Energy-efficient Energy-efficient Data Collection in Wireless Data - - PDF document

▶
energy efficient energy efficient data collection in
SMART_READER_LITE
LIVE PREVIEW

Energy-efficient Energy-efficient Data Collection in Wireless Data - - PDF document

Oct 09, 2022 159 likes •372 views

Energy-efficient Energy-efficient Data Collection in Wireless Data Collection in Wireless Sensor Networks Sensor Networks G. Anastasi Pervasive Computing & Networking Lab. (PerLab) Department of Information Engineering University of

slide-1
SLIDE 1

1

Energy-efficient Data Collection in Wireless Sensor Networks Energy-efficient Data Collection in Wireless Sensor Networks

  • G. Anastasi

Pervasive Computing & Networking Lab. (PerLab) Department of Information Engineering University of Pisa, Italy g.anastasi@iet.unipi.it

PerLab WIRTES 2007 Pisa, July 2, 2007

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Research Topics

Pervasive and Mobile Computing

Wireless Sensor Networks Opportunistic Networking Ad Hoc and Mesh Networks Power management for mobile computing

slide-2
SLIDE 2

2

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Research Topics

Pervasive and Mobile Computing

Wireless Sensor Networks Opportunistic Networking Ad Hoc and Mesh Networks Power management for mobile computing

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

A Sensor Node A Sensor Node

Power Unit Power Unit

Power Generator Power Generator Sensor ADC

Processor Processor Memory Memory

Transceiver Transceiver Location Finding System Location Finding System

Mobilizer Mobilizer

Sensor Node Architecture

  • Small
  • Low power
  • Low bit rate
  • Low cost
slide-3
SLIDE 3

3

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Dense Sensor Networks

  • Several thousand nodes
  • Nodes are tens of feet of each other
  • Densities as high as 20 nodes/m3
  • Multi-hop communication

I.F.Akyildiz, W.Su, Y. Sankarasubramaniam, E. Cayirci, “Wireless Sensor Networks: A Survey” Computer Networks (Elsevier) Journal, March 2002.

Sensor Field

Internet, Internet, Satellite, Satellite, etc etc Sink Sink

Task Manager

Sensor Field Sink Nodes

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Sparse Sensor Networks

Distance between nodes larger than transmission range Data collection through mobile nodes (data mules)

Part of the external environment (buses, cabs, …) Part of the infrastructure (robots, …)

M R

Base Station Data Mule

  • S. Jain, R. Shah, W. Brunette, G. Borriello, S. Roy

“Exploiting Mobility for Energy Efficient Data Collection in Wireless Sensor Networks” ACM/Springer Mobile Networks and Applications,

  • Vol. 11, pp. 327-339, 2006.
slide-4
SLIDE 4

4

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Data Collection through Data Mules

Pros

Increased system lifetime Increased reliability Increased capacity Increased flexibility

Cons

Increased message latency (delay-tolerant applications) Limited scalability (unless multiple mules are used) Physical obstacles may limit mule’s movements Costs of data mule(s)

How to prolong the network lifetime?

Dense Sensor Networks Sparse Sensor Networks

slide-5
SLIDE 5

5

How to prolong the network lifetime?

Dense Sensor Networks Sparse Sensor Networks

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Possible approaches

Energy Harvesting Low-Power Components and Design Topology Management Power Management Management Data compression/aggregation Optimal Sampling Predictive Monitoring …

slide-6
SLIDE 6

6

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Power Management

Sleep Wakeup Scheduling

Switches off the radio subsystem during inactivity periods Sleep/wakeup schedule needed for communication

Adaptive sampling

Reduces the amount of data to be transmitted to the sink Decreases the power consumption for sensing

Adaptive Duty Cycling Sleep/Wakeup Scheduling Adaptive sampling

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Adaptive Sleep/Wakeup

Goal: minimize radio energy consumption Basic Idea of Sleep/Wakeup Schemes

Nodes sleep for most of the time Wake up periodically for transmitting/receiving data MAC or Application-layer protocol

Our Scheme

Application-layer protocol Relies on a routing tree Active periods dynamically adjusted

slide-7
SLIDE 7

7

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Multi-hop Network

~Communication Period ~ .Silence Interval . .Talk Interval . i j

Routing Tree

Data flow from leaves to the root (sink node)

Coordination scheme

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Fully Synch (TinyDB)‏

Pros

Simplicity

Cons

Static scheme Global duty-cycle (low efficiency) Requires clock synchronization

3 1 ... 2 4 ... ...

slide-8
SLIDE 8

8

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Fixed Staggered (TAG)‏

Parent-child talk intervals

Adjacent to reduce sleep-awake commutations Clock synchronization (periodic timestamps sent by the sink node) Pros

Pipeline Suitable to data aggregation

Cons

Static scheme Global parameters

3 1 ... 2 4 ... ... Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Our Proposal: Adaptive Staggered

Adaptive duty cycle

Variable-length talk intervals depending on

Number of children Network traffic Channel conditions Node Joins/Leaves …

1 2 3 4

slide-9
SLIDE 9

9

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Simulation

Simulation Setup

Ns-2 tool IEEE 802.15.4 MAC protocol

Network scenario

50m×50m area 30 nodes randomly deployed

Performance Indices

Average Activity Time (in % wrt always on) Delivery Ratio Average Message Latency Fairness (MRD)

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Simulation Results

Initial phase Sample topology Talk interval adaptation

Talk interval (secs)‏ Communication period #

Steady-state

Adaptation example

Node #

TX Range = 15m, CS Range = 30m, 50m x 50m grid, 30 nodes

slide-10
SLIDE 10

10

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Simulation Results

Global delivery ratio

66,00% 68,00% 70,00% 72,00% 74,00% 76,00% 78,00% 80,00% 82,00% 84,00% Adaptive staggered Fixed staggered (optimal) Fixed staggered (TAG) Alw ays on Delivery ratio

Protocol fairness

0,00% 1,00% 2,00% 3,00% 4,00% 5,00% 6,00% 7,00% 8,00% Adaptive staggered Fixed staggered (optimal) Fixed staggered (TAG) Always on Mean relative deviation (MRD) End to end latency 50 100 150 200 250 300 350 400 450 500 Adaptive staggered Fixed staggered (optimal) Alw ays on Latency (msecs)

Comparison with other approaches

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Power Management

Sleep Wakeup Scheduling

Switches off the radio subsystem during inactivity periods Sleep/wakeup schedule needed for communication

Adaptive sampling

Decreases the power consumption for sensing Reduces the amount of data to be transmitted to the sink

Adaptive Duty Cycling Sleep/Wakeup Scheduling Adaptive sampling

slide-11
SLIDE 11

11

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Snow Sensor

Courtesy by E. Pasero (PoliTo), C. Alippi (PoliMi)

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

An Embedded System to Evaluate the Snow Status

5000 10000 15000 20000 25000 30000 35000 40000 45000 50 Hz 100 Hz 200 Hz 500 Hz 800 Hz 1 KHz 5 KHz 10 KHz 50 KHz 100 KHz

Measuring frequency Equivalent capacity

Snow Snow Snow Snow Snow Snow Ice Ice Ice Ice Ice Ice Ice Ice Air Water

water ice snow

Courtesy by E. Pasero (PoliTo), C. Alippi (PoliMi)

slide-12
SLIDE 12

12

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Snow sensor

Power consumption

Sensing only

Power consumption

Sensing, processing and communication

Courtesy by E. Pasero (PoliTo), C. Alippi (PoliMi)

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Adaptive Sampling Algorithm

Estimate Fmax by considering the initial W samples and set Fc = c * Fmax.; Define Fup = (1 + (c-2)/4) * Fmax and Fdownp = (1 – (c-2)/4) * Fmax; h1=0 and h2=0; for (i=W+1; i < DataLength; i++) { Estimate the current maximum frequency Fcur on the subsequence (i-W, W) if ( | Fcurr - Fup | < | Fcurr - Fmax | ) h1= h1+1; else if ( | Fcurr - Fdown | < | Fcurr - Fmax | ) h2= h2+1; else { h1=0; h2= 0; } if (h1 > h )|| (h2 > h) { Fc = c * Fcurr.; Fup = (1 + (c-2)/4) * Fcurr; Fdown = (1 – (c-2)/4) * Fcurr; } }

slide-13
SLIDE 13

13

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Simulation

Simulation Setup

MatLab Experimental datasets

Network Scenario

Star Topology TDMA communication scheme Adaptive sampling algorithm at Base Station

Perfomance Indices

% of samples wrt fixed over-sampling Mean Relative Error (MRE)

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Simulation Results

Sampling Fraction (17-26%)‏

slide-14
SLIDE 14

14

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Simulation Results

Graphical Comparison

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Simulation Results

Energy Consumption

17% 150 mJ/sample Duty-cycle 3-5% Duty-cycle + Adaptive Sampling 100% Activity ratio 880 mJ/sample Always On Power cons. Power management scheme

slide-15
SLIDE 15

15

How to prolong the network lifetime? How to prolong the network lifetime?

Dense Sensor Networks Sparse Sensor Networks

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Sparse Sensor Networks

Data Mules are resource rich Static sensors are energy-constrained

Power management

Mule-sensor communication

Mule discovery protocol Data transfer protocol

MR

Base Station Data Mule

slide-16
SLIDE 16

16

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Experimental testbed

Investigated factors

mule distance mule speed

Testbed environment

MULE collecting data packets sent by a static node straight path and constant speed

Performance measures

Contact time Packet loss behavior Number of successfully transmitted packets

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Experimental Measurements

Impact of distance between sensor and mule

NUMBER OF PACKETS SUCCESSFULLY TRANSFERRED TO THE MULE FOR DIFFERENT Dy VALUES (v=1 m/sec)

Impact of the MULE’s speed on packet loss and contact time

NUMBER OF PACKETS SUCCESSFULLY TRANSFERRED TO THE MULE AT DIFFERENT SPEEDS (Dy = 15m)

slide-17
SLIDE 17

17

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Energy-efficient Data Collection

Many previous works assume

Circular transmission range Packet loss negligible within the transmission range

[Somasundara-2006] proposes a stop&wait approach

Uses acks for

Reliability (message received) Beaconing (mule still within the transmission range)

No assumption about

Packet loss behavior Mule position and mobility

Starts transmitting upon mule discovery Robust but not efficient

  • A. Somasundara, A. Kansal, D. Jea, D. Estrin, M. Srivastava, “Controllably Mobile Infrastructure

for Low Energy Embedded Networks”, IEEE Trans. on Mobile Computing, Vol. 5, N. 8, August 2006.

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Optimal vs. Naïve Approach

Optimal approach Naïve approach

B dt t Th t t

  • end
  • start

t t

  • start
  • end

= ⋅ −

∫

) ( to subject minimize

[ ]

n end n start t

t ,

B dt t Th

n end n start

t t

= ⋅

∫

) (

such that

n ery dis n start

t t

cov

=

0.2 0.4 0.6 0.8 1 1.2 1.4

  • 80
  • 40

40 80

v=1m/s, Dy=15m v=40km/h, Dy=15m

  • 80
  • 40

40 80 Message Loss Probability time (s)

( ) tstart tend tstart tend

Dx (m)

slide-18
SLIDE 18

18

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Performance Comparison

Average data transfer delay

1 2 3 4 5 6 7 5 10 15 20 25 30 v=40Km/h, Dy=15m

naive, w=1

  • ptimal, w=1

naive, w=2

  • ptimal, w=2

naive, w=4

  • ptimal, w=4

naive, w=8

  • ptimal, w=8

Data-transfer time (s) Messages to transfer (B) Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Performance Comparison

Slowdown in the average data transfer delay

1 2 3 4 5 6 7 8 5 10 15 20 25 v=40Km/h, Dy=15m

w = 1 w = 2 w = 4 w = 8

Slowdown Messages to transfer (B)

slide-19
SLIDE 19

19

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Adaptive Data Transfer (ADT) protocol

Feasible Approximates the optimal protocol

Window-based scheme

Relies upon

Real measurements of the contact time

Assumes that

Message loss probability is minimum around the middle

  • f the contact time

Message loss probability function is (approximately) symmetric wrt the mid contact time

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Simulation Results

Mule’s speed

40 Km/h (1-50) 20 Km/h (51-100 40 Km/h (101-150)

1 2 3 4 5

20 40 60 80 100 120 140

B=10, Dy=15m

Naive Optimal ADT (alfa=0.5)

Data-transfer time (s)

Mule's passage

slide-20
SLIDE 20

20

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Sparse Sensor Networks

Data Mules are resource rich Static sensors are energy-constrained

Power management

Mule-sensor communication

Mule discovery protocol Data transfer protocol

MR

Base Station Data Mule Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Current Research Projects

  • ArtDecO: Adaptive Infrastructures for Decentralized Organizations
  • Funded by: MIUR, FIRB Project (2006-2009)
  • WiSeMaP: Wireless Sensor networks for Monitoring Natural Phenomena
  • Funded by: MIUR, PRIN Project (2006-2008)
  • GeoMon: Monitoraggio delle opere ingegneristiche e prove geotecniche

tramite l’utilizzo delle reti di sensori

  • Funded by: Engisud, Palermo (2007-08)
  • VirtusVini: Monitoraggio della produzione vitivinicola con reti di sensori
  • Funded by: Engisud, Palermo (2007-08)
  • CityNet: Progetto e realizzazione di un’infrastruttura per il monitoraggio

dell’inquinamento dell’aria mediante sensori a terra fissi e collettori mobili

  • Funded by: TDGroup, Pisa (2007)
  • Nautilus
  • Funded by: Consorzio di aziende toscane (2007)
  • Power Management in IEEE 802.16e (WiMax) Networks
  • Funded by Nokia Research Center, Helsinki (2006-07)
slide-21
SLIDE 21

21

Energy-efficient Data Collection in Wireless Sensor Networks WIRTES 2007

Recent Publications on WSN

  • C. Alippi, G. Anastasi, C. Galperti, F. Mancini, M. Roveri, “Adaptive Sampling

for Energy Conservation in Wireless Sensor Networks for Snow Monitoring Applications”, submitted to the IEEE Int’l Workshop on Mobile Ad-hoc and Sensor Systems for Global and Homeland Security (MASS-GHS 2007).

  • G. Anastasi, M. Conti, A. Passarella, L. Pelusi, “Mobile-relay Forwarding in

Opportunistic Networks”, chapter in Adaptive Processing in Wireless Networks (M. Ibnkahla, Editor), Taylor and Francis, New York (USA), accepted for publication.

  • G. Anastasi, M. Conti, M. Di Francesco, A. Passarella, “How to prolong the

Lifetime of Wireless Sensor Networks” chapter 6 in Mobile Ad Hoc and Pervasive Communications, (M. Denko, L. Yang, Editors), ASP, to appear.

  • G. Anastasi, M. Conti, E. Monaldi, A. Passarella, “An Adaptive Data-transfer

Protocol for Sensor Networks with Data Mules, Proceedings of the IEEE International Symposium on a World of Wireless, Mobile, and Multimedia Networks (WoWMoM 2007), Helsinki (Finland), June 18-21, 2007.

  • G. Anastasi, M. Conti, E. Gregori, A. Passarella, “Motes Sensor Networks in

Dynamic Scenarios: a Performance Study for Pervasive Applications in Urban Environments”, International Journal of Ubiquitous Computing and Intelligence,

  • Vol. 1, N.1, April 2007. Invited Paper.
  • G. Anastasi, M. Conti, M. Di Francesco, A. Passarella, “An Adaptive and Low-

latency Power Management Protocol for Wireless Sensor Networks”, Proc. ACM International Workshop on Mobility Management and Wireless Access (MobiWac 2006), Torremolinos (Spain), October 2006.