An Application-Specific Protocol Architecture for Wireless

Microsensor Networks

Wendi HeinzelmanAnantha Chandrakasan and Hari Balakrishnan

Massachusetts Institute of Technology

Microsensor Networks

• Remote monitoring of the environment– Surveillance– Machine diagnostics

• Relevant parameters:– System lifetime (energy efficiency)– Quality– Latency

Base station

A

B D

E

f(A-G)C

FG

The MIT µ-AMPS Project

• Develop energy-optimized solutions – Node architecture and radio hardware– OS, algorithms and protocols

• Assumptions:– Base station far from nodes– All nodes energy-constrained– Locally, data correlated

Goal: Design protocols for high quality and energy and spectrum efficiency

ApplicationTransportNetwork

Physical

MACData-link

Application-SpecificArchitecture

General-Purpose Protocol Architectures

• Direct transmission– Power scales as Rn

– Bandwidth limitations

• Routing– MTE (Min. Trans. Energy) Routing– Multi-step communication– Short lifetimes for close nodes

• Clustering– Cellular model– Needs high-energy cluster-head

A�B�C if:dAB

2+dBC2<dAC

2

AB C

R

LEACH: Energy-Efficient Protocol Architecture

• Low-Energy Adaptive Clustering Hierarchy– Adaptive, self-configuring cluster formation– Localized control for data transfers– Low-energy medium access– Application-specific data aggregation

Base station

Cluster-head

Dynamic Clusters

• Cluster-head rotation to evenly distribute energy load• Adaptive clusters

– Clusters formed during set-up– Scheduled data transfers during steady-state

Time•••

START START START

Set-up Frame RoundSteady-state

Cluster-heads = •

Distributed Cluster Formation Algorithm

Using Pi(t)

Choose CH with “loudest” announcement

Cluster-headNodes

Non-CHNodes

Node icluster-head ?Yes No

Wait forcluster-head

announcements

Send Join-Requestmessage to chosen

cluster-head

Announcecluster-head status

Wait forJoin-Request

messages

Steady-stateoperation for

t=Tround seconds

Autonomous decisions lead to global behavior

• No global control• Flexible, fault-tolerant

Distributed Cluster Formation

ktN

i

=∗= ∑=

1)(P CH] E[#1

i

Ci(t) = 1 if node i a CH in last r rounds

• Assume nodes begin with equal energy• Design for k clusters per round• Want to evenly distribute energy load a Each node CH once in N/k rounds

−=0

)/mod(* )(Pi kNrkNk

t0 )(Ci =t

1 )(Ci =t

k = system param.(Analytical optimum)

• Can determine Pi(t) with unequal node energy

Unequal Initial Energies

High-energy nodes CH more often than low-energy nodes

ktE

tEt

total

i

)()(

)(Pi = Ei(t) = energy of node i at time tEtotal(t) = total energy in system at time t

Cluster-head nodesEi(t) ≈ Eo – X

Pi(t) ≈ 0

Non-cluster-head nodesEi(t) ≈ Eo

Pi(t) ≈ k/(N-kr)

If nodes begin with Eo

Etotal ≈ Eo(N-kr) + (Eo-X)kr

Set-up Steady-state

LEACH Steady-State

• Cluster-head coordinates transmissions– Time Division Multiple Access (TDMA) schedule– Node i transmits once per frame

• Cluster-head broadcasts TDMA schedule • Low-energy approach

– No collisions– Maximum sleep time– Power control

Clusters formedTime

Slot for node i

Slot for node i •••Frame

Inter-Cluster Interference

• Transmission in different clusters can collide– Nodes minimize transmission power– Each cluster has unique spreading code – Distributed solution to minimize interference

BA

C

Cluster B

Cluster C

LEACH Medium Access

ADV• CSMA

• Large power, small messages

ADV Join-REQ SCHTime

Steady-stateSet-up

Join-REQ• CSMA

• Large power, small messages

SCH• DS-SS code

• Power to reach all members

CH CHNon-CH

Data Transfer to Cluster-Head• TDMA slot (with DS-SS)

• Power to reach CH, large messages

Application-Specific Data Aggregation

• Clusters exhibit spatial locality– Local data aggregation– Common signal enhanced/noise reduced

• Computation vs. communication tradeoff– Depends on cost of computation and communication– Signal processing within the network

Energy for beamforming (J/bit/signal)

Tot

al e

nerg

y (J

)No DataAggregation

Local Data Aggregation

**

Base Station Cluster Formation (LEACH-C)

All Nodes Base Station

• Only nodes with Ei > µEeligible

• Get optimal clusters for comparison• Requires communication with base station• Need GPS or other location-tracking method

• Small packets• Large energy

Wait forinformation from

nodes

Determine optimalclusters and send info

to nodes

Send {(xi,yi),Ei} to BS

Wait forcluster

information

Clusters Formed

Simulation Framework

• Extensions to ns network simulator– Computation and communication energy models

• Radio energy

• StrongARM processor beamforming [A. Wang]

– New node states– Medium access– LEACH, LEACH-C, MTE routing, static clustering

Transceiver Tx Amplifier Transceiver

Eelec* kk bit packet εamp* k * d2 Eelec* k

d

Transmitter Receiver

k bit packet

Simulation Parameters

500 bytesData size

5 nJ/bit/signalBeamforming cost

100 pJ/bit/m2Transmit amplifier

50 nJ/bitRadio electronics

100 kbpsBit rate

50 µsProcessing delay

Base Station

75 meters

100

met

ers

100 meters

100 nodes

Optimum Number of Clusters

Ave

rage

ene

rgy

per

roun

d (J

)

Number of clusters (k)

k=7

k=9

k=5

k=3

k=1

k=11

• Too few clusters a cluster-head nodes far from sensors

• Too many clusters a not enough local signal processing

Analytical Optimum

δγ +=− kE CHnon

1

k clusters _ N/k nodes/cluster:

toBSopt d

MNk

6=

N=100M=10075 < dtoBS < 185

_ 2 < kopt < 6

• Simulation agrees with theory

βα +=kN

ECH dtoBS

d2toCH α 1/k

Data per Unit Energy

Am

ount

of d

ata

rece

ived

at B

S

Energy Dissipation (J)

Static-Clustering

LEACH-C

LEACH

MTE

• LEACH achieves order of magnitude more data per unit energy – 2 hops v. 10 hops average– Data aggregation successful

Nodes begin withlimited energy

Network Lifetime

Num

ber

of n

odes

aliv

e

Amount of data received at BS

Static-Clustering

LEACH-C

LEACHMTE

• LEACH delivers over 10 times amount of data for any number of node deaths

• Rotating cluster-head effective

Summary

• Microsensor network protocols must be designed for– Bandwidth efficiency– Energy efficiency– High quality

• Application-specific protocol architecture beneficial – Most efficient use of limited resources– Ideas extendable to other application spaces