Tackling  Self  Interference,  Cross-­Technology  Interference  and  Channel  Fading  in  WSN

Mohammad  MobashirSchool  of  Computing,  National  University  of  Singapore

mobashir@comp.nus.edu.sg

Scalable  Data  Dissemination

Motivation

This  leads  to  severe  interference  of  the  following  two  kinds

Self  InterferenceSame  technology  devices  transmitting  synchronously

Cross-­‐Technology  InterferenceDifferent  Technologies  sharing  the  2.4GHz  ISM  band

With  a  push  towards  “IoTification”,  the  world  foresees  a  sharp  rise  in  the  number  of  devices  having  wireless  

networking  capabilities  

2.4  GHz  ISM  Band

Synchronous  TransmissionMultiple  transmitters  to  send  same  data  concurrently  to  multiple  receivers.Eliminates  channel  contention.  Allows  fast  data  dissemination.• Glossy,  LWB,  Splash,  Chaos,  P3,  etc.

Scalability  ProblemReliability  drops  for  large  number  of  concurrent  transmitters.• Difficult  to  maintain  precise  synchronization  (<0.5μs)• Due  to  combination  of  software, hardware and signal  propagation  delays.

0

0.2

0.4

0.6

0.8

1

2 3 4 5 6 7

Pack

et R

ecep

tion

Rat

io (P

RR

)

Number of concurrent transmitters

0

0.2

0.4

0.6

0.8

1

2 3 4 5 6 7

Pack

et R

ecep

tion

Rat

io (P

RR

)

Number of concurrent transmitters

0

0.2

0.4

0.6

0.8

1

2 3 4 5 6 7

Pack

et R

ecep

tion

Rat

io (P

RR

)

Number of concurrent transmitters

0

0.2

0.4

0.6

0.8

1

2 3 4 5 6 7

Pack

et R

ecep

tion

Rat

io (P

RR

)

Number of concurrent transmitters

Indriya Twist Indoor Outdoor

ApproachExploit  Capture  Effect over  Multi-­‐Channels• Receiver can decode a packet in the presence of interfering signals if the

signal from the sender is sufficiently  stronger  than that from the interferers

0

0.2

0.4

0.6

0.8

1

0 5 10 15 20 25 30 35

Pack

et R

ecep

tion

Ratio

(PRR

)

6 (dBm)

IndoorOutdoor

IndriyaTwist

0

0.2

0.4

0.6

0.8

1

1 2 3 4 5 6 7 8

Pack

et R

ecep

tion

Rat

io (P

RR

)

Number of concurrent transmitters

-84.09

-83.02

-80.93

-70.89

-70.50

-70.04

-56.97-52.47

Requires  much  less  stringent  timing  requirement  (<160 μs)  than  that  of  synchronous  transmission (<0.5 μs)

Design

Root

S4S1 S2 S3

R1 R2 R3 R4 R5

Nodes S1 S2 S3 S4

R1 G B W W

R2 G B W W

R3 W G B W

R4 W B B G

R5 W W W G

1

Nodes S1 S2 S3 S4

S1 0 1 0 0

S2 1 0 1 1

S3 0 1 0 1

S4 0 1 1 0

1

S2S1

S3S4

Root

S4S1 S2 S3

R1 R2 R3 R4 R5

Conflict GraphLink Classification MatrixTopology Construction Channel Assignment

Graph ColoringG: Si Good for Rj

B: Si Bad for Rj

W: Si Weak for Rj

Channel 1Channel 2Channel 3

Evaluation

0

20

40

60

80

100

Glossy:Ch26 Glossy:Ch22 Glossy:Ch19 Glossy:Ch20 Glossy:Ch14 Glossy:Ch15 Syncast

Rel

iabi

lity

%

Day 1 - Day 5 Variations

Day1 Day2 Day3 Day4 Day5

Robust  Data  Collection  in  Dynamic  Urban  EnvironmentCross-­‐Technology  Interference2.4GHz  ISM  band  is  shared  by  a  number  of  Wireless  Technologies

WiFi ZigBee Bluetooth

Availability  of  ZigBee  Channels

Expectation

For  Robust  Data  Collection  in  Urban  Environment• No  CTI-­‐Free  channel  is  be  expected• Channel  Estimation  in  such  environment  is  expensive• Exploit  Spatial  and  Channel  Diversity to  provide  enough  communication  

opportunities  so  that  channel  estimation  becomes  redundant

Approach

The  Internet  Of  Things

Phone Microwave

AcknowledgementThis work is supported in part by the Agency for Science, Technology andResearch (A*STAR), Singapore, under SERC Grant 1224104049.

Source:  Cisco  IBSG  projections,  UN  Economic  &  Social  Affairs

11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

1 6 112 1412 133 4 5 7 8 9 10

i

802.15.4 ith Channel 802.11 b/g ith Channel802.15.4 Channels

orthogonal to 1-6-11 802.11 b/g Channels

1 6 11

2 7 12

3 8 13

4 9 14

5 10

2.4 GHz ISM Band

2.400 GHz 2.490 GHz

Ove

rlapp

ing

of W

iFi

and

ZigB

ee C

hann

els

Poss

ible

WiF

iCh

anne

l Ass

ignm

ents

i

Reality  in  Urban  Environments

Presence of CTI significantly reducesthe number of “good” communicationopportunities or communication links