MAC protocol issues beyond the passive RFID systemwinner.ajou.ac.kr/publication/data/invited/20051110_AutoIDLAB_JAE-HYUN-KIM.pdf · MAC protocol issues beyond the passive RFID system

MAC protocol issues MAC protocol issues beyond the passive RFID systembeyond the passive RFID system

AutoAuto--ID Labs China, Japan, and KoreaID Labs China, Japan, and KoreaShanghai MeetingShanghai Meeting

Fudan University

Nov. 10, 2005Nov. 10, 2005

JaeJae--Hyun KimHyun Kim

WWireless ireless IInformation nformation aaNNdd NNetwork etwork EEngineering ngineering RResearch Lab. esearch Lab. School of Electrical Engineering School of Electrical Engineering

AjouAjou univuniv., Korea., Korea

2Jae-Hyun Kim

AgendaAgenda

Research areasAchievements in RFIDOn-going researchEPC Class structure and research issuesResearch plan and proposalReference

3Jae-Hyun Kim

Research areasResearch areas

MAC Protocols Research Performance analysis of the IEEE 802.16 MAC protocolPerformance analysis of the IEEE 802.16 MAC protocol

End-to-end QoS ResearchEndEnd--toto--End QoS provisioning for 3GPP/3GPP2/ WLAN/Home End QoS provisioning for 3GPP/3GPP2/ WLAN/Home NetworkNetwork

Handover ResearchInterInter--RAT Handover (WCDMA to CDMA 2000)/Roaming RAT Handover (WCDMA to CDMA 2000)/Roaming algorithmsalgorithms

Security (Authentication) Research Wireless Access Control in Home NetworkWireless Access Control in Home Network

Sensor Network Research EnergyEnergy--efficient MAC Protocol / Network life time analysisefficient MAC Protocol / Network life time analysis

Anti-collision Research LowLow--powered and fast antipowered and fast anti--collision algorithm in RFID systemcollision algorithm in RFID system

4Jae-Hyun Kim


Performance analysis of the IEEE 802.16 MAC protocolMain ideaMain idea

The optimal initial backoff window according to the number of users and traffic characteristics

Initial Ranging Period

Data Transmission Period of SS

Contention Period ●●●

Gap between SSs

Gap between transmission and reception

Access Collision Access CollisionRequest Bandwidth

Request Bandwidth

( 1)(1 )1

Nt t

NS p p −⎛ ⎞= −⎜ ⎟⎝ ⎠

( )

01

0

1

1

2(1 )(2 ) 1

2 1

11 12 1 1

Mnc

cMnc

Mcc

Mc c

W p FD pCPp

M pp F FF CPp CP CPp

+=

+

+

⎛ ⎞−= × − ×⎜ ⎟−⎝ ⎠

⎛ ⎞+ ⎛ ⎞+ × − + × − + −⎜ ⎟ ⎜ ⎟⎜ ⎟− − ⎝ ⎠⎝ ⎠

∑

Throughput Delay Drop Probability

1Md cp p +=

N: the number of users, pt: transmission probability, pc: collision probability, CP: contention period, F : Frame size, M : Maximum retry

5Jae-Hyun Kim


ResultsResultsPropose the optimal initial backoff window using the cost function

AchievementsAchievements"The Analysis of the Optimal Contention Period for Broadband Wir"The Analysis of the Optimal Contention Period for Broadband Wireless Access eless Access Network," in Proc. PWN05, Kauai island in Hawaii, USA, Mar. 12, Network," in Proc. PWN05, Kauai island in Hawaii, USA, Mar. 12, 2005, pp. 2152005, pp. 215--220.220."Performance Enhancement of the Binary Exponential "Performance Enhancement of the Binary Exponential BackoffBackoff Algorithm in the IEEE Algorithm in the IEEE 802.16 MAC Protocol," submitted to VTC'06 spring.802.16 MAC Protocol," submitted to VTC'06 spring."Optimization of the Initial "Optimization of the Initial BackoffBackoff Window for Improving Performance of the BWA Window for Improving Performance of the BWA MAC Protocol," submitted to ICC'06.MAC Protocol," submitted to ICC'06.

1 2

3

_ ( ) _ _ ( ) _ ( )_ _ ( )

Cost Function N WF Nor inv throughput N WF Nor delay NWF Nor prob drop N

= × + ×+ ×

Number of users

Optimal Initial Backoff Window

2~6 W0 = 4

6~16 W0 = 8

16~70 W0 = 16

70~100 W0 = 32

The system is sensitive to the delay (WF1=0.1, WF2=0.7, WF3=0.2)- Optimal initial backoff window is 4

The system is affected by the packet drop probability (WF1=0.1, WF2=0.2, WF3=0.7)

6Jae-Hyun Kim


End-to-End QoS Performance evaluation of Next Generation Wireless Network

Main ideaMain ideaModeling of network elements and service trafficModeling of detailed protocol for each network elements

7Jae-Hyun Kim


ResultsResultsSimulation methodology for large scale network simulation Performance evaluation for voice service (ADPCM, PCM), data service (HTTP1.0, HTTP1.1, FTP) according to access technology (ATM, IP)Performance analysis of bandwidth gain for AAL2 using simulator and Markov chain and optimization of Timer_CU

AchievementsAchievements" " QoSQoS parameter mapping method " US Patent pendingparameter mapping method " US Patent pending""Analysis of Bandwidth Gain over Various Timer CU of AAL2 for VoiAnalysis of Bandwidth Gain over Various Timer CU of AAL2 for Voice Traffic ce Traffic MultiplexingMultiplexing," IEEE Trans. on ," IEEE Trans. on VehVeh. . TechnolTechnol., vol. 54, No.4, pp.1438., vol. 54, No.4, pp.1438--1446, Jul., 2005.1446, Jul., 2005."Performance Modeling and Evaluation of Data/Voice Services in W"Performance Modeling and Evaluation of Data/Voice Services in Wireless Networks,ireless Networks,””submitted to Wireless Network.submitted to Wireless Network.""EndEnd--toto--end Wireless Performance Simulator : end Wireless Performance Simulator : ModellingModelling Methodology and Methodology and PerformancePerformance," Lecture Note on Computer Science 3420 : Networking ," Lecture Note on Computer Science 3420 : Networking -- ICN2005, pp.258ICN2005, pp.258--267, Apr., 2005. 267, Apr., 2005. ""EndEnd--toto--end User Perceived Application Performance in 3G+ Networksend User Perceived Application Performance in 3G+ Networks," in Proc. IEEE ," in Proc. IEEE ICC'04, Vol. 4, Paris, France, Jun. 20ICC'04, Vol. 4, Paris, France, Jun. 20--24, 2004, pp. 2337 24, 2004, pp. 2337 -- 2341.2341.

8Jae-Hyun Kim


End-to-end User-perceived Application Performance in 3GPP2

FTP Download time, web page response time, channel throughput, FTP Download time, web page response time, channel throughput, user user goodputgoodput, TCP performance, , TCP performance, VVoice packet delayoice packet delay

0

50

100

150

200

250

mse

c

TDM tandemG.711 (64kbps)

ATM BackboneG.711 (64kbps)

AAL1

ATM BackboneG.726 (32kbps)

AAL2

IP BackboneG.726 (32kbps)

IP backboneVocoder Bypass

Technologies

Uplink Avg. Delay BackBone Avg. Delay Downlink Avg. Delay

E2E One Way Delay (ms)

64 kbps (G.711)

32 kbps(G.726)16kbps(G.728)

0 94 87 50 93 86

100 92 85 150 90 83 200 87 80 250 80 73 300 74 67 350 68 61 400 63 56 450 59 52

45 0

9 0 - 1 0 0 B e s t q u a l i t y ;

u s e r s v e r y s a t is f ie d

8 0 - 9 0 H ig h q u a l i t y ; u s e r s s a t is f i e d

7 0 - 8 0 M e d iu m q u a l i t y ; s o m e u s e r s d is s a t is f i e d

6 0 - 7 0 L o w q u a l i t y ; m a n y u s e r s d is s a t is f i e d

5 0 - 6 0 P o o r q u a l i t y ; n e a r ly a l l u s e r s d is s a t is f i e d

L e s s t h a n 5 0

U n a c c e p t a b le q u a l i t y

Voice Quality Scores (R value)

Uplink : MT – BTS – MSC/RNCBackbone : MG – Switch/Router – MGDownlink : MSC/RNC - BTS - MT

9Jae-Hyun Kim

1 10 20 30 40 50 60 70 80 90 100 120 140 1600

2

4

6

8

10

12

14

16

Number of concurrnt voice users

Ave

rage

ban

dwid

th g

ain(

%)

Timer CU= 0 ms (Analysis)Timer CU= 1 ms (Analysis)Timer CU= 2 ms (Analysis)Timer CU= 3 ms (Analysis)Timer CU= 4 ms (Analysis)Timer CU= 5 ms (Analysis)Timer CU= 6 ms (Analysis)Timer CU= 7 ms (Analysis)Timer CU= 0 ms (Simulation)Timer CU= 1 ms (Simulation)Timer CU= 2 ms (Simulation)Timer CU= 3 ms (Simulation)Timer CU= 4 ms (Simulation)Timer CU= 5 ms (Simulation)Timer CU= 6 ms (Simulation)Timer CU= 7 ms (Simulation)

Voice multiplexing issues in 3GPP2Packing densityPacking density Bandwidth gainBandwidth gain

The symbols present simulation results and the curves mean analyThe symbols present simulation results and the curves mean analytical results. tical results. Maximum bandwidth gain with AAL2 is about 18% higher than the baMaximum bandwidth gain with AAL2 is about 18% higher than the bandwidth ndwidth gain without AAL2gain without AAL2

1 10 20 30 40 50 60 70 80 90 100 120 140 160

84

86

88

90

92

94

96

98

100

Number of concurrent voice user

Pac

king

den

sity

in N

ode-

B(%

)

Timer CU=0 ms (Analysis)Timer CU=1 ms (Analysis)Timer CU=2 ms (Analysis)Timer CU=3 ms (Analysis)Timer CU=4 ms (Analysis)Timer CU=5 ms (Analysis)Timer CU=6 ms (Analysis)Timer CU=7 ms (Analysis)Timer CU=0 ms (Simulation)Timer CU=1 ms (Simulation)Timer CU=2 ms (Simulation)Timer CU=3 ms (Simulation)Timer CU=4 ms (Simulation)Timer CU=5 ms (Simulation)Timer CU=6 ms (Simulation)Timer CU=7 ms (Simulation)


10Jae-Hyun Kim


QoS mapping, packet scheduler, and CAC algorithm in Home Network

ClassifierClassifierClassify the each service traffic in IP layer

Class 1 Class 1 –– ReservationReservation--based algorithmbased algorithmWLAN

Negotiate with HC using TSPEC negotiationHCCA or EDCA(AC_VI, AC_VO)

IEEE 1394Negotiate the channel resource using request frame to Bus managerIsochronous mode

EthernetMark with highest priority in the TCI field

CACClass 1 traffic load traffic exceeded its share of threshold, stop accepting new call of Class 1

Class 2, 3, 4 Class 2, 3, 4 –– PriorityPriority--based algorithmbased algorithmThe packets in the queue are service by increasing the order of their deadline

IEEE 1394 IEEE 802.11 ethernet

Classifier CAC

Class 1 Class 2 Class 3 Class 4EDF SchedulerReservation

Control signal (to CAC)Control signal (from CAC)Data channel

Voice Video

( ) ( , )( ) ( , )( ) ( )total reservation

W i J i jD i a i jR j R j

= +−

Deadline :

i : number of class j : number of superframe W : weighting factor J : packet length Rtotal : total bandwidth Rreservation : reserved bandwidth a: packet arrival time

11Jae-Hyun Kim


Channel throughputTraffic load per service classTraffic load per service class

Throughput per nodeThroughput per node

Same the Same the throughput throughput per nodeper node

Node_1, Node_2, Node_1, Node_2, Node_3Node_3

NodeNode_0_0

NodeNode_0_0

NodeNode_1_1

NodeNode_2_2

NodeNode_3_3


NodeNode_0_0

NodeNode_1_1

Node_Node_22

NodeNode_3_3


Node_0 Node_1 Node_2 Node_3

Interarrivaltime 80 80 80 80

Service class

Real Block

Interactive

best effort best effort

W_1 0.462 0.23 0.154 0.154

W_2 0.662 0.23 0.054 0.054

Same Weighting Factor

12Jae-Hyun Kim


Inter-RAT Handover(WCDMA to CDMA 2000)/RoamingMain ideaMain idea

Channel modeling, mobility modeling, and network element modeling

Network model for Inter-RAT handoverWCDMA to CDMA2000 handover procedure

13Jae-Hyun Kim


0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

S ce na r io 1 S ce na r io 2

(Th _ t r=18)

S c ena r io 3

(Th _m=20)

S ce na r io 4

(Th _m=19,

Th _ t r=18)

S ce na r ios (pa rame te r )

Co

st

functi

on r

esult

s..

.

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

S ce na r io 1 S ce na r io 2

(Th _ t r=18)

S c e na r io 3

(Th _m=18)

S c ena r io 4

(Th _m=18,

Th _ t r=18)

S c ena r ios (p rame te r )

Co

st

functi

on r

esult

s..

.

ResultsResultsOptimal Inter-RAT Handover/Roaming scenario and Parameter sets

A viewpoint of users and service providers

AchievementsAchievements"Performance Evaluation of WCDMA"Performance Evaluation of WCDMA--toto--CDMA2000 Handover," accepted to KICSCDMA2000 Handover," accepted to KICS’’05 05 fall.fall."Performance Evaluation of WCDMA"Performance Evaluation of WCDMA--toto--CDMA2000 Handover," submitted to VTCCDMA2000 Handover," submitted to VTC’’06 06 spring.spring."WCDMA"WCDMA--toto--CDMA2000 Handover Technologies : Scenarios and Performance," CDMA2000 Handover Technologies : Scenarios and Performance," submitted to IEEE Communication Magazine.submitted to IEEE Communication Magazine.

Optimal parameter set for users Optimal parameter set for service-providers

14Jae-Hyun Kim


Wireless Access Control in Home NetworkMain ideaMain idea

Propose authentication scenarios and authentication protocolsPrevent the invalid wireless user from accessing to Home NetworkSecure and convenient wireless LAN access mechanisms

An example of access control in Home NetworkAn example of access control in Home Network

AchievementsAchievements"A Secure Wireless LAN Access Technique for Home Network," submi"A Secure Wireless LAN Access Technique for Home Network," submitted to VTC'06 tted to VTC'06 springspring"Wireless LAN Access Technique and Authentication Protocol for H"Wireless LAN Access Technique and Authentication Protocol for Home Network," ome Network," accepted by KICS'05 fall.

이웃집 무선 네트워크

DTV

WLAN

Residential Gateway

UWB Network

UWB

Web PAD

WLAN Network

Digital AudioNotebook

PDA

Media Center

PC(방2)

Health care

Home Automation

ZigBee 난방

접근접근

접근접근

접근접근 제어제어

이웃집 무선 네트워크

DTV

WLAN

Residential Gateway

UWB Network

UWB

Web PAD

WLAN Network

Digital AudioNotebook

PDA

Media Center

PC(방2)

Health care

Home Automation

ZigBee 난방

접근접근

접근접근

접근접근 제어제어Access control

Wireless Network

access

access

Warm

accepted by KICS'05 fall.

15Jae-Hyun Kim


Energy-efficient MAC Protocol / Network life time analysisMain ideaMain idea

Based on the IS-MAC protocol, but no routing tableThree thresholds in buffer for transmission

Timer to switch to sleep modeNo RTS and CTS signalDf

ComparisonComparison

ContentionTime

F-Filter ACK Time for Data Transmission

Timer

Active Time

Ack : the time for acknowledgement signal,filterF : the transmission time for filter signal, : timer to switch to sleep modeIST

IS filterT F Ack> +

HIGH MEDIUM LOW

Buffer Length

Transmission

D

H

J

SSourceNode

AB C

EF

G

I

Sink NodeD

H

J

SSourceNode

AB C

EF

G

I

Sink Node

Directional flooding protocol ISF protocol

16Jae-Hyun Kim


ResultsResultsAverage energy consumption per node

AchievementsAchievements"IS"IS--MAC Based Flooding Routing Protocol for Sensor Networks," in MAC Based Flooding Routing Protocol for Sensor Networks," in Proc.Proc. ACM PEACM PE--WASUNWASUN’’05, Montreal, Canada, Oct. 1005, Montreal, Canada, Oct. 10--13, 2005, pp.7913, 2005, pp.79--83. 83. "The Energy"The Energy--Efficient Algorithm for a Sensor Network," Lecture Note on CompuEfficient Algorithm for a Sensor Network," Lecture Note on Computer ter Science 3391 : Information Networking, pp. 293Science 3391 : Information Networking, pp. 293--302, Feb., 2005.302, Feb., 2005."IS"IS--MAC Based Flooding Protocol for Sensor Networks," in MAC Based Flooding Protocol for Sensor Networks," in Proc.Proc. IT International IT International Student Fair 2005, pp.167Student Fair 2005, pp.167--170, Aug., 2005.170, Aug., 2005."Numerical Analysis of Multiple and Single Path Schemes for Sens"Numerical Analysis of Multiple and Single Path Schemes for Sensor Networks," in or Networks," in Proc.Proc.JCCI 2005, JCCI 2005, DaeDae--GuGu, p. 127, Apr.,2005. , p. 127, Apr.,2005.

Case 1 : Varying the number of total nodesCase 1 : Varying the number of total nodes Case 2 : Varying the number of source nodesCase 2 : Varying the number of source nodes

17Jae-Hyun Kim


Development of low-powered and fast anti-collision algorithm in RFID system

Main ideaMain ideaMathematic analysis and the simulation of the anti-collision algorithmsDevelopment of the enhanced anti-collision algorithm

Tag collision problem Network model for the simulation

18Jae-Hyun Kim

Achievements in RFIDAchievements in RFID

Developed anti-collision algorithmsNovel AntiNovel Anti--collision Algorithms for Fast Object Identification in collision Algorithms for Fast Object Identification in RFID systemRFID system

ALOHA-based algorithm (ISO 18000-6 type A)Propose the method estimating the number of tags

Fast Wireless AntiFast Wireless Anti--collision Algorithm in Ubiquitous ID System collision Algorithm in Ubiquitous ID System Bit-by-bit binary tree-based algorithm (AutoID Class 0)Propose the method using bit sequence information of tag ID

A Novel Tag identification algorithm for RFID System using UHFA Novel Tag identification algorithm for RFID System using UHFBinary-tree based algorithm (AutoID Class 1)Propose the method using the 8-bit information (not 3 bit) transferred from tagsPropose the method using bit sequence information of tag ID

19Jae-Hyun Kim


Probabilistic slotted ALOHA (ISO 18000-6 Type A)

TAG4(0101)

TAG3(0011)

TAG2(1010)

TAG1(1011)

STATE

2nd REQSlot4Slot3Slot2Slot11st REQREADER

Frame size = 4

IDLE1011 COLL 0101

0011

1010

1011

0101

0011

1010

Frame size = ?

Need to vary the Frame size for the number of tags

20Jae-Hyun Kim


Dynamic Framed Slotted ALOHA I/II AlgorithmsEnhance the performance of the algorithm defined in ISO 18000Enhance the performance of the algorithm defined in ISO 18000--6 6 Type AType ABasic conceptBasic concept

Two Tag Estimation Methods (TEM)Ratio of the number of collided slots to the frame size

Number of tags related with collision in a slot

0 100 200 300 400 500 600 700 800 900 1000

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Number of tags

Col

lisio

n R

atio

L=8L=16

L=32

L=64

L=128

L=192

L=256

L=320

L=512

L=640

11 1 1 .

1

Number of collided slotsFrame size

n

ration

CL L

= = − − +−

⎛ ⎞ ⎛ ⎞⎜ ⎟ ⎜ ⎟⎝ ⎠ ⎝ ⎠

Prob. that there is the collision in a slot

1- Prob. that a tag transfers its ID successfully= =rate

coll

idle coll

CP

P P+

_

12.3922 .tags

opt rate

CC

==

Number of estimated tags 2.3922 collM= ×Where means the number of collided slots in a frame

collM

Process for tag estimation

21Jae-Hyun Kim


Performance evaluation

AchievementsApplying for the domestic patentApplying for the domestic patent"Novel Anti"Novel Anti--collision Algorithms for Fast Object Identification in RFID systcollision Algorithms for Fast Object Identification in RFID system," in em," in ProcProc. . ICPADS2005, Fukuoka, Japan, Jul. 20ICPADS2005, Fukuoka, Japan, Jul. 20--22, 2005, pp. 6322, 2005, pp. 63--67. 67. "Dynamic Framed Slotted ALOHA Algorithm using Fast Tag Estimatio"Dynamic Framed Slotted ALOHA Algorithm using Fast Tag Estimation method for RFID n method for RFID System," accepted by CCNC2006.System," accepted by CCNC2006.

Identification time vs. number of tags

0

20

40

60

80

100

120

140

100 200 300 400 500 600 700 800 900 1000

Nu mbe r o f t ags

Ide

nti

fic

ati

on

tim

e

2.3922 V ogt 1 V ogt 2

DFSA Slot - 128 Slot - 256

4

4.5

5

5.5

6

6.5

800 900 1000

Number of tags

Ide

nti

fic

ati

on

tim

e

2.3922 Vogt 1 Vogt 2 DFSA

22Jae-Hyun Kim


Deterministic binary tree (AutoID Class 0)

REPLY

TAG

X(0)

READER

REPLY

TAG READERTAGREADERTAGREADERREADER

TAG3(100)

TAG2(011)

TAG1(001)

STATE

X(0)REPLYCMDREPLYX(0)CMD

0

0

X

0

1

1

0

X 001 X

1 1

23Jae-Hyun Kim


Improved bit-by-bit binary tree algorithm (IBBT)Enhance the performance of the algorithm defined in AutoID classEnhance the performance of the algorithm defined in AutoID class 00Basic conceptBasic concept

By using bit Sequence information obtained from the tag, the unnecessary repetition of a reader’s ID request command can be reduced.

When there are tags whose last bit of ID is just differentWhen there are tags whose ID is sequential

REQ all bit

X0XX

1st REQ

(0)

0001

X(0)

1st REQ

X(0)

2nd REQ 0010

(0)

1st REQ

(1)

3rd REQ 1010

X(0)

1011

(1)

3rd REQ

(1)

(0)

2nd REQ

(0)

2nd REQ

(0) (1)

X0XX

X0XX

Basic procedure of IBBT

24Jae-Hyun Kim


0 20 40 60 80 100 120 140 160 180 2000

1000

2000

3000

4000

5000

6000

7000

8000

the number of tags

the

num

ber o

f ite

ratio

ns

BBT-36bit(analysis)BBT-36bit(simulation)BBT-25bit(analysis)BBT-25bit(simulation)BBT-19bit(analysis)BBT-19bit(simulation)IBBT-36bit,19bit(analysis)IBBT-36bit(simulation)IBBT-19bit(simulation)

Performance evaluation

Achievements"Improved Bit"Improved Bit--byby--bit Binary Tree Algorithm in Ubiquitous ID System," in bit Binary Tree Algorithm in Ubiquitous ID System," in ProcProc. IEEE PCM . IEEE PCM 2004, Tokyo, Japan, Nov. 292004, Tokyo, Japan, Nov. 29--Dec. 3, 2004, pp. 696Dec. 3, 2004, pp. 696--703. 703. "Fast Wireless Anti"Fast Wireless Anti--collision Algorithm in Ubiquitous ID System ," in collision Algorithm in Ubiquitous ID System ," in ProcProc. IEEE VTC 2004, . IEEE VTC 2004, L.A., USA, Sep. 26L.A., USA, Sep. 26--29, 2004. 29, 2004. "Novel Bit"Novel Bit--byby--bit Binary Tree Algorithm in Ubiquitous ID System," in bit Binary Tree Algorithm in Ubiquitous ID System," in ProcProc. WTC2004, Seoul, . WTC2004, Seoul, Korea, Sep. 12Korea, Sep. 12--15, 2004, pp. 51.15, 2004, pp. 51.

25Jae-Hyun Kim


Probabilistic binary tree (AutoID Class 1)Can select specific tags for identificationCan select specific tags for identificationUse 8 bin slots (3 bit info.) to singulate the tagsUse 8 bin slots (3 bit info.) to singulate the tags

Reader TAG

COMMANDPOINTER

VALUE

REQ.

STATUS

TAG 1

TAG 2

Bin 0(000)

Bin 1(001)

Bin2(010)

Bin 3(011)

Bin 4(100)

Bin 5(101)

Bin 6(110)

Bin 7(111)LENGTH

TAG 3

PingID0000 00000000 0100

1010

(1010001110101010)

(1010010101001010)

(1010010010011010)

00111010

01010100

0100100101001101

10100101

(1010001110101010)

(1010010101001010)

(1010010010011010)

IDLE IDLE SUCC IDLE IDLE SUCC IDLE IDLE

PingID0000 00000000 01111010010

IDLE SUCC COLL IDLE IDLE IDLE IDLE IDLE

(1010001110101010)

(1010010101001010)

(1010010010011010)

TAG 1 sends ITM ( Full ID )(1010001110101010)

(1010010101001010)

(1010010010011010)

ScrollID0000 00000000 01111010001

(1010001110101010)

(1010010101001010)

(1010010010011010)

26Jae-Hyun Kim


Fast Anti-collision Algorithm in EPC CLASS 1 UHF Enhance the performance of the algorithm defined in AutoID ClassEnhance the performance of the algorithm defined in AutoID Class 11Basic conceptBasic concept

By using 8-bit ( not 3bits ) information received from the tagReduce the number of the bin slot

By using bit Sequence information of whole ID obtained from the tag after reader’s receiving ScrollAllID command

When tags having a sequential ID, identification time can be decreasedBy applying the proposed tree-search algorithm

Fast tag inventory is possible

27Jae-Hyun Kim


Performance evaluationRandom tag IDRandom tag ID

AchievementsApplying for the domestic patent

20 40 60 80 100 120 140 160 180 2000

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

The number of used tags

Tag

iden

tific

atio

n tim

e(se

c)

Conventional EPC CLASS 1(analysis)Conventional EPC CLASS 1(simulation)Proposed algorithm(analysis)Proposed algorithm(simulation)Proposed algorithm using ScrollAllID(analysis)Proposed algorithm using ScrollAllID(simulation)

Applying for the domestic patent

28Jae-Hyun Kim


Performance evaluationSequential tag IDSequential tag ID

Achievements"Anti"Anti--collision algorithm using Bin slot in RFID System," accepted by collision algorithm using Bin slot in RFID System," accepted by IEEE TENCON '05IEEE TENCON '05"A Novel Tag identification algorithm for RFID System using UHF,"A Novel Tag identification algorithm for RFID System using UHF," accepted in EUC" accepted in EUC--05, 05, Nagasaki, Japan, 6Nagasaki, Japan, 6--9 Dec., 2005.

20 40 60 80 100 120 140 160 180 2000

1

2

3

4

5

6

The number of used tags

Tag

iden

tific

atio

n tim

e(se

c)

Conventional EPC CLASS 1Proposed algorithmProposed algorithm using ScrollAllID

100 120 140 160 180 2000.35

0.4

0.45

0.5

0.55

0.6

0.65

0.7

0.75

9 Dec., 2005.

29Jae-Hyun Kim

OnOn--going researchgoing research

Study on EPC Class 1 Gen 2 protocolDevelop the simulator for performance analysis of Class 1 Gen 2Develop the simulator for performance analysis of Class 1 Gen 2

Project with Samsung ThalesActual environment parameters applied to the simulatorMoving speed of tags and signal strength of the readerInterference ( tag-to-tag and reader-to-reader )

Network model for Gen 2

30Jae-Hyun Kim

EPC Class structure and issuesEPC Class structure and issues

Class IV

Class III

Class II

Class 0/ Class I

Class V

Passive tags and the reader talks first

Passive tags and reader talks first

Passive Field programmable tag, Encryption

Semi-Passive tags with sensors

Reader talks first, Encryption, etc.

Active tags and Encryption.

Both a reader and tags can start communication

Class IV capabilities and

Communication with passive tags

-. Energy-efficient MAC protocol-. Adaptive routing protocol for reliability-. Auto-configuration of the topology

-. Priority-dependent scheduling-. QoS provisioning-. Reliable data transmission

-. Simple and low-powered MAC-. Security and privacy problems

-. Tag and reader anti-collision algorithms-. Fast tag identification

-. Tag and reader anti-collision algorithms-. Fast tag identification

31Jae-Hyun Kim

Research plan and proposalResearch plan and proposal

Phase 1 : Focus on Class II to III (2005.11 ~ 2006. 10)Tag antiTag anti--collision algorithms for EPC Class 2 and 3collision algorithms for EPC Class 2 and 3

Tree based, probability based, and etc.Reader antiReader anti--collision algorithms for EPC Class 2 and 3collision algorithms for EPC Class 2 and 3

TDMA, Color-wave, and etc.LowLow--powered MAC for EPC Class 3powered MAC for EPC Class 3

Phase 2 : Focus on Class III to IV (2006.11 ~ 2007. 10)Simple and lowSimple and low--power MAC for EPC Class 3power MAC for EPC Class 3PriorityPriority--based scheduling schemebased scheduling schemeQoSQoS provisioning for environmental sensing data or emergent dataprovisioning for environmental sensing data or emergent dataReliable data transmission under the interfering environmentReliable data transmission under the interfering environment

32Jae-Hyun Kim

Research plan and proposalResearch plan and proposal

Phase 3 : Focus on Class IV to V (2007.11 ~ )Sensor MAC protocolSensor MAC protocol

Network life timeEnergy-efficient MAC protocolAdaptive routing protocol for reliability

High channel efficiency for multiple readers and tagsHigh channel efficiency for multiple readers and tagsAutoAuto--configuration of the topologyconfiguration of the topology

33Jae-Hyun Kim

E-mail : [email protected]

Home : Http://madang.ajou.ac.kr/~jkim

Thank you !!

MAC protocol issues beyond the passive RFID systemwinner.ajou.ac.kr/publication/data/invited/20051110_AutoIDLAB_JAE-HYUN-KIM.pdf · MAC protocol issues beyond the passive RFID system

Documents