songhua xing, cyrus shahabi and bei pan infolab university of southern california los angeles, ca...
TRANSCRIPT
![Page 1: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/1.jpg)
Songhua Xing, Cyrus Shahabi and Bei Pan
InfoLab
University of Southern California
Los Angeles, CA 90089-0781
http://infolab.usc.edu
Continuous Monitoring of Nearest Neighbors on Land Surface
![Page 2: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/2.jpg)
2
Outline
Motivation Related Work Preliminaries Surface Index based Approach Performance Evaluation Conclusion and Future Work
![Page 3: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/3.jpg)
3
Motivation
Yosemite National Park
![Page 4: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/4.jpg)
4
Motivation
How to continuously monitor the animals over surface?
![Page 5: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/5.jpg)
5
Motivation
How to continuously monitor the animals over surface?
Problem To continuously monitor k Nearest Moving Neighbor based on the Surface Distance. A variance of Static Surface k Nearest
Neighbor Query Why is this problem challenging? Why is this problem important?
![Page 6: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/6.jpg)
Motivation
6
Why is this problem challenging? Huge size of surface model
Millions of terrain data for a region of 10km×10km
Triangular Meshes
![Page 7: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/7.jpg)
Why is this problem challenging? Huge size of surface model
Millions of terrain data for a region of 10km×10km
Costly surface distance computation Tens of minutes on a modern PC for a terrain of
10,000 triangular meshes
Motivation
7
Chen-Han Algorithm: O(n2)
*Shortest paths on a polyhedron: CHEN, J., HAN, Y., Computational Geometry 1990
![Page 8: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/8.jpg)
Why is this problem challenging? Huge size of surface model
Millions of terrain data for a region of 10km×10km
Costly surface distance computation Tens of minutes on a modern PC for a terrain of
10,000 triangular meshes
Motivation
8
Chen-Han Algorithm: O(n2)
*Shortest paths on a polyhedron: CHEN, J., HAN, Y., Computational Geometry 1990
![Page 9: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/9.jpg)
Why is this problem challenging? Huge size of surface model
Millions of terrain data for a region of 10km×10km
Costly surface distance computation Tens of minutes on a modern PC for a terrain of
10,000 triangular meshes
Motivation
9
Chen-Han Algorithm: O(n2)
A
B
1
2
3
4
AB
1
2
3
4C
AB
1 2
3 4
A B1 2
3
4
C
Un
fold
ing Case 1
Case 2
Case 3
Case 4 ……
*Shortest paths on a polyhedron: CHEN, J., HAN, Y., Computational Geometry 1990
![Page 10: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/10.jpg)
Why is this problem challenging? Huge size of surface model
Millions of terrain data for a region of 10km×10km
Costly surface distance computation Tens of minutes on a modern PC for a terrain of
10,000 triangular meshes
No efficient surface index structure for moving objects Shortest Path Tree on road network can not apply Existing Surface index structures are designed for
static environment
Motivation
10
![Page 11: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/11.jpg)
Online Map
Geo-realistic Game
Battlefield Simulator
Motivation
11
Why is this problem important? The availability of the data
(imagery, elevation) of Earth surface at very high resolution
The quick geo-realistic rendering of terrain surface on the computer display (e.g., Google Earth™)
The prevalence of GPS equipped sensors
June 2009: the most complete terrain map of the Earth's surface has been published through the collaboration between NASA (USA) and the ASTER (Japan).
99% Coverage1.3 Million Images30-Meter Resolution
![Page 12: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/12.jpg)
Applications
Motivation
12Resource Exploration
Caravan in Desert
Disaster Response
Go Camping
Ski
Mars Exploration
![Page 13: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/13.jpg)
13
Outline
Motivation Related Work Preliminaries Surface Index based Approach Performance Evaluation Conclusion and Future Work
![Page 14: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/14.jpg)
14
Related Work
Conventional kNN
Reverse kNN
Time-aware kNN
Visible kNN
Euclidean Space Road Networks Surface
Spatial Database
kNN Query Processing
![Page 15: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/15.jpg)
15
Related Work
Conventional kNN
Reverse kNN
Time-aware kNN
Visible kNN
NN Query: Roussopoulos et al., SIMGOD 1995
Influences Set: Korn et al., SIMGOD 2000
Efficient Method for Maximizing Bichromatic Reverse Nearest Neighbor: Wong et al., VLDB 2009
Continuous NN Search: Tao et al,. VLDB 2002
Lazy Updates : Cheema et al,. VLDB 2009
VkNN Query: Nutanong et al., DASFAA 2007
Euclidean Space Road Networks Surface
Spatial Database
kNN Query Processing
![Page 16: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/16.jpg)
16
Related Work
Query Processing in SNDB : Papadias et al., VLDB 2003
V-based kNN in SNDB: Shahabi et al., VLDB 2004
Scalable Network Distance Browsing in Spatial Databases: Samet et al., SIGMOD 2008
RNN in Large Graphs: Yiu et al., TKDE 2006
CNN Monitoring in RN: Mouratidis et al., VLDB 2006
Euclidean Space Road Networks Surface
Spatial Database
kNN Query Processing
Conventional kNN
Reverse kNN
Time-aware kNN
![Page 17: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/17.jpg)
17
Related Work
Conventional kNN
SkNN Query : Deng, Zhou ..., ICDE 2006, VLDB J.
Euclidean Space Road Networks Surface
Spatial Database
kNN Query Processing
![Page 18: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/18.jpg)
18
Related Work
Conventional kNN
SkNN Query : Deng, Zhou ..., ICDE 2006, VLDB J.
Indexing Land Surface: Shahabi et al., VLDB 2008
Euclidean Space Road Networks Surface
Spatial Database
kNN Query Processing
![Page 19: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/19.jpg)
19
Outline
Motivation Related Work Preliminaries Surface Index based Approach Performance Evaluation Conclusion and Future Work
![Page 20: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/20.jpg)
20
Preliminaries
Continuous nearest neighbor monitoring in road networks (Mouratidis et al, VLDB 06) Shortest Path Spanning Tree (Dijkstra Tree)
1. Low fan-out2. Tall and thin
BOUNDARY
![Page 21: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/21.jpg)
Observation 1: SE Tree is fat and short.
Observation 2: These shortest surface paths rarely share common edges.
Observation 3: These shortest surface paths do not cross each other.
21
Preliminaries
Surface Expansion Tree (SE Tree)
source
![Page 22: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/22.jpg)
Preliminaries
sp2
p1
p3p4
p5
p6
Old Result Boundary
More incoming objects Original Result
Boundary contains more than k objects
Identify the kth object and shrink the result boundary
New Result Boundary
Continuous Monitoring (k = 3)
![Page 23: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/23.jpg)
Preliminaries Continuous Monitoring (k = 3)
sp2
p1
p3p4
p5
p6
Result Boundary
Expansion Boundary
More outgoing objects Original Result
Boundary contains less than k objects
Compute the Expansion Boundary
Expand SE tree to Expansion boundary and identify the kth object inside the search area
DRAWBACKS Expansion is slow; Search Area may be large; SE Tree is fat and short.
Search Region
![Page 24: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/24.jpg)
24
Outline
Motivation Related Work Preliminaries Surface Index based Approach Performance Evaluation Conclusion and Future Work
![Page 25: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/25.jpg)
Surface Index based Approach
25
Intuition Motivated by Partitioning Shortest Paths on
road networks Scalable Network Distance Browsing in Spatial
Databases: Samet et al., SIGMOD 2008 (best paper award)
X
V3
V1
V2
V4
V5
V6
Is there a way to partition paths on surface?
The shortest path from X to any vertex of one region must be inside that region.
![Page 26: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/26.jpg)
Surface Index based Approach
26
Intuition Motivated by Partitioning Shortest Paths on
road networks Scalable Network Distance Browsing in Spatial
Databases: Samet et al., SIGMOD 2008 (best paper award)
source
Too many partitions
![Page 27: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/27.jpg)
Surface Index based Approach
27
Surface Shortest Path Container Built based on the Observations that shortest
paths do not cross each other and hardly share common edges.
edges opposite to the source point
![Page 28: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/28.jpg)
Surface Index based Approach
Angular Surface Index (ASI) Hierarchically index these containers
ASI Tree
ADVANTAGES Size: Much Smaller Regularity: Well
Balanced Efficiency: Tall and Thin
![Page 29: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/29.jpg)
Surface Index based Approach
Query Processing The core is the same as using SE Tree Search is localized in Containers only
p2
p1
p3
p4
p5
p6s
p2p3
p4
p5
p1
p6s
Search Region
![Page 30: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/30.jpg)
Surface Index based Approach
More Details Definitions/Properties of Containers,
Equidistance Line and ASI-tree Please refer to Section 5.1, 5.2 and 5.3 in the paper.
Constructions of Containers Please refer to Section 5.1.2 in the paper.
Algorithms for Query Processing Please refer to Section 5.4 in the paper.
![Page 31: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/31.jpg)
31
Outline
Motivation Related Work Preliminaries Surface Index based Approach Performance Evaluation Conclusion and Future Work
![Page 32: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/32.jpg)
32
Performance Evaluation Real World Dataset http://data.geocomm.com
Eagle Peak (EP) at Wyoming State, USA 10.7km×14km.
Bearhead (BH) at Washington State, USA Similar size as above.
Bearhead (BH)Eagle Peak (EP)
![Page 33: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/33.jpg)
33
Performance Evaluation The impact of k and Object Density
OBSERVATIONS The ASI based
Algorithm has a much better performance .
Response Time increases as the value of k increases.
Response Time decreases as the object density increases.
![Page 34: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/34.jpg)
34
Performance Evaluation The impact of Object Agility and Speed
OBSERVATIONS The ASI based
Algorithm has a much better performance .
Response Time increases as the object agility increases.
Response Time remains almost unaffected with the object speed.
![Page 35: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/35.jpg)
35
Performance Evaluation Synthetic Dataset
Create terrains while varying the roughness Response Time increases as the roughness increases.
![Page 36: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/36.jpg)
36
Outline
Motivation Related Work Preliminaries The Naïve Approach Surface Index based Approach Performance Evaluation Conclusion and Future Work
![Page 37: Songhua Xing, Cyrus Shahabi and Bei Pan InfoLab University of Southern California Los Angeles, CA 90089-0781 Continuous Monitoring](https://reader035.vdocuments.us/reader035/viewer/2022062519/56649d015503460f949d3f05/html5/thumbnails/37.jpg)
37
Conclusion and Future Work
Contribution First study the Continuous kNN query on Surface; Show the concept of expansion tree for land surface does
not work as SE-tree suffers from intrinsic defects: it is fat and short;
Propose a superior approach that partitions SE-Tree into hierarchical chunks of pre-computed surface distances which overcome the above deficiency.
Experimentally verify the applicability of proposed methods.
Future work Study the Continuous kNN query problem for arbitary
moving query point.