A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

42
4/9/2015 1

Transcript of A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

Page 1: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

4/9/2015 1

Page 2: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

Project Guide :- Ms. Diya Thomas

Panel Members :- Project Members :-

Mr. Febin P Jacob Maria Lucy Thomas

Ms. Amitha Mathew Ponnu Jose

Priyanka Jacob

4/9/2015 2

Page 3: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

Introduction

Literature Survey

System Specification

Assumption and Dependencies

Flow charts and Pseudo code

Conclusion

Future Enhancements

Bibliography

4/9/2015 3

Page 4: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

Efficient routing in WSNs is an important field that

needs to be addressed.

This project aims to provide solution to Dead-End

problem of Location-Based Routing in Wireless

Sensor Networks.

4/9/2015 4

Page 5: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

A message falls into a local minimum using greedy

forwarding in WSNs.

PROPOSED ALGORITHM :-

Can generate loop-free short paths with higher

delivery ratio and lower energy consumption.

Can handle large-scale networks

4/9/2015 5

Page 6: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

Greedy Forwarding

Next hop is the neighbor that gets the packet

closest to destination.

destinationsource

4/9/2015 6

Page 7: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

Dead-End Problem in Greedy Forwarding

Base Station

SourceSource

Successful Routing using GF

4/9/2015 7

Page 8: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

4/9/2015 8

Page 9: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

Reactive routing protocol

Builds routes between nodes only as desired by source nodes.

A broadcast route discovery mechanism

RREQ (Route Request packet) broadcasting to find a route RREP (Route Reply packet) is used to set up forward path

Dynamic establishment of route table entries

4/9/2015 9

Page 10: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

Destination sequence number

For updating the latest information for a route Ensures that only most efficient route is used

Routing loops avoided using (source-addr, broadcast-id) pair in RREQ packet

4/9/2015 10

Page 11: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

Basic message set in

AODV includes:

Hello - For link status

RREQ - Route Request

RREP - Route Reply

RERR - Route Error

4/9/2015 11

Page 12: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

4/9/2015 12

Route Request (RREQ) broadcast flood

Source

Destination

RREQ

Page 13: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

4/9/2015 13

Route Reply (RREP) Propagation

Destination

Source

RREP

Page 14: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

4/9/2015 14

Route Reply (RREP) Propagation

Destination

Source

Packet Forwarding path

Page 15: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

1• High route discovery latency.

2• No reuse of routing information.

3• Overhead on the bandwidth.

4 • Vulnerable to misuse.

4/9/2015 15

Page 16: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

Sensors can figure out who its closest neighbors are (usingbeacons).

2 forwarding methods:-

- Greedy forwarding

- Perimeter forwarding

To calculate a path, GPSR uses a greedy forwardingalgorithm.

Perimeter forwarding will route around the perimeter of theregion.

4/9/2015 16

Page 17: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

4/9/2015 17

Greedy Forwarding Perimeter Forwarding

greedy fails

have left local minimagreedy works greedy fails

Page 18: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

GREEDY FORWARDING FAILURE

Perimeter Forwarding

4/9/2015 18

Apply the right-hand rule to traverse the edges of a void:-

- Pick the next anticlockwise edge.

Page 19: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

1• Generates longer paths with loops.

2• Risk of losing data packets is high.

3• Increases computational complexity.

4• Increases traffic burden.

4/9/2015 19

Page 20: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

Hardware Interfaces:

Processor : Pentium P3 or higher

Ram : 2 GB recommended

HDD : 4 GB

Software Interfaces:

Platform : Ubuntu

Interface : NS-2

4/9/2015 20

Page 21: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

4/9/2015 21

Page 22: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

Base Station

A

B C

D E

I

concave node

K

J

L

H

G

F

4/9/2015 22

Page 23: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

A

B C

DE

F

Shadow Area

Bright Area Base Station

IK

J

L

H

G

4/9/2015 23

Page 24: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

5

12

7

16

2018

8

15

16

2019

23

A

B C

DE

F

Euclidean distance to the BS

19+∆

∆ set as the average Euclidean

distance between neighboring

sensor nodes

Base Station

22

IK

J

L

H

G

4/9/2015 24

Page 25: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

5

12

7

16

2018

8

15

16

2019

23

A

B C

DE

F

22+∆

22

25

Base Station

IK

J

L

H

G

4/9/2015 25

Page 26: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

5

12

7

16

20

8

15

16

2019

23

A

B C

DE

F

23+∆

22

25 26

Base Station

IK

J

L

H

G

4/9/2015 26

Page 27: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

5

12

7

16

20

8

15

16

2019

23

A

B C

DE

F

26+∆

2625

2229

Base Station

IK

J

L

H

G

4/9/2015 27

Page 28: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

Sensor nodes are stationary and are arranged in a 2D

sensing field.

They exchange packets through Greedy Forwarding.

Euclidean distance is considered as the cost factor.

Node A’s cost is lesser than its neighbors' cost for the

Phase 2 to initiate.

4/9/2015 28

Page 29: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

4/9/2015 29

Page 30: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

4/9/2015 30

THE SHADOW SPREAD FUNCTION RUNNING AT NODE X

Shadow_Spread(node x)

status(x) = bright;

while active do

if receive beacon(status, location) from a neighbour y then

Refresh the neighbour set N(x) by updating status and location information of y;

if DecideShadow(x) then

status(x) = shadow;

else

status(x) = bright;

end if

if beacon timeout occurs then

Copy the status and location information of node x to Beacon(status, location) and broadcast

end if

Page 31: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

4/9/2015 31

AUXILLIARY FUNCTION DECIDE - SHADOW

Bool DecideShadow(node x)

shadow = true

for each node z€N(x) do

if((status(z)==bright)&&(distance(z)<distance(x)) then

return false;

end if

return shadow;

Page 32: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

Check

whether

each N(x) is

shadow

Set status as bright

Copy status and location information of

node x to Beacon(status, location) and

broadcast

Stop

Start

Set status as shadowyes

no

Node x receives beacon

messages from neighbors

N(x)

4/9/2015 32

Page 33: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

4/9/2015 33

THE COST SPREAD FUNCTION RUNNING

AT EACH SHADOW NODECost-Spread(node x)

cost(x) = distance(x);

while active do

if receive beacon(cost) from a neighbor y then

Refresh the neighbor set N(x) by updating the

cost of y;

if (cost(x)<= minimum cost of neighbor set N(x)) then

cost(x) = maximum cost of neighbours + ∆;

end if

end if

if beacon timeout occurs then

copy cost of node x to beacon (cost) and send out;

end if

Page 34: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

Shadow

nodes check

whether its’

N(x) has a

larger cost

Establish path to a neighbor with the lowest

cost and hence deliver packet to BS

Stop

Start

Increment the cost until each shadow

node has a N(x) with lesser cost

yes

no

Calculate the Euclidean

distance of each node to the

base station

4/9/2015 34

Page 35: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

4/9/2015 35

A

B C F

Base Station

IK

J

L

H

G

D E

Case 1 :

Bright node as source

Page 36: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

4/9/2015 36

A

B C F

Base Station

IK

J

L

H

G

D ECase 2 :

Shadow node as source

Page 37: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

• Current methods for Dead-end problem are insufficient.

• PAGER does not require a node to memorize paths.

• It constructs loop-free paths with lengths close to shortest

path.

• Packet delivery ratio is high.

4/9/2015 37

Page 38: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

• To consider other parameters as the cost factor.

• Algorithm can be extended to multiple base stations.

• Mobility constraint could be overcome!

4/9/2015 38

Page 39: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

[1] Le Zou, Mi Lu, and Zixiang Xiong, “A Distributed

Algorithm for the Dead-end Problem,” IEEE Trans. On

vehicular technology, vol. 54, no.4, July 2005.

[2] L.Akyildiz, et al., “A Survey on Sensor Networks,”

IEEE Communications Magazine, Aug 2002, pp. 102-114.

[3] I. Stojmenovic, “Position-Based Routing in Ad Hoc

Networks,’’ IEEE Communications Magazine, July 2002,

pp. 128-134.

4/9/2015 39

Page 40: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

[4] Karp B., Kung H.T, “Greedy Perimeter Stateless Routing

for Wireless Networks,” Proceedings of 6th Annual

ACM/IEEE International Conference of Mobile Computing

and Networking; Boston, MA, USA. August, 2000, pp. 243–

254.

[5] C. Intanagonwiwat, R. Govindan, D.Estrin, J.Heidemann,

and F. Silva, “Directed diffusion for wireless sensor

networking,” IEEE/ACM Trans. Netw., vol. 11, no. 1, pp. 2–

16, Feb, 2003.

4/9/2015 40

Page 41: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

4/9/2015 41

Page 42: A DISTRIBUTED ALGORITHM FOR THE DEAD-END PROBLEM IN WSNs

4/9/2015 42