1 Indexing the Positions of Continuously Moving Objects Speaker: Chia-Hsiang Hsu Date: 2006/10/16

2 Reference Simonas Saltenis, Christian S. Jensen, ScottT. Leutenegger, Mario A. Lopez: Indexing the Positions of Continuously Moving Objects. SIGMOD Conference 2000: [DBLP: conf/sigmod/SaltenisJLL00]Simonas SaltenisChristian S. JensenScottT. LeuteneggerMario A. Lopez SIGMOD Conference 2000

3 Outline Introduction Problem Statement Structure and Algorithms Performance Experiments Conclusion

4 Introduction The rapid and continued advances in positioning systems, e.g., GPS, wireless communication technologies, and electronics in general promise to render it increasingly feasible to track and record the changing positions of objects capable of continuous movement.

5 Continuous movement poses new challenges to database technology. Capturing continuous movement with this assumption would entail either performing very frequent updates or recording outdated, inaccurate data, neither of which are attractive alternatives.

6 We use one linear function per object, with the parameters of a function being the position and velocity vector of the object at the time the function is reported to the database. Updates are necessary only when the parameters of the functions change.

7 Problem Statement

8

9 Moving Points and Resulting Leaf-Level MBRs

10 Query Types

11 Query Examples for One-Dimensional Data

12 Structure and Algorithms The position of a moving point is represented by a reference position and a corresponding velocity vector-(x,v) in the one dimensional case, where x=x(t ref ). We let t ref be equal to the index creation time, t l.

13 A tradeoff exists between how tightly a bounding rectangle bounds the enclosed moving points or rectangles across time and the storage needed to capture the bounding rectangle. It would be ideal to employ time-parameterized bounding rectangles that are always minimum, but the storage cost appears to be excessive.

14 Minimum Bounding Internals

15 AB Minimum Bounding Internal

16 Update

17

18 Querying

19

20

21

22

23 Performance Experiments With the exception of one experiment, the simulated objects in the scenario move in a region of space with dimensions 1000*1000 kilometers.

24

25

26 Conclusion