Affiliations: Department of Computer Engineering, Pusan National University, Busan, Republic of Korea
Note: [] Corresponding author: Dr. Gihong Kim and Dr. Bonghee Hong, Department of Computer Engineering, Pusan National University, Busan, Republic of Korea. E-mails: [email protected] (Gihong Kim); [email protected] (Bonghee Hong).
Note: [] Corresponding author: Dr. Gihong Kim and Dr. Bonghee Hong, Department of Computer Engineering, Pusan National University, Busan, Republic of Korea. E-mails: [email protected] (Gihong Kim); [email protected] (Bonghee Hong).
Abstract: In RFID systems, it is important to track and trace locations of tags attached to physical objectsin many applications such as automated manufacturing, inventory tracking and supply chain management. Queries for tracing tag locations are among the most challenging requirements in RFID based applications. For efficient query processing, trajectories of tags can be modeled and indexed as a line in a three-dimensional domain, with the axes being the tag identifier, the reader identifier, and the time. In a different way of a moving object index, the ranges of coordinates for each domain are quite different so that the distribution of query regions is biased. Since this biased distribution of query regions produces many overlaps with index nodes, many node accesses are occurred at answering queries. To reduce unnecessary overlaps, the splitting scheme of an index should be changed. Previous indexes for tag trajectories, however, do not consider the biased distribution of query regions. To solve this problem, we propose a new biased splitting method based on R*-tree. For efficient insertion of tag trajectories, our method derives the weighted margin for each node by using weights of each axis and margin of nodes. With the weighted margin, we can choose the subtree which has the minimum cost for inserting trajectories and splitting nodes. Proposed splitting method reduces the cost of region query by reducing overlapped area of query region and MBRs. We also evaluate the performance of our method and compare it with the previous methods. Our experiments show that the index based on the proposed splitting method considerably improves the performance of queries than the index based on the previous methods.
