Modeling Timevariant User Mobility in Wireless Mobile Networks

1
Modeling Time-variant User Mobility in Wireless
Mobile Networks

• Wei-jei Hsu et al.
• UFL, INRIA, USC
• Presenter Zheng Guo

2
Outline

• Introduction
• Related Work
• Description of Time-variant Mobility Model
• Derivation of Theoretic Expressions
• Hitting Time Meeting Time
• Performance Evaluation
• Conclusions and Discussions

3
Introduction

• Requirement of mobility-assisted routing in
  wireless mobile networks (DTN)
• Lack of realistic mobility model
• Objective
• Time-variant community mobility model
• Skewed location visiting preferences periodical
  re-appearance
• Hitting time meeting time

4
Related Work

• Current mobility pattern
• IID mobility pattern random walk
• Trace-based pattern
• Existing models
• WLAN association model
• Inter-encounter time distribution

5
Description of Time-variant Mobility Model

• Mobility characteristics
• Skewed location visiting
• Periodical re-appearance

6
Time-variant community mobility model

• Time structure
• Normal movement periods (NMP)
• Concentration movement periods (CMP)
• Communication size

7
Time-variant community mobility model (cont.)

• Local epoch roaming epoch
• Two-state Markov model

8
Parameters of model
9
Probabilities of different states
10
Derivation of Theoretic Expressions (hitting time)

• Unit time slice Bernoulli trial with Ph
• Overall geometric distribution

11
Derivation of Theoretic Expressions (hitting
time) (cont.)
12
Derivation of Theoretic Expressions (hitting
time) (cont.)
13
Derivation of Theoretic Expressions (meeting time)
14
Derivation of Theoretic Expressions (meeting
time) (cont.)
15
Derivation of Theoretic Expressions (meeting
time) (cont.)
16
Derivation of Theoretic Expressions (meeting
time) (cont.)
17
Derivation of Theoretic Expressions (meeting
time) (cont.)
18
Validation of Theory with Sim.
19
Validation
20
Relative Error
21
Relative Error (cont.)
22
Performance Evaluation
23
Performance Evaluation (cont.)
24
Fine-tuning Model

• Six-tier communities
• Three distinct time periods

25
Conclusions Discussions

• Time-variance mobility model
• Match existing traces characteristics
• Mathematical expressions
• Community based?
• Trajectory encounter?
• Bernoulli distribution?
• Random community selection in very time slice ?