Weighted insert scheme based on R?-tree for tracing moving objects with an RFID tag

To track and trace large amount of moving objects, index structure is needed. A previous study proposed an RFID index structure based on R*-tree. The trajectories of tags are modeled as a line segment in a three-dimensional domain and each domain is the reader identifier, the tag identifier and the time. But the previous study has limitation. R*-tree is suitable for the query that the range ratio of x and y is similar, so the construction of MBR is square-shaped. But in RFID environment, the query region is very narrow because the relative size of the tag identifier is very small than reader identifier. Because of square-shaped MBR, many node accesses are occurred at answering RFID query. We solve the problem by weighted insertion method. Simple describing of the main idea is making the MBR similar to the query region. This method reduces the cost of node accesses by minimizing of overlapped node of the MBRs and the query. With the experiment result, we show that proposed method considerably improves the performance of queries.