, there will be numerous repeated bits between the electronic tags and the reader, greatly increasing the search time and the possibility of errors in the operation. The dynamic binary search algorithm as the improved version of binary search algorithm not only inherits its superiority but also has it optimized in light of its deficiencies, contributing to the higher success rate and more accurate operation of the anti-collision of RFID tags. The search process of the dynamic binary search algorithm is the same to that of the binary search algorithm except for the difference in the execution of the second step, where only the data from the highest collision bit to the highest order are sent every time the query REQUEST is sent out from the reader. A new REQUEST is obtained without sending the bits 1 from the highest collision bit to the final bit during the non-dynamic search. The tag is not required to send the high-order similar to the query command of the reader. (Table 1) Table 1 Dynamic binary search recognition table First search Second search Third search REQUEST 11111111 10 1010 Tag A 10110010 110010 Tag B 10100011 100011 0011 Tag C 11100011 Collision bit decoding results 1x1x001x 1x001x Identification tag No No B After three times of searches with the dynamic binary search algorithm, the tag B is correctly recognized by the reader. Compared with the basic binary search algorithm, the dynamic binary search algorithm is significantly improved in the efficiency of operation. Under the premise of equal number of searches, the dynamic binary search algorithm can effectively reduce the output of the reader query command, especially the transmission of 19

the tag serial number. Through the simplification of these operations, the dynamic binary search algorithm can greatly shorten the operation time, thus effectively improving its operating efficiency. 3.3 Bit-locking backward dynamic binary search algorithm The data redundancy generated in the search of binary algorithm is the key factor influencing its efficiency. The dynamic binary search algorithm can improve the data redundancy caused by the search process without reducing the number of reader searches, thus enhancing the overall operation efficiency. When a tag is correctly identified, the reader resends the search command to query all the remaining tags, which means the query returns to the root node of the binary tree. This backward binary algorithm can not only effectively improve the data redundancy but also reduce the searches due to the return to the binary tree root. As can be seen in table 2, the backward binary search retains the upper-storey node with the method of grouping during its implementation. It inherits the three steps of the binary search algorithm in terms of operation links, through which B can be accurately recognized. The query REQUEST is reset by taking another subset according to the highest collision bit of group 0 in the table for query of unread tags in the group, i.e. the REQUEST 10111111 can be used to recognize the tag A in group 0. When the tags in group 0 are fully recognized, the query demand of the reader returns to the root node to recognize the tag C in group 1 by resending 11111111. The binary search method requires 6 searches, which indicates that the recognition of n tags requires 2n (log n +1) searches; while the bit-locking backward binary search method only requires 5 searches for 1 tag and (2n-1) searches for recognition of n tags. When the number of tags is larger, the reduction of search times is more evident, as shown in table 3, which can be obtained by calculation and figure 7. Table 2 Lock back binary search table First search Lock grouping Second search Third search REQUEST 11111111 0 and 1 groups 10111111 10101111 Tag A 10110010 10110010 Tag B 10100011 10100011 10100011 Tag C 11100011 The 0 group Collision bit decoding results 1 x 1 x 0 0 1 x 1 0 1 1 0 0 1 0 101x001x 1 0 1 1 0 0 1 1 Identification tag NO The 1 group NO B 1 1 1 1 0 0 1 1 Table 3 Binary algorithm and backward binary algorithm query times Tag count 6 11 16 21 26 31 36 41 46 51 56 Method 1query method 21 49 80 113 148 184 222 260 300 340 381 Method 2query method 11 21 31 41 51 61 71 81 91 101 111 Figure 7. Binary search method and lock back search search times 20

3.4 Jumping dynamic binary search algorithm The dynamic binary search algorithm can improve the efficiency of the operation by reducing the data redundancy during the actual operation, and the bit-locking backward binary search algorithm can enhance the operation efficiency by reducing the number of queries, both of which compensate and improve the deficiencies of the binary search algorithm from different perspectives based on the guarantee of the rationality of the basic operation. The jumping dynamic binary search algorithm can extract and integrate the advantages of the first two excellent binary search algorithms, thus enabling it to become the most advanced algorithm in the RFID tag anti-collision algorithms. It maximizes the operation efficiency under the insurance of operation rationality and reduces the data bits of the reader query command and the tag serial number to be sent on the basis of the execution of the bit-locking backward method. The recognition process of the jumping dynamic search is shown in table 4. Table 4 Jump dynamic search recognition First search Lock grouping Second search Third search REQUEST 11111111 0 and 1 groups 10 1010 Tag A 10110010 110010 Tag B 10100011 100011 0011 Tag C 11100011 The 0 group 1 0 1 1 0 0 1 0 Collision bit decoding results 1 x 1 x 0 0 1 x 1x001x 1 0 1 1 0 0 1 1 Identification tag NO The 1 group NO B 1 1 1 1 0 0 1 1 The three searches in Figure 10 are circularly conducted according to the basic three steps of dynamic binary search. Upon the completion of the first search, the bit-locking grouping is carried out to mark the upper-layer nodes. The tag B is returned to group 0 after being recognized for the implementation of the next two steps. The fourth step: the reader query command is 1011 and returns to group 0; the tag A is correctly recognized by sending the last four bits 0010 apart from the query demand, after which it is commanded to sleep. The fifth step: reset the query demand to be 11111111, and the tag C in group 1 is recognized. The jumping dynamic binary search algorithm requires fewer searches on average than the basic binary search and the bit-locking backward binary search and less data transmission than dynamic search method, and it possesses the overall performance much higher than that of the basic binary search algorithm, the dynamic binary search algorithm and the bitlocking backward binary search algorithm, integrating the advantages of the former three algorithms. 4. CONCLUSION Through the research on the binary search algorithm, it is found that the basic binary search algorithm is characterized by accuracy and especially larger effectiveness in avoiding RFID tag collision compared with ALOHA algorithm. However, the basic binary search algorithm will produce data redundancy in the detection process, and the overall operation efficiency is also influenced because of the relatively large number of queries. In light of the deficiencies of the basic binary search method, the dynamic binary search method is proposed to optimize the problem of data redundancy, and the bit-locking backward binary search method is put forward to solve the problem of large number of queries. The two algorithms both improve the basic binary search algorithm to a certain extent, enhancing the operation efficiency while guaranteeing the operation accuracy. The dynamic binary search algorithm is the most advanced one at present, which inherits the advantages of the dynamic binary algorithm and the bit-locking backward binary search algorithm, solving the problem of data redundancy and greatly reducing the queries of the reader by integrating the advantages of the former two algorithms. They can play a role in the jumping binary search algorithm, and thus greatly optimize the RFID tag anti-collision algorithm, promoting the extensive application and development of RFID technology in social fields. ACKNOWLEDGMENTS This work is supported by the National Natural Science Foundation of China under Grant 61531016. 21

REFERENCES Fan W.J., Zhang S.S., Tian X. (2012). Research on Anti-collision Algorithm of RFID based on backward binary search, Computer Applications and Software, 05, 191-194. Gao J.H, Zheng X.Y. (2012). RFID. Binary search anti-collision algorithm, Computer Measurement and Control, 10, 2754-2756. He X.T., Zheng W.F. (2011). RFID. Binary collision suppression algorithm based on two points superposition, Journal of South China Normal University (NATURAL SCIENCE EDITION), 03, 61-65. Li Z.J., Zhou J., Qiao J., Wu W.J. (2013). Research and optimization of adaptive fractal coding RFID anticollision algorithm, Journal of Communications, 09, 185-190. Mo L. (2010). Research on RFID bit shielded binary search anti-collision algorithm, Journal of Hebei University of Science and Technology, 05, 458-462. Wang L., Jin J., Zhang J. (2011). Jumping binary tree search anti-collision algorithm and its analysis, Microelectronics and Computer, 07, 146-148. Wu J.X., Huang S.Y., Li C.Y., Ho H. (2009). Improved RFID binary search anti-collision algorithm, Computer Engineering and Applications, 05, 229-231, 235. Xia Z.G., He Y.G., Hou Zhou state. (2010). Computer engineering and application of anti-collision algorithm, A Binary Tree Detection Label, Vol. 20, pp. 245-248. Yan X.L., Chen Q.K., Hao J.T. (2013). A new anti-collision algorithm of microcomputer system based on dynamic binary, 09, 2148-2151. Zhang W.X., Aung Z.M., Yin X.Z. (2012). An improved backward binary search RFID multi tag anticollision algorithm, Journal of HeFei University of Technology (NATURAL SCIENCE EDITION), 07, 919-921, 934. 22