Top Banner
Dynamic Route Construction for Mobile Collectors in Wireless Sensor Networks Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1
20

Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

Dec 15, 2015

Download

Documents

Abigail Bourns
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

1

Dynamic Route Construction for Mobile Collectors in Wireless Sensor NetworksPresentation by: Drew WichmannPaper by: Samer Hanoun and Saeid Nahavandi

Page 2: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

2

Wireless Sensor Networks (WSNs)

Set of small nodesDistributed in spaceMonitor conditionsBuilt with

Transceiver Microcontroller Sensor Energy source

Page 3: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

3

Power Consumption

Biggest Issue

Forwarding

Bottlenecking Unbalanced distribution

Solution Mobile Sinks

Page 4: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

4

Mobile Sinks

Mechanical data carrier Robot Unmanned Aerial Vehicle (UAV)

Physically approach sensorsRequires routes

Random Static Dynamic

Page 5: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

5

Problem Formulation

Assumptions Known locations Sensor nodes stationary Uniformly distributed Sleep when full Mobile collector▪ Sufficient energy▪ Sufficient memory

Page 6: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

6

Problem Formulation (Continued)

Page 7: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

7

Route Construction Algorithm

1. Build a fully connected graph G(V,E) of all sleeping sensors

2. Select a vertex r (center of sensing field) to be a root vertex

3. Compute a minimum spanning tree T for G from root r

4. Let L be the list of vertices visited in a DFS on T

5. Generate the Hamiltonian cycle H that visits the vertices in the order L

6. Follow the Hamiltonian cycle H as the constructed route

Page 8: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

8

Example

Page 9: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

9

Build connected graph

Page 10: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

10

Select root

Root r

Page 11: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

11

Compute Minimum Spanning Tree

Page 12: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

12

Create List

ABFGCDE

Page 13: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

13

Find Hamiltonian Cycle

Page 14: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

14

Follow Hamiltonian Cycle

Page 15: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

15

Simulation

Parameters 150 Sensors (uniformly distributed)

100m x 100m area

1 KB buffer

50000 time units

Page 16: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

16

Simulation Parameters (continued)

Events occur at centerRi = i * R1

R1 = 10msensingrangei =

[ baserate * (i-1) + 1, baserate * i ]

baserate = 2 seconds

Page 17: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

17

Results

20 different independent networksCompared with closest neighborMetrics

Sleeping Time Number of Sleeping Sensors Sleeping Time per Request Distance Travelled per Request

Page 18: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

18

Results

Page 19: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

19

Results

Page 20: Presentation by: Drew Wichmann Paper by: Samer Hanoun and Saeid Nahavandi 1.

20

Conclusion

No dependency on data generation rates

Minimize sleeping times

Better performance than closest neighbor

Shows effect of speed and number of collectors

Future work Cooperation between collectors Real-time requests