1

Core-PC: A Class of Correlative Power Control Algorithms for Single Channel Mob

ile Ad Hoc Networks

Jun Zhang and Brahim Bensaou

The Hong Kong University of Science and Technology

TWC 07

Outline

Introduction Correlative Power Constraints Core-PC Performance Evaluation Conclusion

Introduction

The energy supply in wireless devices is limited by their battery capacity.

From measurements in real systems,– Packet processing only consumes a small fraction– The energy is consumed by the

transmission reception listening to the channels

Introduction

It is important to design power control algorithms that– Improving network throughput– Reducing energy consumption

Goal

To design power controlled MAC protocols – Throughput better than IEEE 802.11– Energy consumption smaller than IEEE 802.11

Introduction

All previous works on power control only consider the assignment of the transmission power of each frame separately.

Introduction

The authors derive a set of equations that correlate the transmission powers of RTS, CTS, DATA and ACK frames.

The authors derive a class of adaptive power control algorithms (Core-PC).

Correlative Power Constraints-- Basic Framework and Definitions

The transmission zone– The received power level of a frame from node i i

n its transmission zone is higher than or equal to κ.

The carrier sensing zone– The received power level of a frame from node i i

n its carrier sensing zone is higher than or equal to η.

Correlative Power Constraints-- Noise Level Estimation

When node A is receiving the CTS reply from a node B, we assume– The channel at node A is idle.– The node’s NAV is always set and the node is

silenced whenever it is in the transmission zone.– The thermal noise level is negligible.– The propagation model is the two ray ground

propagation model.

4

1),(d

rtg Path loss

Correlative Power Constraints-- Noise Level Estimation

Ravg : average transmission range

Pavg : average transmission power Δ : the density of simultaneous

transmitters outside A’s transmission zone

Δ is upper bounded by 2

1

avgR

Correlative Power Constraints-- Noise Level Estimation

ARTSR avg

avgANoise dx

xR

xPP

,

42,

2

According to the two ray ground propagation model

4/1)/( avgavg PR

ANoiseBARTS

avgANoise P

P

PP ,,

~

42

2

x

P

R

x avg

avg

Correlative Power Constraints -- Requirements on Power Assignment for Frame Reception

BABRTSP ,

ABACTSP ,

The received power of RTS from node A to node B at location B

The received power of the CTS at node A must also fulfill the SIR requirement.

ANoiseABACTS PP ,, SIR threshold

),( ABgP

PP

avg

BARTSAB

CTS

The transmission power of CTS from node B to node A

A BRTS

CTS

Correlative Power Constraints -- Requirements on Power Assignment for Frame Reception

A B

),( ABgP

PP

avg

BARTSAB

CTS

),( ABgP

PP

avg

ABCTSBA

DATA

),( ABgP

PP

avg

BADATAAB

ACK

Correlative Power Constraints -- Requirements on Power Assignment for Frame Reception

To simplify the notation,

),(/),(min rsgrsP

),(/,max rsgP rs

xframe outgoingan ofpower ion transmissthe:xP

xframe incomingan ofpower ion transmissthe:xP

Correlative Power Constraints -- Requirements on Power Assignment for Frame Reception

))(),((min xrxsPPx

),( ABgP

PP

avg

BARTSAB

CTS

)(),(max OrOsavg

O PPPP I

Correlative Power Constraints -- Feasibility of Power Assignment

)),,(( max max,min

RTS

avgrsCTS

P

PPrsPP

)),,(( max max,min

CTS

avgrsDATA

P

PPrsPP

)),,(( max max,min

DATA

avgrsACK

P

PPrsPP

Correlative Power Constraints -- Feasibility of Power Assignment

maxmin ),( PrsP

rsavgP

P,

max

max2, PPP RTSavgrs

The minimal possible power assignment for a DATA frame in a successful 4-way handshake is

)),(,max( min,max rsPPP rsavg

Core-PC

Core-PC

In Algorithm 1, different combination of (PRTS,Pavg) lead to different power assignment algorithms.

Core-PC

Three alternative approaches may be adopted for setting the PRTS.– (a) Simple scenario

PRTS=Pmax

– (b) Symmetric scenario PRTS=PCTS

– (c) Minimum power scenario The RTS frame is transmitted at a power level such that th

e DATA transmission power is minimized.

Core-PC

Similarly, there are different ways of choosing the value of Pavg.

– (A) Worst case scenario Pavg=Pmax

– (B) Node-related adaptive scenario (i) initially Pavg=Pmax

(ii) all other nodes transmit their DATA frame at the same power level

– (C) Network-related adaptive scenario Pavg=0.9Pavg+0.1Pt transmission power

of a captured frame

Performance Evaluation

Simulator: NS-2 Routes: AODV Data rate: 11Mbps κ: 3.652e-10 Watts η: 1.559e-11 Watts ζ: 10dB Pmax: 0.2818 Watts

Rmax: 250m

Performance Evaluation-- Single Hop Scenario

50m

(I) 200m

(II) 150m

CBR traffic is generated and carried in UDP datagrams with a packet size of 512 bytes

Two packet sending rate:– 200 packets/s per sender– 400 packets/s per sender

Performance Evaluation-- Single Hop Scenario

Rate = 200 PKT/SECEND 200m

Performance Evaluation-- Single Hop Scenario

Rate = 400 PKT/SECEND 200m

Performance Evaluation-- Single Hop Scenario

Rate = 400 PKT/SECEND 150m

Performance Evaluation-- Multi-Hop Static Scenario

49 nodes 400*400 m2 area 7 flows are set between randomly chosen en

d-to-end source-destination pairs Packet size is 512 bytes

Performance Evaluation-- Multi-Hop Static Scenario

Performance Evaluation-- Multi-Hop Static Scenario

Conclusion

The correlations among the required transmission power of RTS/CTS/DATA/ACK– derived constraints that ensure the correct delivery

Using these constraints– The authors developed a class of correlative power

control algorithms

Simulations have shown the algorithm achieves– Higher throughput– Lesser energy consumption

Thank you