mitigating the reader collision problem in rfid networks with
TRANSCRIPT
Mitigating the Reader Collision Problem in RFID Networks with Mobile Readers
Presented ByShailesh M. Birari
Guided ByProf. Sridhar Iyer
2
Basic Working of RFID system
Uses radio frequency to identify & track items in supply chain and manufacturing
RFID readers and tags Active and Passive tags
Motivation for Mobile Readers
Cost : “Always on” Fixed reader may be an overkill
Convenience : Easy, faster deployment No wiring installation hassles
Example applications : Searching a particular book in library Counting the items on the shelves in a super market Showing the list of items in the vicinity of the
customer in a super market
Scenario under consideration
Super market, library Each customer has a RFID reader Readers form an ad hoc network All readers have unrestricted mobility Readers often join and leave the network All tags are passive
Why a new protocol ?
TDMA : Interfering readers transmit in different timeslot Time synchronisation required Timeslot distribution is inefficient in a mobile
network CSMA : Sense channel before transmitting
RFID suffer from hidden terminal Collision happen at the tags and hence
collision detection is not possible by carrier sensing at the readers alone
Why a new protocol ?
FDMA : Interfering readers transmit at different frequency Tags do not have tuning circuitry Adding tuning circuitry to the tags will increase
the cost CDMA :
Requires complex circuitry at tags which will increase the cost of passive tags
CTSCTS
Why a new protocol ? (contd..)
RTS-CTS : Additional collision avoidance for CTS from
tags
A CTS from all the tags is required to ensure collision avoidance
RTS RTS T2T1 CTS CTSR1
T3RTST1
RTSR1RTSRTS R2T2
PULSE Protocol
Assumptions Dual channel : data and control channel Data channel : reader-tag communication Control channel : reader-reader communication A reader can receive simultaneously on both
channels but transmit on only one channel at a time
No inter-channel interference
PULSE Protocol Example
T1
Query
Beacon
Query QueryQuery QueryQuery QueryQuery
Beacon
R2T2 T3R1
QueryQuery QueryQuery
R1’s Read Range R2’s Read Range
PULSE Protocol Overview
Before communicating, a reader listens on the control channel for any beacon for Tmin time
If no beacon on the control channel for Tmin , start communication on the data channel
Reader periodically transmits a beacon on the control channel while communicating with the tags
Tmin
Contend_backoff
R1 chooses 2 BI, R2 chooses 5BI, R3 chooses 3BI
Tread2 1
5 4
3 2
TreadR1
R2
R3
2
5
3
Tmin
Tmin
Tmin
Tmin
Tmin3
2
5
1Tread
R1 chooses 3BI
Delay before beaconing
R2
R1
R3
R1‘s control channel Sensing range
R1‘s beacon range
Wait for control channel to get idle and then send beaconR1, R2, R3 are not in each others beacon rangeBoth R1 and R3 are communicating with tagsTransmit beacon immediatelyR1, R2 & R3 are communicating with the tagsChoose a small delay and then transmit
Simulation Setup (contd..)
Performance Metrics:
timeTotal
readers) all(by ly successfulsent queries Total Throughput System
readers allby collided) l(successfusent queries Total
100 readers allby ly successfulsent queries Total Efficiency System
Beacon Range Factor (BRF):
Poweron Transmissi Channel Data
Poweron Transmissi Channel Control BRF
Beacon Interval (BI) : interval after which beacon is sent Compared Protocols : CSMA, Colorwave, Aloha
System Throughput
25 Reader Topology :System Throughput with 25 Readers
0
1000
2000
3000
4000
5000
6000
7000
Aloha CSMA Colorw ave Pulse (BRF =28)
Mac protocols
Syste
m T
hro
ug
hp
ut
(Qu
eri
es/s
eco
nd
)
Static Readers
Mobile Readers
Pulse shows throughput improvement in both static and mobile networks
System Throughput (contd..)
Varying the number of readersSystem Throughput with Varying Number of Readers
0
1000
2000
3000
4000
5000
6000
7000
4 9 16 25 36 49 64
Number of Readers
Sys
tem
Th
rou
gh
pu
t (Q
uer
ies/
seco
nd
)
Aloha(Static)
CSMA(Static)
PULSE(Static)(BRF = 28)
Colorwave(Static)
Aloha(Mobile)
CSMA(Mobile)
PULSE(Mobile)(BRF = 28)
Colorwave(Mobile)
Pulse shows throughput improvement even at dense network of 64 readers
System Efficiency
25 Reader TopologySystem Efficiency with 25 Readers
0
10
20
30
40
50
60
70
80
90
100
Aloha CSMA Colorw ave Pulse (BRF =28)
Mac protocols
Sy
ste
m E
ffic
ien
cy
(Pe
rce
nta
ge
)
Static Readers
Mobile Readers
Pulse has system efficiency of above 95% which means Pulse is able to detect and avoid most of the collisions successfully
Optimal Beacon Interval (BI)
Effect of Beacon Interval on 25 reader topology
System Throughput with 25 Readers topology
0
1000
2000
3000
4000
5000
6000
7000
1 5 10 15
Beaconing Interval (msec)
Sys
tem
Th
rou
gh
pu
t (
Qu
erie
s/se
con
d)
Static Readers
Mobile Readers
System Efficiency with 25 Readers topology
98.4
98.6
98.8
99
99.2
99.4
99.6
99.8
100
1 5 10 15
Beaconing Interval (msec)
Syste
m E
ffic
ien
cy
(Perc
en
tag
e)
Static Readers
Mobile Readers
Variation in Beacon Interval does not show too much of difference in both system throughput and efficiency.
Optimal BRF
Throughput Vs BRF (Static Readers)System Throughput Vs BRF with Static Readers
0
1000
2000
3000
4000
5000
6000
7000
20 24 28 32
BRF for Pulse
Sys
tem
Th
rou
gh
pu
t (Q
uer
ies/
seco
nd
)
4 Readers
9 Readers
16 Readers
25 Redaers
36 Readers
49 Readers
64 Readers
BRF of 28 shows highest system throughput in almost all the networks
Optimal BRF (contd..)
Throughput Vs BRF (Mobile Readers)System Throughput Vs BRF with Mobile Readers
0
1000
2000
3000
4000
5000
6000
20 24 28 32
BRF for Pulse
Syste
m T
hro
ug
hp
ut
(Qu
eri
es/s
eco
nd
) 4 Readers
9 Readers
16 Readers
25 Redaers
36 Readers
49 Readers
64 Readers
BRF of 28 shows highest system throughput in almost all the networks
Optimal BRF (contd..)
Effect of Density of readers on networks with different BRFs
System Efficiency with Varying Readers
0
20
40
60
80
100
120
Number of Readers
Syste
m E
ffic
ien
cy
(Perc
en
tag
e)
Static BRF = 20
Static BRF = 24
Static BRF = 28
Static BRF = 32
Mobile BRF = 20
Mobile BRF = 24
Mobile BRF = 28
Mobile BRF = 32
Networks with BRF=28 maintain its efficiency above 95% even when the number of readers is increased to 64
Performance Modeling
Assume a beacon transmission is heard by all the readers
Backoff Decrement Interval: Interval after which backoff value is decremented
May contain a successful transmission by other reader
May contain a collision May be empty
Performance Modeling (contd..)
Cycle : Duration between two successful Tread transmission
by a reader Consists of BDIs
Calculate the average duration of a BDI Calculate the average number of BDIs in a
cycle Calculate the average duration of a cycle
Comparison
Comparison results
Comparison of Analysis and Simulation Results
Simulation
Analysis
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
4 9 16 25 36 49 64
Number of Readers
Sys
tem
Th
rou
gh
pu
t (
Qu
erie
s/se
con
d)
Analysis
Simulation
Conclusion
Mobile Readers reduce cost and improve convenience
Pulse shows an improvement in both the dimensions, system throughput and system efficiency
Pulse is effective even in dense mobile networks
References
[1] Daniel W. Engels. The Reader Collision Problem. Technical Report, epcglobal.org, 2002.
[2] J. Waldrop, D. W. Engels, and S. E. Sarma. Colowave: An anticollision algorithm for the reader collision problem. In IEEE Wireless Communications and Networking Conference (WCNC), 2003.
[3] QualNet Simulator 3.6. http://www.qualnet.com
[4] O. Tickoo and B. Sikdar. Queuing Analysis and Delay Mitigation in IEEE 802.11 Random Access MAC based Wireless Networks. In IEEE INFOCOM, 2004.
Existing Work ETSI EN 302 208 (CSMA):
Sense the data channel for 100msec before communicating the with tags
Colorwave (TDMA) : Readers randomly select a timeslot to transmit Chooses a new timeslot if collision and announce it to
neighbors UHF Gen 2 Standard (FDMA):
Separate reader transmissions and tag transmissions spectrally
Readers collide with readers and tags collide with tags
Approaches Considered
Registration at the access point (query response) Transmit Neighbour information to AP along
with request to transmit AP scans the status of the neighbours and
responds accordingly
Centralised graph coloring at Access Point All nodes transmit neighbour information to the
AP AP applies a graph coloring to allocate time-
slots
Approaches Considered (contd.)
Interesting Features of RCP Readers may not be in each others sensing range; Tag cannot select a particular reader to
respond(unlike cellular systems) None of the readers can read the tag The passive tags, where the collision may take
place, are not able to take part in the collision resolution as in hidden terminal problem
Reduces the read rate of the RFID system