An Adaptive Pre-detection Based RFID Tag Anti-collision Scheme

Chiu-Kuo Liang and Yuan-Cheng Chien

Dept. of Computer Science and Information Engineering, Chung Hua University, Hsinchu, Taiwan, R.O.C.

Keywords: Tag Anti-Collision, Hybrid Query Tree, Pre-Detection Query Tree.

Abstract: One of the research areas in RFID systems is a tag anti-collision protocol; how to reduce identification time

with a given number of tags in the field of an RFID reader. There are two types of tag anti-collision

protocols for RFID systems: tree based algorithms and slotted aloha based algorithms. Many anti-collision

algorithms have been proposed in recent years, especially in tree based protocols. However, there still have

challenges on enhancing the system throughput and stability due to the underlying technologies had faced

different limitation in system performance when network density is high. Particularly, the tree based

protocols had faced the long identification delay. Recently, a Hybrid Hyper Query Tree (H

QT) protocol,

which is a tree based approach, was proposed and aiming to speedup tag identification in large scale RFID

systems. The main idea of H

QT is to track the tag response and try to predict the distribution of tag IDs in

order to reduce collisions. In this paper, we propose a pre-detection tree based algorithm, called the

Adaptive Pre-Detection Based Query Tree algorithm (APDBQT), to avoid those unnecessary queries. Our

proposed APDBQT protocol can reduce not only the collisions but the idle cycles as well by using pre-

detection mechanism. The simulation results show that our proposed technique provides superior

performance in high density environments. It is shown that the APDBQT is effective in terms of increasing

system throughput and minimizing identification delay.

1 INTRODUCTION

Radio Frequency IDentification (RFID) is an

automatic technology that guarantees to advance

modern industrial practices in object identification

and tracking, asset management, and inventory

control (Vogt, 2002). Recently, several identification

systems such as barcodes and smart cards are

incorporated for automatic identification and data

collection. However, these systems have several

limits in read rate, visibility, and contact. RFID

systems are a matter of great concern because they

provide fast and reliable communication without

requiring physical sight or touching between readers

and tags.

One of the areas of research is the speed with

which a given number of tags in the field of RFID

readers can be identified. For fast tag identification,

anti-collision protocols, which reduce collisions and

identify tags irrespective of occurring collisions, are

required (Vogt, 2002; Myung and Lee, 2005; Myung

et al., 2006; Law et al., 2000; Lee et al., 2005;

Capetanakis, 1979; Zhou et al., 2003). There are two

types of collisions: reader collisions and tag

collisions. Reader collisions indicate that when

neighboring readers inquire a tag concurrently, so

the tag cannot respond its ID to the inquiries of the

readers. These collision problems can be easily

solved by detecting collisions and communicating

with other readers. Tag collisions occur when multi

tags try to respond to a reader simultaneously and

cause the reader to identify no tag. For low-cost

passive RFID tags, there is nothing to do except

response to the inquiry of the reader. Thus, tag anti-

collision protocols are necessary for improving the

cognitive faculty of RFID systems.

In general, the tag anti-collision techniques can

be classified into two categories, aloha-based and

tree-based protocols. Aloha-based approaches use

time slot to reduce collision probability, such as

Framed-Slotted aloha algorithm (Vogt, 2002; Park et

al., 2007), dynamic framed slotted aloha algorithm

(Lee et al., 2005). Tags randomly select a particular

slot in the time frame, load and transmit its

identification to the reader. Once the transmission is

collided, tags will repeatedly send its id in next

interval of time to make sure its id is successfully

recognized. Aloha-based protocols can reduce the

collision probability. However, they have the tag

starvation problem that a particular tag may not be

identified for a long time. For the consideration of

Liang C. and Chien Y..

An Adaptive Pre-detection Based RFID Tag Anti-collision Scheme.

DOI: 10.5220/0005268800690076

In Proceedings of the 4th International Conference on Sensor Networks (SENSORNETS-2015), pages 69-76

ISBN: 978-989-758-086-4

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

performance, when number of RFID tag increased,

the tag collision rate will be increased as well; this

may result a low tag recognition rate.

The tree-based schemes use a data structure

similar to a binary search algorithm, such as binary

tree splitting protocol (Myung et al., 2006), query

tree (QT) algorithm, and tree working algorithm

(Capetanakis, 1979; Feng et al., 2006). An RFID

reader consecutively communicates with tags by

sending prefix codes based on the query tree data

structure. Only tags in the reader’s interrogation

zone and of which ID match the prefix respond. The

reader can identify a tag if only one tag respond the

inquiry. Otherwise the tags responses will be

collided if multiple tags respond simultaneously.

Although tree based protocols deliver 100%

guaranteed read rates, but they have relatively long

identification delay. Recently, a hybrid query tree

protocol (HQT) (Ryu et al., 2007) was proposed and

aiming to reduce transmission overhead by using 4-

ary search tree mechanism and slotted backoff

mechanism, in order to speed up tag identification

and to increase the overall read rate and throughput

in large-scale RFID systems. The main idea of the

HQT technique is to reduce the number of collisions

during the identification phase. In the 4-ary search

tree mechanism, the prefix string of a collided query

will be extended by 2-bits next time, unlike of 1-bit

in the QT protocol. This way, collisions can be

reduced substantially. Furthermore, the HQT

protocol was aiming to reduce the idle cycles by

using a slotted backoff mechanism. When a tag

responds to a reader, it sets its backoff timer using a

part of its ID. If there is a collision (multiple tags

respond), the reader can partially deduce how the

IDs of tags are distributed and potentially reduce

unnecessary queries.

Based on the HQT protocol, a H

QT protocol

(Kim and Lee, 2009) was proposed and aiming to

reduce the idle cycles and improve the performance

of tag identification. Although the H

QT technique

performs better than the HQT technique in reducing

the number of idle cycles, it still has some idle

cycles, which cannot be reduced during the tag

identification process. In this paper, we proposed a

pre-detection based protocol, called Adaptive Pre-

Detection Based Query Tree (APDBQT) protocol, to

eliminate those unnecessary idle cycles. To evaluate

the performance of our proposed technique, we have

implemented our proposed APDBQT scheme along

with previous proposed methods, HQT and H

protocols. The experimental results show that the

proposed technique presents significant

improvement in most circumstance.

The remainder of this paper is organized as

follows: Related work is discussed in Section II. In

Section III, the tree based tag identification

algorithm is introduced as preliminary of this study.

In Section IV, our proposed algorithm, the APDBQT

algorithm is presented. Performance comparisons

and analysis of the proposed technique will be given

in Section V. Finally, in Section VI, some

concluding remarks are made.

2 RELATED WORK

Many research results for collision avoidance have

been presented in literature. The existing tag

identification approaches can be classified into two

main categories, the Aloha-based anti-collision

scheme (Vogt, 2002; Law et al., 2000; Lee et al.,

2005; Park et al., 2007; Klair et al., 2007) and the

tree-based scheme (myung et al., 2006; Capetanakis,

1979; Zhou et al., 2003; Feng et al., 2006). RFID

readers in the former scheme create a frame with a

certain number of time slots, and then add the frame

length into the inquiry message sent to the tags in its

vicinity. Tags response the interrogation based on a

random time slot. Because collisions may happen at

the time slot when two or more tag response

simultaneously, making those tags could not be

recognized. Therefore, the readers have to send

inquiries contiguously until all tags are identified.

As a result, Aloha-based scheme might have long

processing latency in identifying large-scale RFID

systems (Law et al., 2000). In (Vogt, 2002), Vogt et

al. investigated how to recognize multiple RFID tags

within the reader’s interrogation ranges without

knowing the number of tags in advance by using

framed Aloha. A similar research is also presented in

(Zhen et al., 2005) by Zhen et al. In (Klair et al.,

2007), Klair et al. also presented a detailed

analytical methodology and an in-depth qualitative

energy consumption analysis of pure and slotted

Aloha anti-collision protocols. Another anti-collision

algorithm called enhanced dynamic framed slotted

aloha (EDFSA) is proposed in (Lee et al., 2005).

EDFSA estimates the number of unread tags first

and adjusts the number of responding tags or the

frame size to give the optimal system efficiency.

In tree-based scheme, such as ABS (Myung et al.,

2006), Improved Bit-by-bit Binary-Tree (IBBT)

(Choi et al., 2004) and IQT (Sahoo et al., 2006),

RFID readers split the set of tags into two subsets

and labeled them by binary numbers. The reader

repeats such process until each subset has only one

tag. Thus, the reader is able to identify all tags. The

SENSORNETS2015-4thInternationalConferenceonSensorNetworks

adaptive memoryless tag anti-collision protocol

proposed by Myung et al. (Myung and Lee, 2005) is

an extended technique based on the query tree

protocol. Choi et al. (Choi et al., 2004) also

proposed the IBBT (Improved Bit-by-bit Binary-

Tree) algorithm in Ubiquitous ID system and

evaluate the performance along three other old

schemes. The IQT protocol (Sahoo et al., 2006) is a

similar work approach by exploiting specific prefix

patterns in the tags to make the entire identification

process. Recently, Zhou et al. (Zhou et al., 2007)

consider the problem of slotted scheduled access of

RFID tags in a multiple reader environment. They

developed centralized algorithms in a slotted time

model to read all the tags. With the fact of NP-hard

(Zhou et al., 2007), they also designed

approximation algorithms for the single channel and

heuristic algorithms for the multiple channel cases.

Although tree based schemes have advantage of

implementation simplicity and better response time

compare with the Aloha based ones, they still have

challenges in decreasing the identification latency.

In this paper, we present an enhanced tree based tag

identification technique aims to coordinate

simultaneous communications in large-scale RFID

systems, to speedup minimize tag identification

latency and to increase the overall read rate and

throughput. Simulation results show that our

proposed technique outperforms previous techniques.

3 TREE BASED

ANTI-COLLISION SCHEMES

In this section, we present two tree-based anti-

collision techniques, namely HQT algorithm (Ryu et

al., 2007), and the H

QT algorithm (Kim and Lee,

2009), that are most related to our work.

3.1 Hybrid Query Tree Algorithm

In the environment with high tags density, collision

may happen very frequently while using the query

tree algorithm, and due to that, a lot of query time

will be wasted. By using the 4-ary search query tree

mechanism, HQT can enable the prefix to increase

two bits at a time from 1 bit. In this way, some

collisions occurred in QT protocol can be reduced in

HQT protocol. However, the drawback of the 4-ary

search tree mechanism is the increasing number of

idle cycles. To resolve this problem, HQT protocol

introduces the slotted backoff mechanism. The

slotted backoff mechanism is a technique that makes

tags respond to the transmit prefix after waiting a

certain time, instead of immediately respond. When

a tag responds to a reader, it sets its backoff timer

using a part of its ID. The backoff time of each tag is

determined from the 2-bits, which follow the prefix

of tag ID identical to the query prefix string. For

example, tags do not defer their response if it is ‘00’.

If it is ‘01’, ‘10’, or ‘11’, tags will defer 1, 2, or 3

backoff time slots until they respond to the reader,

respectively. Fig. 1 shows the operation of the

slotted backoff mechanism in HQT algorithm.

Figure 1: The slotted backoff mechanism in HQT.

3.2 Hybrid Hyper Query Tree

Algorithm

The main problem in HQT algorithm is that those

idle cycles between busy slots cannot be reduced. To

resolve the problem, the H

QT algorithm uses a

different slotted backoff mechanism. The backoff

time of each tag is determined from the 3-bits, which

follows the prefix of tag ID identical to the prefix.

Unlike the mechanism used in HQT, the H

counts the number of ‘1’ in the following 3-bits and

uses this number as the selected time slot for tags to

respond. Fig. 2 shows the tag selecting its response

slot based on tag ID.

Figure 2: H

QT algorithm.

Fig. 3 depicts an example of the query tree

structure of identifying 5 tags with 6-bits ID length

using H

QT algorithm. The process of the

identification is as follows: First of all, the reader

sends request command with the empty-prefix to the

AnAdaptivePre-detectionBasedRFIDTagAnti-collisionScheme

tags. In this case, tags A, C and D will delay one

time slot to respond since the first 3-bits of their tag

IDs contain only one ‘1’, as shown in Fig. 4.

Similarly, tags B and E will delay two time slots to

respond due to the number of ‘1’ in the first 3-bits if

their tag IDs is 2. In this case, since no tag responds

immediately, it means that there is no tag whose first

3-bits of their tag IDs match ‘000’. Therefore, there

is no need for reader to send the prefix string ‘000’.

Similarly, since no tag responds after 3 time slot

delay, the reader does not need to send the prefix

string ‘111’. Therefore, the idle cycles can be

eliminated.

Next, the reader receives tag IDs from tags A, C

and D after one time slot delay. At this moment, the

reader is aware that the pattern of the first 3-bits of

tags A, C and D is ‘X0X’, in which ‘X’ represents a

collision bit. Thus, the reader recognizes that the

first 3-bits of tag A, C and D may be ‘001’ or ‘100’,

which will be added into the queue for re-

transmission. Similarly, the reader is aware that the

bit pattern of the first 3-bits of tags B and E is ‘XX1’

after two time slots delay. Thus, the reader will put

prefix strings ‘011’ and ‘101’ into the queue for re-

transmission.

Next, the reader then sends the request command

with prefix string ‘001’. At this moment, only tag A

responds after 1 time slot delay. In this case, tag A is

identified by the reader. Table 1 summarizes the

detail steps of communication between the reader

and the tags with the example shown in Fig. 3.

Figure 3: An example of H

QT algorithm.

Figure 4: The response of tags after reader’s empty prefix

request in Figure 3.

Table 1: Communication steps of Figure 3.

Step

HQT H

Broadcast Status

Broadcas

Status

1 empty Collision empty Collision

2 00 Identify Tag A 001 Identify Tag A

3 01 Identify Tag B 100 Collision

4 10 Collision 011 Identify Tag B

5 1000 Collision 101 Identify Tag E

6 1001 Idle 100001

Identify Tag C

7 1010 Identify Tag E 100010

Identify Tag D

8 100001 Identify Tag C

9 100010 Identify Tag D

4 THE PROPOSED SCHEME

Recall that, in H

QT algorithm, the idle cycles can

be reduced substantially. However, there still have

some collision time slots. As a result, the reader has

to spend more time slots to resolve the collisions.

Due to that, it will take more time to complete the

tag identification process. In this paper, we proposed

a pre-detection scheme to eliminate the collision

time slots and idle cycles.

4.1 Adaptive Pre-Detection based

Query Tree Algorithm

We proposed an APDBQT algorithm, which uses

pre-detection technique to realize the precise

distribution of tag IDs. Once the distribution of tag

IDs has been obtained, the reader broadcasts such

message to tags and each tag is aware of the exact

time slot to response. As a result, tags respond to the

reader in different time slots and collisions can be

avoided. Furthermore, since each tag realizes its

corresponding time slot to respond, no empty time

slot exists.

In our proposed APDBQT algorithm, after the

reader send the request command to tags, the

operations during the tag response period can be

partitioned into three phases: the pre-detection phase,

the broadcasting phase and the tag response phase,

as shown in Fig. 5. The purposes of three phases

design can be explained as follows: In pre-detection

phase, the reader can realize the distribution of tag

IDs by collecting the responses from tags. Then, in

broadcasting phase, the reader will send the

distribution information to tags so that each tag is

aware of the time slot to send its ID to reader.

Finally, in the tag response phase, the responses

from tags are arranged into a sequence of time slots

so that collisions and empty slots can be avoided.

SENSORNETS2015-4thInternationalConferenceonSensorNetworks

Figure 5: The tag response cycle of our proposed scheme.

After the reader sends a query string (i.e. a prefix

string) and a parameter m to tags and then waits for

tags to response, in the pre-detection phase, each tag

whose tag ID matches with the prefix string sent

from the reader will respond on the pre-detection

time slot according to its following m-bits of its tag

ID. In order to collect the response information from

tags, we allocate 2

short time slots in pre-detection

phase for tags to respond and the time slots can be

numbered as the binary representation of m bits

respectively. Fig. 5 shows an example with m = 2.

Therefore, the pre-detection phase in Fig. 5 consists

of four time slots, namely the ‘00’, ‘01’, ‘10’, and

‘11’ time slots respectively. We also adapt the m-ary

search tree mechanism such that each tag whose tag

ID matches with the prefix string sent from the

reader will respond on the corresponding time slot

depending to its following m-bits of its tag ID. It

should be noticed that, in this phase, each tag

responds a 4-bits random number (RN) to reader

instead of the whole tag ID. The reasons for tags of

using 4-bits random numbers to respond are as

follows: First, it can reduce the time for reader to

realize the distribution of tag IDs, compared with the

response of whole tag IDs. Second, the status of

each time slot can be precisely identified with high

probability. If no tag responds in a time slot, then the

reader can correctly identify such time slot as an idle

cycle, which can be eliminated during the tag

response phase. If only one tag responds, the reader

can also correctly identify such time slot as a

successful cycle. Therefore, the reader will allocate a

time slot to receive the response from that tag in the

tag response phase. If more than one tag responds,

since the tags respond 4-bits random numbers, a

collision cycle can be identified by the reader by

checking the received different random numbers.

Although, there still has some chance for a reader to

receive the same random number from different tags,

however, the probability of successful collision

detection is very high. Therefore, by using our pre-

detection mechanism, the distribution of tag IDs can

be correctly obtained with high accuracy.

Meanwhile, the reader monitors and records the

response status from tags in each time slot during the

pre-detection phase. The reader uses a ‘0’-bit to

represent the time slot when no tag responds or more

than one tag respond. On the other hand, the reader

uses a ‘1’-bit to represent the time slot when only

one tag responds. Therefore, after the pre-detection

phase, the reader can use an m-bits string to

represent the status of 2

time slots in the pre-

detection phase. Then, during the broadcasting phase,

the reader broadcasts the m-bits string to tags and by

receiving the binary bit string, each tag can realize

the exact time slot to respond by counting the

number of ‘1’ in the received binary bit string from

the start bit to its corresponding bit. Then, the tag

can respond its tag ID to reader in the tag response

time slot by finding the correct time slot to respond.

For example, in Fig. 5, the tag which responds on

the ‘11’ time slot in the pre-detection phase can

realize that it can only send its tag ID on the third

time slot in the tag response phase since it receives

the binary bit string ‘1101’ sent from the reader and

there are three ‘1’s from the beginning to its

corresponding ‘1’.

After the tag response phase, the reader will re-

calculate the value of m depending on the idle slot

ratio in previous pre-detection phase. The idle slot

ratio is defined as the number of idle time slots to

the total number of time slots in previous pre-

detection phase. Thus, as the idle slot ratio gets high,

meaning that the density of tag distribution gets low

and vice versa. Therefore, the reader may allocate

more time slots in pre-detection phase by letting m

larger in the next cycle of identification when the

reader is aware of that the density of tags is getting

higher. Similarly, the reader may allocate less time

slots in pre-detection phase by letting m smaller in

the next cycle of identification when the reader is

aware of that the density of tags is getting lower.

4.2 An Example

To facilitate the understanding of our proposed

algorithm, an example is given as follows.

Fig. 6 depicts the example of the process of

identifying 9 tags with 6-bits of tag IDs, namely

A:‘000010’, B:‘001001’, C:‘001010’, D:‘011001’,

E:‘100001’, F:‘100010’, G:‘101011’, H:‘110011’,

and I:‘110111’, respectively, by using APDBQT

protocol. The process of identification is as follows:

First of all, the reader sends the request command

with the empty-prefix and m = 2 to the tags. In this

case, all tags respond to this request command and

the time slot for a tag to respond is depending on the

first 2-bits of its tag ID. In this example, tags A, B

and C will respond in ‘00’ time slot, tag D will

respond in ‘01’ time slot, tags E, F, and G will

AnAdaptivePre-detectionBasedRFIDTagAnti-collisionScheme

respond in ‘10’ time slot, and tags H and I will

respond in ‘11’ time slot, as shown in Fig. 6(a).

Suppose that the random numbers for tags to

respond are all different. It can easily be seen that,

since there is only one tag response for ‘01’ time slot,

the reader will mark the time slot as ‘1’.

Furthermore, since it has more than one tag

responses in ‘00’, ‘10’, and ‘11’ time slots, the

reader will mark these time slots as ‘0’. It should be

noticed that as the reader realizes that there is no idle

time slot in the pre-detection phase, the value of m

will be increased to 3. Meanwhile, as the reader

(a)

(b)

(c)

(d)

Figure 6: An example pf our APDBQT algorithm.

recognizes the collision time slot, the corresponding

prefix bit string will be added into a queue for

further requesting. In this example, ‘00’, ‘10’ and

‘11’ bit strings will be added into the queue, along

with the value of m. After the pre-detection phase,

the reader will mark all time slots as ‘0100’ and

broadcast it to tags. After tags receive the message,

tag D realizes its own time slot to respond.

Therefore, tag D will be identified. In the meantime,

all other tags recognize that the status of their time

slot is ‘0’, which means that they do not need to

send their tag IDs to reader at that time slot, as

shown in Fig. 6(a). After identifying tag D, the

reader sends another request command and m from

queue, which is the ‘10’ bit string and m = 3 in this

example as shown in Fig. 6(b). In this cycle, tags A,

B and C are identified. In the next round, as shown

in Fig. 6(c), tags E, F, and G can be identified. In the

last round, as shown in Fig. 6(d), tags H and I can be

identified.

4.3 Comparison of Tag Identification

Methods

To facilitate the understanding of the performance of

our proposed algorithm, we compare the

identification process between previous H

QT and

our proposed APDBQT algorithms by using the

example in Fig. 6.

Table 2 shows required prefixes and steps for

identifying all 9 tags by using different methods. In

Table 2, the H

QT scheme needs 10 steps to

complete the identification process while in our

proposed APDBQT scheme, only 4 steps are needed.

Thus, our proposed APDBQT protocol reduces

identification overhead efficiently and achieves

better performance than H

QT scheme.

Table 2: Communication steps of Figure 6.

Step

QT APDBQT

Broadcast Status Broadcast Status

1 empty Identify Tag A empty Identify Tag D

2 001 Collision 00

Identify Tags A,

B and C

3 100 Collision 10

dentify Tags E, F

and G

4 011 Identify Tag D 11

Identify Tags H

and I

5 101 Identify Tag G

6 110

Identify Tags H

and I

7 001001 Identify Tag B

8 001010 Identify Tag C

9 100001 Identify Tag E

10 100010 Identify Tag F

SENSORNETS2015-4thInternationalConferenceonSensorNetworks

5 PERFORMANCE

EVALUATION

To evaluate the performance of the proposed

technique, we implemented the APDBQT scheme

along with the HQT algorithm and the H

algorithm. In the interrogation zone, we increase the

number of tags from 500 to 4000. All tags are

randomly generated in a uniform distribution

manner. The lengths of the tag IDs used in each

experiment are 96 bits. It should be noticed that

some overhead are not taken into account in our

simulation due to the communication latency and the

propagation delay from the signal processing on the

channel.

Fig. 7 shows the number of queries needed for

reader to complete the tags identification. We can

observe that, as the number of tags increases, each

algorithm increases linearly due to the number of

collision increases. However, our proposed

APDBQT scheme requires less number of queries

compared with other schemes.

Fig. 8 shows the number of idle cycles generated

by each algorithm during the tag identification

process. We can observe that, both H

QT and our

proposed APDBQT algorithm can eliminate all idle

cycles regardless the number of tags increases.

Figure 7: No. of queries required to complete

identification.

Figure 8: No. of idle cycles generated by each algorithm.

Fig. 9 shows the number of collisions generated

by each algorithm during the tag identification

process. We can observe that our proposed

APDBQT algorithm generates much fewer collisions

than both HQT and H

QT algorithms. Due to the

pre-detection mechanism, most collisions can be

detected in the pre-detection phase, there are only a

few time slots wasted in the tag response phase.

Figure 9: No. of collisions generated by each algorithm.

Figure 10: The time required to complete tag identification.

Fig. 10 shows the total time required for each

algorithm to complete the tag identification process.

We can observe that our proposed APDBQT

algorithm needs less time than both HQT and H

algorithms to complete tag identification. Thus, the

APDBQT algorithm outperforms the HQT and

QT algorithms.

6 CONCLUSIONS

With the emergence of wireless RFID technologies,

identifying high density RFID tags is a crucial task

in developing large scale RFID systems. Due to the

nature of large scale RFID systems, many collisions

may occur during the process of tag identification. In

this paper, we proposed a nearly collision-free tag

identification algorithm to reduce the iteration

AnAdaptivePre-detectionBasedRFIDTagAnti-collisionScheme

overhead efficiently. By using the random numbers

for tags to respond in the pre-detection phase, many

unnecessary collided inquiries can be reduced and

the efficiency of tag identification can be

significantly improved. To evaluate the performance

of proposed techniques, we have implemented the

APDBQT technique along with previous HQT and

QT algorithms. The experimental results show

that the proposed technique provides considerable

improvements on the latency of tag identification. It

is also shown that the APDBQT is effective in terms

of increasing system throughput and efficiency.

REFERENCES

Vogt, H., 2002. Efficient Object Identification with

Passive RFID Tags. In Proc. Inter. Conf. on Pervasive

Computing, Pervasive’ 02. pp. 98-113, London, UK,

2002, Springer-Verlag.

Myung, J., Lee, W., 2005. An adaptive memoryless tag

anticollision protocol for RFID networks. In IEEE

INFOCOM’05, Poster Session.

Myung, J., Lee, W., Srivastava, J., 2006. Adaptive binary

splitting for efficient RFID tag anti-collision. In IEEE

Communication Letter, vol. 10, no. 3, pp. 144–146.

Law, C., Lee, K., Siu, K. Y., 2000. Efficient Memoryless

Protocol for Tag Identification. In proceedings of the

International Workshop on Discrete Algorithms and

Methods for Mobile Computing and Communications.

Lee, S., Joo, S., Lee, C., 2005. An enhanced dynamic

framed slotted ALOHA algorithm for RFID tag

identification. In Proc. MobiQuitous, pp. 166-172.

Capetanakis, J. I., 1979. Tree algorithms for packet

broadcast channels. In IEEE Trans. Information

Theory, vol. 25, pp. 505-515.

Zhou, F., Jin, D., Huang, C., Hao, M., 2003. Optimize the

Power Consumption of Passive Electronic Tags for

Anti-collision Schemes. In Proc. The 5th Inter. Conf.

on ASIC, Vol. 2, pp.1213-1217.

Park, J., Chung, M., Lee, T. J., 2007. Identification of

RFID Tags in Framed-Slotted ALOHA with Robust

Estimation and Binary Selection. In IEEE

Communications Letters, vol. 11, no.5, pp. 452-454.

Feng, B., Tao, L. J., Bo, G. J., Hua, D. Z., 2006. Id-binary

tree stack anti-collision algorithm for RFID. In IEEE

Transactions on Computers, pp. 207–212.

Ryu, J., Lee, H., Seok, Y., Kwon, T., Choi, Y., 2007. A

Hybrid Query Tree Protocol for Tag Collision

Arbitration in RFID systems. In IEEE International

Conference on Communications (ICC-07), pp. 24–28.

Kim, T. H., Lee, S. J., 2009. A Hybrid Hyper Tag Anti-

Collision Algorithm in RFID System. In ICACT, pp.

1276-1281.

Klair, D. K., Chin, K. W., Raad, R., 2007. An

investigation into the energy efficiency of pure and

slotted aloha based RFID anticollision protocols. In

proceedings of the IEEE WoWMoM’07, Finland.

Zhen, B., Kobayashi, M., Shimizui, M., 2005. Framed

aloha for multiple RFID objects Identification. In

IEICE Trans. on Comm, E88-B(3), pp. 991–999.

Choi, H. S., Cha, J. R., Kim, J. H., 2004. Improved Bit-by-

bit Binary Tree Algorithm in Ubiquitous ID System.

In proceedings of the IEEE PCM2004, Tokyo, Japan,

pp. 696–703.

Sahoo, A., Iyer, S., Bhandari, N., 2006. Improving RFID

System to Read Tags Efficiently. In KRSIT Technical

Report.

Zhou, Z., Gupta, H., Das, S. R., Zhu, X., 2007. Slotted

Scheduled Tag Access in Multi-Reader RFID Systems.

In proceedings of the IEEE International Conference

on Networks Protocols (ICNP), pp. 61–70.

SENSORNETS2015-4thInternationalConferenceonSensorNetworks