Energy-efficient collision-free medium access control for wireless

sensor networksVenkatesh RajendranKatia ObraczkaGarcia-Luna-Aceves

Department of Computer Engineering University of California, Santa Cruz

ACM SenSys’03 Speaker: Yung-Lin Yu

Outline• Introduction

• TRAMA– NP(Neighbor Protocol)– SEP(Schedule Exchange Protocol)– AEA(Adaptive Election Algorithm)

• Experimental setup

• Simulation results

• Conclusion

Introduction• The deployment of sensor network usually

done in ad-hoc manner– Self-organize into a multi-hop wireless network

• Nodes may be difficult to recharge

• Nodes recharging may not be cost effective

• Major challenge– Self adaptive to changes in traffic– Prolongs the battery life

TRAMA• Overview

– TRAMA consists of three components• NP (Neighbor Protocol)

• SEP (Schedule Exchange Protocol)

• AEA (Adaptive Election Algorithm)

NP• Nodes can only join during random access

periods– Main function of random access periods is node

additions and deletions– All nodes must be transmit or receive state– The most energy consumption

NP (cont.)• Using signaling packets to gather

neighborhood information– During the random access period

• Updates about its one-hop neighborhood– Added or deleted– Keep-alive– Time out a neighbor

SEP• SEP establishes and maintains schedule info

rmation

• The information is periodically broadcast

• Each node has a SCHEDULE_INTERVAL

• Winning slots– Node computes in the interval [t,t+ SCHEDUL

E_INTERVAL]• Last winning slot reserved for broadcasting the no

de’s schedule for the next interval

SEP (cont.)• Schedule packets

– Nodes announce their schedule via Schedule packets

• Using bitmap to transmit schedule packets– The length of bitmap is the number of one-hop

neighbors– Eg.

• A node has 4 one-hop neighbors with identities 14,7,5,4

• If broadcast , bitmap : 1111

• If multicast to 14 and 5, bitmap: 1010

SEP (cont.)• ChangeOver slot

– The slot after which all the winning slots go unused

• The maximum sleep periods– ChangeOver slot to l

ast winning slot

SEP (cont.)

• A summary of a node’s schedule– Sent with every data packet– Summary help minimize the effects of packet lo

ss– In order not to excessive overhead, the schedule

summary is 6 bytes

AEA• Purpose

– To Decide node’s state (TX, RX, SL)– Re-use slots

AEA (cont.)

CB

D

100

95 79

200

tx

A lost

ASK

Inconsistency problem

AEA (cont.)• Node u is tx(u)

– u wants to transmit• Let u.state = TX

• Let u.receiver = u.reported.rxId

– u gave up transmit• Call HandleNeedTransmissions

• tx(u) belongs to N1(u)– tx(u).announcedReceiver = u

• Let u.state = RX

– Else u.state = SL

AEA (cont.)• atx(u) hidden from tx(u)

– atx(u).announcedReceiver = u• Let u.state = RX

– Else u.state = SL

• HandleNeedTransmissions– ntx(u) = u

• Let u.state = TX• Let u.receiver = u.reported.rxId

– atx(u).announcedReceiver = u• Let u.state = RX

– Else u.state = SL

Experimental setup• Simulation platform

– Qualnet

• Physical layer model– TR1000

• 50 nodes are uniformly distributed over a 500m x 500m area

• 6 one-hop neighbors on average• 17 two-hop neighbors on average• 2 different types of traffic load

– Synthetic data generation– Data gathering application

Simulation results• Synthetic traffic

Average packet delivery ratio for synthetic traffic

Average queuing delay for synthetic traffic

Simulation results (cont.)

Simulation results (cont.)• Data gathering application

Simulation results (cont.)

Simulation results (cont.)

Conclusion• TRAMA achieves

– Energy-savings comparable to S-MAC– Delivery guarantees comparable to NAME

• TRAMA has higher delay– It Suited for

• Not delay sensitive

• High delivery guarantees

• Energy efficiency