A New Spatial Index Structure for Efficient Query Processing in Location Based Services Speaker:...

Post on 18-Jan-2016

215 views 0 download

Tags:

Transcript of A New Spatial Index Structure for Efficient Query Processing in Location Based Services Speaker:...

1

A New Spatial Index Structure for Efficient Query Processing in Location Based Services

Speaker: Yihao JhangAdviser: Yuling Hsueh

2010 IEEE International Conference on Sensor Networks, Ubiquitous, and Trustworthy Computing

2

Outline

• Introduction• Related work

– Grid Index– B+-tree

• ISGrid• Query Processing• Experiment• Conclusion

3

Introduction

• A new spatial index structure.• ISGrid provides better efficient query

processing than R-tree.• ISGrid is a grid structure that

provides direct accesses to data and uses Minimum Boundary Rectangle(MBR) as a leaf node.

4

Grid index

• Grid is a regular tessellation of a 2-D surface that divides it into a series of contiguous cells, which can then be assigned unique identifiers and used for spatial indexing purposes.

5

B+-tree

• B+-tree is a tree structure. It usually employed in database or file operating system.

• It has the link to point to the closer data and allow quick sequence read the data.

6

ISGrid

• Configuration of ISGrid

7

ISGrid(cont.)

8

ISGrid(cont.)

• How to choose neighbor nodes?– Traditional: the order of the distance. (x)– Best method: Voronoi Diagram

9

Query Processing

• k-NN Queries– STEP 1: Searching the nearest leaf node

to the query point using the grid index.– STEP 2: Searching the k-NNs through

visiting the neighbor node entry.

10

Query Processing(cont.)

STEP1

STEP2

11

Query Processing(cont.)

• Range Queries– STEP1: Searching the nearest leaf node

to the query point using the grid index.– STEP2: Searching the objects within a

certain range using the neighbor node information.

12

Query Processing(cont.)

STEP1

STEP2

13

Experiment

• Performance of k-NN query processing.

14

Experiment(cont.)

• Performance of continuous k-NN by CNNS.

15

Conclusions

• Authors proposed an index structure, called ISGrid.

• ISGrid provides efficient continuous k-NN query processing in the environment for static objects and moving queries.

16

Thank you for Listening!