An Improved IEEE 802.15.4 Superframe Structure with Minimum Delay and Maximum CFP Link Utilization

Abstract

An efficient superframe structure for 802.15.4 Medium Access Control (MAC) layer that operates over 868, 915 and 2400MHz, is proposed in this paper. In this superframe structure, Contention Free Period (CFP) precedes the Contention Access Period (CAP) and more number of slots are used in the same CFP period as of original 802.15.4 standard. As CFP precedes the CAP, the communication delay for the CFP traffic is exceptionally reduced. The Beacon frame is fine-tuned to achieve the above said superframe structure and makes it backward compatible with the original standard. Due to large number of small slots in CFP, more small amount of data requesting nodes can be assigned CFP space for communication. The analytical results show that our proposed superframe structure has nearly 50% less delay, accommodates almost double the number of CFP requesting nodes and has better link utilization compared to the original

Full text

Ad Hoc & Sensor Wireless Networks, Vol. 35, pp. 151–171 ©2017 Old City Publishing, Inc.
Reprints available directly from the publisher Published by license under the OCP Science imprint,
Photocopying permitted by license only a member of the Old City Publishing Group.
An Improved IEEE 802.15.4 Superframe
Structure with Minimum Delay and
Maximum CFP Link Utilization
AHMAD NASEEM ALVI1,S.H.AHMED2,∗,M.A.YAQUB2,N.JAVAID1,
SAFDAR H. BOUK2AND DONGKYUN KIM2
1Dept. of Electrical Engineering, COMSATS Inst. of Info. Technology, Islamabad, Pakistan
E-mail: naseem alvi@comsats.edu.pk; nadeemjavaid@comsats.edu.pk
2School of Computer Science and Engineering, Kyungpook National University, Daegu, Korea
E-mail: yaqub@knu.ac.kr; bouk@knu.ac.kr; dongkyun@knu.ac.kr
Received: May 07, 2014. Accepted: December 08, 2016.
An efficient superframe structure for 802.15.4 Medium Access Control
(MAC) layer that operates over 868, 915 and 2400MHz, is proposed in
this paper. In this superframe structure, Contention Free Period (CFP)
precedes the Contention Access Period (CAP) and more number of slots
are used in the same CFP period as of original 802.15.4 standard. As
CFP precedes the CAP, the communication delay for the CFP traffic
is exceptionally reduced. The Beacon frame is fine-tuned to achieve
the above said superframe structure and makes it backward compati-
ble with the original standard. Due to large number of small slots in
CFP, more small amount of data requesting nodes can be assigned CFP
space for communication. The analytical results show that our proposed
superframe structure has nearly 50% less delay, accommodates almost
double the number of CFP requesting nodes and has better link utiliza-
tion compared to the original 802.15.4 standard for all three frequency
bands.
Keywords: Superframe structure, contention free period, contention access
period, beacon frame, 802.15.4.
1 INTRODUCTION
Wireless Sensor Networks (WSNs) have been main focus of global research
community due to their wide range of applications. The major application
∗Contact author: E-mail: hassan@knu.ac.kr
151

152 AHMAD NASEEM ALVI et al.
Beacon
Conte nti on Acce ss pe riod
(CAP)
Conte nti on Free Period
(CFP)
Inactive Period
Beacon
Active Period
Beacon Interval (BI)
Superframe Duration(SD)
FIGURE 1
802.15.4 Beacon enabled mode Superframe format.
areas of WSN include healthcare, military, environmental monitoring, civil
engineering, etc. [1]. WSN comprises standalone, autonomous and tiny bat-
tery operated wireless nodes with limited energy, computation, processing
and communication capabilities. These constraints led the emergence of new
physical and Medium Access Control (MAC) stack named IEEE 802.15.4,
because the existing communications stacks i.e. IEEE 802.11 and 802.16,
were not designed to function under these constraints. IEEE 802.15.4 stan-
dard was designed for Low Rate Wireless Personal Area Network (LR-
WPAN) applications where low data rate, higher reliability with less power
consumption is required [2]. The standard is suitable for fixed as well as for
low cost mobile wireless nodes with limited battery power with high reliabil-
ity [3].
In most of the applications, sensor nodes are intended to operate
autonomously on the battery, therefore, the WSN protocols should be energy
efficient to prolong node’s life-time. The large proportion of battery con-
sumption is due to communication (transmission and reception) over wire-
less radio component in a node. The MAC protocol is responsible to control
the radio transceiver and in 802.15.4, MAC layer keeps sensor nodes in sleep
state for more than 99% with duty cycle even less than 0.1%.
IEEE 802.15.4 operates at three different frequency bands such as
868MHz, 915MHz and 2.4GHz and works either in a Beacon enabled or Non-
Beacon enabled mode. The Beacon enabled mode is divided into two main
sections, active and inactive period, as shown in Figure 1. All WSN nodes
communicate during active period and remain in sleep mode during later
inactive period to conserve energy. The active period of Beacon enabled mode
consists of Contention Access Period (CAP) and optional Contention Free
Period (CFP). Each superframe in this mode is divided in to 16 equal dura-
tion time slots. One or more slots are reserved for the Beacon frame because
its size may vary due to number of remaining data frames for the associ-
ated nodes. The Beacon frame is generated by the PAN coordinator and con-
tains information about frame structure, next Beacon, network, and pending
messages.

NEW SUPERFRAME STRUCTURE WITH MINIMUM DELAY AND MAXIMUM CFP LINK 153
The CAP consists of maximum 16 or minimum 9 slots. In CAP, nodes
contend to access medium by following slotted CSMA/CA mechanism [4].
On the other hand, the maximum number of slots in CFP can be up to 7
and are known as Guaranteed Time Slots (GTS). Nodes having critical data
requests are allocated Guaranteed Time Slot (GTS) by the coordinator. The
nodes that are allocated GTS can explicitly carry out communication during
their allocated period to the PAN coordinator. All research related to max-
imize throughput and reduce communication delay of the traffic in CFP by
considering the above mentioned scenario mentioned in the 802.15.4 stan-
dard [2]. However, this standard has some limitations for GTS allocations.
1. The cumulative delay from GTS allocation till transferring of data causes
a significant delay, which is inappropriate for time sensitive WSN appli-
cations.
2. Due to limited number of CFP time slots maximum 7 nodes can be allo-
cated GTS.
In this work, we propose an efficient superframe structure for 802.15.4
that is backward compatible with the original standard and compared it with
the current superframe structure in all three frequency bands. This new super-
frame structure significantly minimizes the delay and can assign GTS slots up
to 14 nodes without compromising the CFP duration. The proposed Super-
frame structure is suitable to support communication within the medium size
Personal Area Networks (PANs) [5]. The medium size PAN scenarios con-
tain more than 7 nodes in the PAN. Examples of PAN scenarios include home
automation, industrial monitoring, hospital storage or ward monitoring, struc-
tural monitoring, etc.
Major Contributions
The major contributions of our proposed MAC frame are as follows:
rAn efficient super-frame structure where CFP precedes the CAP along
with accommodating more number of GTS requesting nodes than the orig-
inal standard.
rDelay of GTS traffic is minimized due to this superframe structure.
rGTS utilization has also been improved by reducing the CFP slot size to
its half as of the original IEEE 802.15.4 standard.
rOur proposed superframe format is fully compatible with the IEEE
802.15.4 standard.
rFairly support communication within medium or large size PANs.
Rest of the paper is organized in the following manner: Section 2 high-
lights the previous work by different authors. Section 3 briefly discusses the
proposed superframe format along with the necessary modifications in the

154 AHMAD NASEEM ALVI et al.
Frequency Modulation Symbols / Bits / Symbol Data rate
Band (MHz) Scheme sec symbol Duration (sec) (bits/sec)
868 −868.6 BPSK 20000 1 50 ∗e−620000
902 −928 BPSK 40000 1 25 ∗e−640000
2400 −2483.5 O-QPSK 62500 4 16 ∗e−6250000
TABLE 1
Frequency Bands with Data Rates
Beacon frame fields. The numerical estimators for delay and link utilization
for the proposed superframe format are also presented in this section. Numer-
ical results of our proposed scheme are compared with the original IEEE
802.15.4 Beacon enabled mode in Section 4. In last, Section 5 concludes the
paper.
2 IEEE 802.14.5 STANDARD OVERVIEW AND RELATED WORK
2.1 IEEE 802.15.4 standard overview
Institute of Electrical and Electronic Engineering (IEEE) finalized 802.15.4
standard for low rate, low power and low cost, wireless personal area network
(Lo-WPAN). The standard operates in three different frequency bands such
as 868MHz, 915MHz and 2400MHz. 868Mhz and 915MHz use BPSK mod-
ulation scheme, however, 2400MHz uses O-QPSK modulation scheme with
data rates of 20Kbps, 40Kbps and 250Kbps respectively. These frequency
bands with their respective data rates are shown in Table 1. In each frequency
band, standard offers non-beacon enabled mode and beacon enabled mode.
During non-beacon enabled mode, nodes follow unslotted CSMA/CA
mechanism in order to access the channel, however superframe structure is
introduced in beacon enabled mode. Beacon enabled mode consists of active
period and an optional in-active period. Active period also known as Super-
frame Duration (SD) comprise of 16 equal sized time slots. SD initiates with
beacon frame generated by the coordinator followed by Contention Access
Period (CAP) and optional Contention Free Period (CFP). Out of these 16
time slots, minimum 9 time slots are reserved for beacon frame and CAP,
whereas, maximum 7 slots can be allocated for CFP.
Beacon frame is used for synchronization and possesses the information
of arrival of next beacon frame. Duration from start of current beacon frame
till the arrival of next beacon is known as Beacon Interval (BI’) and can be
calculated by equation 1:
BI=960 ×2BO (1)
where, BO ranges from 0 to 14.

NEW SUPERFRAME STRUCTURE WITH MINIMUM DELAY AND MAXIMUM CFP LINK 155
During CAP, all nodes contend to access the medium by following slotted
CSMA/CA mechanism in order to send their request to the coordinator. Dur-
ing CFP, TDMA based Guaranteed Time Slots (GTS) are allocated to nodes
for time sensitive data transmission. In this work, we focus on GTS allocation
mechanism.
In IEEE 802.15.4 standard, nodes requiring certain bandwidth in order to
transmit their data, send GTS request commands to coordinator during CAP.
Coordinator upon receiving all these requests decides about the allocation of
CFP slots to requesting nodes on first come first serve basis. If available CFP
slots remain less than the requesting slots by a node then that CFP slots are
not assigned to that node, contrary requested GTS are allocated to the request-
ing nodes. Coordinator take cares that allocated CFP slots would not reduce
the CAP length from aMinCAPLength value. Coordinator informs about suc-
cessful slots allocation in the GTS descriptor field available in next beacon
frame and nodes can send their data by using these allocated CFP slots in the
following superframe.
Delay of a node is calculated, when a node has a data to send till the time
when it successfully transmit its information to the coordinator. A noticeable
delay is observed, when a node generates GTS request till the allocation of
its CFP slots. This delay in most of the cases longer than the Beacon Inter-
val. This delay become more prominent when value of BO increases, which
is not affordable for time critical applications. During a PAN, maximum 7
CFP slots can be allocated for GTS allocation, which are insufficient as PAN
size increases. At the same time, these limited 7 slots are not efficiently uti-
lized due to varying data traffic. The slot utilization inefficiency rises with
the increase in slot size. These limitations in the existing standard regarding
GTS allocation have been improved in this work.
2.2 Related Work
The performance analysis of IEEE 802.15.4 standard in different prospects
has been carried out in many research studies. These research studies include
performance of CAP as well as CFP. Some interesting algorithms have also
been proposed by different researchers to improve the efficiency in terms of
power consumption, better link utilization and delay minimization. Valero et
al.. [6] propose an incrementally deployable energy efficient scheme based on
IEEE 802.15.4 standard for better energy conservation. In [7], Li et al. intro-
duced a novel approach by utilizing a synchronous low power listening tech-
nique in order to minimize power consumption. In [8], Kajima and Harada
addressed the power wastage issue in superframe structure due to periodic
transmission and introduce a turn off beacon employment. In it, authors opti-
mize the allocated active period by proposing a dynamic re-association pro-
cedure for energy conservation.

156 AHMAD NASEEM ALVI et al.
The standard is widely used in health care systems where Quality of Ser-
vice (QoS) is to minimize the delay for emergency messages. In [9]– [11],
authors address the delay minimization problems by proposing different ideas
during CAP, whereas improvement in GTS mechanism are proposed in [12]–
[16]. Kobayashi and Sugiyura [9] optimized traditional CSMA/CA mech-
anism and propose a timing group division method for faster communica-
tion by minimizing delays. In [10], the author exhibited QoS improvement in
IEEE 802.15.4 standard in terms of decrease in latency. These developments
are achieved by adopting an improved Binary Exponential Backoff algorithm,
which avoids collision. In [11], a backofff control mechanism is introduced
for cluster based WSN. Authors claim that, the scheme not only minimize
the delay but also improves the throughput of the system.
J. Chen et al. [12] introduced Explicit GTS Sharing and Allocation
Scheme (EGSA) for real time communication applications, where tighter
delay bounds are required. A multi-hop communication scheme in GTS
mechanism of IEEE 802.15.4 standard is proposed in [13], which follows
superframe structure and claim for decrease in delay and better packet deliv-
ery ratio as of the standard. An Unbalanced GTS Allocation Scheme (UGAS)
is proposed in [14]. In this scheme, Link utilization is increased by intro-
ducing different duration time slots for different bandwidth requirements.
Authors claim that UGAS improves the bandwidth utilization by 30% as of
the standard. Feng Xia et al. [15] proposed an Adaptive and Real-Time GTS
Allocation Scheme (ART-GAS) for such applications, where time sensitive
and high-traffic is required and compatible with IEEE 802.15.4 standard.
Authors claim that proposed scheme increases the bandwidth utilization as
of the standard. In [16], authors developed an admission control and schedul-
ing algorithm for body area network and claim for 100% compliance in time
constraints.
Many solutions have been proposed to efficiently allocate GTS slots to
the requesting nodes [17]– [20]. These schemes try to improve QoS in LR-
WPANs by minimizing delay, increasing throughput and allocating CFP slots
to more number of nodes than the predefined limit in the standard. However,
most of the previous work follow the same superframe structure of 802.15.4
standard and just try to increase or shrink the size of GTS area or slots to opti-
mize the GTS utilization. In result, those schemes fail to minimize the delay
of GTS traffic due to intrinsic gap of CAP between Beacon and the CFP. Sim-
ilarly, most of the schemes compromise the CFP duration to increase the GTS
efficiency. At the same time most of the work is analyzed on only 2400MHz
frequency bands and 868MHz and 915MHz frequency bands are ignored in
their work. However in this work, we compared our proposed superframe
structure with the superframe structure of the current standard in all three
frequency bands.

NEW SUPERFRAME STRUCTURE WITH MINIMUM DELAY AND MAXIMUM CFP LINK 157
Beacon
Conte ntion Acces s perio d
(CAP)
Beacon Interval = δ +(960 x 2BO)
Superframe Duration = Beacon + ( 960 x 2SO)
Superframe Order (SO) = 0 ≤ SO ≤ BO ≤ 14
Beacon Order(BO) = 0 ≤ BO ≤ 14, If BO=15, there will be no beacon.
Duty Cycle = (δ + (960 * 2SO)) / ( δ + (960 * 2BO)), wher e δ is B eacon D uration in Sym bols.
Inactive Period
Beacon
Active Period
0
01
1
23
2
45
3
67
4
89
5
1011
6
1213
(CFP)
δ
Proposed CFP Slots
Original CFP Slo ts
FIGURE 2
Proposed Superframe Format.
3 PROPOSED SUPERFRAME STRUCTURE FOR IEEE 802.15.4
In this section we discuss the novel MAC protocol for WSN. The superframe
format of the proposed MAC protocol is shown in Figure 2. It shows that
the CFP starts immediately after Beacon frame and followed by the CAP.
However, when there is no GTS request then CAP will commence right after
Beacon frame. Though maximum duration of CFP in superframe is same
as of IEEE 802.15.4 standard however maximum number of CFP slots have
been extended to 14 equal slots instead of 7 maximum slots used in IEEE
802.15.4 standard. Each of the CFP slots is exactly one half duration of the
CAP slot. The main advantage behind this scheme is to minimize unnecessary
delay faced by nodes that communicate during GTS due to presence of CAP
between Beacon and GTS. In addition, it further increases the efficiency in
terms of throughput. To achieve this superframe format and compatibility
with the existing 802.15.4 standard, we proposed some changes in parameters
of the existing standard.
In our proposed scheme, superframe slots are of two durations, slot dura-
tion in CAP is same as computed by SO in the existing standard and CFP slot
size is exactly one half of the CAP. The superframe contains minimum 9 CAP
slots and maximum 14 CFP slots excluding the Beacon frame. This exclu-
sion of Beacon frame will help without compromising the aminCAPlength
parameter as the minimum CAP length will never be less than 540 Symbols.
The Superframe Duration (SD) depends upon the value of Superframe Order
(SO), aNumSuperframeSlot (NSS) and aBaseSlotDuration (BSD) as:
SD =δ+(NSS×BSD ×2SO) (2)
where, NSS = 16, SO ranges between 0 and Beacon Order (BO) 0 ≤SO ≤
BO.BSD in above expression is computed as:
BSD =3×aUnitBackoffPeriod (3)

158 AHMAD NASEEM ALVI et al.
Default value of aUnitBackoffPeriod is 20 Symbols. The Beacon Duration in
symbols, δSTD, in Eq. (2) is computed in equations 4:
δSTD =(m+3∗n)×SST D(Symbols) (4)
where S868 =8, S915 =8, and S2400 =2, and nis the number of nodes that
have been granted GTS and mis the length of Beacon frame, refer Beacon
frame format in Figure 3(a), in bytes without GTS list field. The Beacon
Interval (BI) and Duty Cycle (DC) of the proposed scheme for its different
frequency bands are estimated as follows:
BI =δSTD +(NSS×BSD ×2BO) (5)
DC =SD/BI (6)
Time duration of the next Beacon arrival (Beaconstart) is calculated since
the expiry of current Beacon and is computed by knowing the value of BO
through following formula as:
Beaconstart =NSS×BSD ×2BO (7)
Nodes compute the beginning of CAP (CAP
Start ) by multiplying each CFP
slot duration with number of CFP slots (NCFP) mentioned in Final CAP Slot
of Superframe Specification Field by following formula as:
CAP
Start =15 ×2SO+1×NCFP (8)
To achieve the proposed superframe format, the Beacon frame format along
with the superframe specification and GTS fields has been modified and are
shown in Figure 3(a), 3(b) and 3(c), respectively.
Bits (b8to b11)inSuperframe specification field represent the Final CFP
Slot, which indicate the start of the CAP. However, in the original 802.15.4
standard these bits express the Final CAP Slot. Similarly in GTS Filed, we
extended the GTS Direction field to 2 bytes to augment 14 slots in CFP
period. If Nbps is the number of bits can be transmitted during each CFP
slot and Dis the data required to be sent, then each GTS requesting node
calculates the number of CFP slots NGT S in order to send Dbits of data in
proposed scenario by following formula as:
(NGTS)=D
Nbps (9)

NEW SUPERFRAME STRUCTURE WITH MINIMUM DELAY AND MAXIMUM CFP LINK 159
212 2/8 2
Variable 2
Contr ol
Frame
Source
PAN ID
Source
Address
Superfram e
Specs GTS Field
Beacon
Seq. No.
Pending
Address Field
Beacon
Payload FCS
GTS Spe cification GTS Li stGTS Direction
b
0
~ b
3
= GTS Descriptor Count
(Max. va lue is 14)
b
4
~ b
6
= Reserved
b
7
= GTS Permit
b
0
~ b
15
= Device short address
b
16
~ b
19
= GTS Starting Slot
b
20
~ b
23
= GTS Length
12 3
Beacon Order Superframe order Final CFP Slot
Battery life
Extension
Reserved
PAN
Coordinator
Association
Permit
b
14
= 1 Beacon transmitted by PAN Coordinator.
b
14
= 0 Beacon transmitted by Coordinator.
b
15
=1 PAN coordinator accepted association.
b
15
=0 PAN coordinator did not accept association.
b
8
- b
11
b
12
b
0
- b
3
b
4
- b
7
b
0
,~ b
13
= GTS Direction Mask
(0  Transmit | 1  Receive)
b
14
, b
15
= Reserved
b
13
b
14
b
15
Variable Variable
GTS Field
Superframe Specifications Field
Beacon Frame
Byte(s)
Byte(s)
(a)
(b)
(c)
FIGURE 3
Proposed Superframe Format.
Where, Nbps for 868MHz and 915MHz is 15 ×2SO+1and for 2400MHz is
15 ×2SO+3.
3.1 Delay Calculation
Transmission delay of a node is calculated when a node has data to send till
the end of CFP slot where it successfully sends its data to PAN coordinator. A
node ihas data Dijust at the beginning of the Beacon frame, its time required
to complete its transmission by the end of the allocated GTS is calculated as:
Di=BI +(b=i

b=1
Kb×ts) (10)
Here,
Di= Time required to send data by node i,
Kb= Number of slots allocated to node iand its preceding nodes
ts= Duration in seconds of each CFP slot, which is a multiple of number
of symbols per slot in proposed model (Nspsp) and time required in seconds
for each symbol (tes) is calculated as:
ts=Nspsp ×tes (11)
Here,
Nspsp =15 ×2SO+1(868MHz,915MHz and 2400MHz)
tes =50 ×e−6(868MHz)
tes =25 ×e−6(915MHz)
tes =16 ×e−6(2400MHz)

160 AHMAD NASEEM ALVI et al.
If pnodes have been successfully assigned GTS, then total delay of the
network is calculated as:
Dpmax =
i=p

i=1
[BIi+(b=i

b=1
Kb×ts)] (12)
However, in the current IEEE802.15.4 standard, Delay of a node iis calcu-
lated as:
Di=BI+SD−(b=i

b=1
Kb−1×to) (13)
Here,
BI=960 ×2BO
SD=960 ×2SO
to=Duration in seconds of each CFP slot, which is a multiple of number
of symbols per slot in current standard (Nspso) and time required in seconds
for each symbol (tes) as shown in equation 14:
to=Nspso ×tes (14)
where, Nspso for all frequency bands is 15×2SO+2.
If pnodes have been successfully assigned GTS, then total delay of the
network in current standard is calculated as:
Domax =
i=p

i=1
[(BI+SD)i−
b=i

b=1
Kb−1×ts] (15)
3.2 Link Utilization
It has been observed that significant amount of bandwidth is wasted during
CFP. This wastage becomes more significant as CFP slot size increases. By
increasing CFP slots will help in accommodate data in an efficient manner.
If Didata is required to be sent by node i then time required to transmit this
data tdto PAN coordinator is estimated as:
ti=Di
C(16)
Here Cis the data rate through which node communicate. If Kiis the number
of CFP slots required to send Didata, then it is computed as:
Ki=Di
Nbps (17)

NEW SUPERFRAME STRUCTURE WITH MINIMUM DELAY AND MAXIMUM CFP LINK 161
If a node irequires Kislots in transmitting its data Dito PAN-coordinator
then link utilization (Ui) for node iis calculated as:
Ui=ti
Ki×ts(18)
If pnodes are successfully allocated CFP slots, then the link utilization,
UCFP, for pnodes is computed as:
UCFP =
p

i=1
ti
Ki×ts(19)
However, link utilization of same node ifor current standard, Uoi, is calcu-
lated as:
Uoi=ti
Ko×to(20)
here kois number of CFP slots required to send data during CFP in current
standard. If qnodes have been successfully assigned CFP slots, then link
utilization UCFP
oduring a specific BI is calculated as:
UCFP
o=
q

i=1
ti
Ko×to(21)
4 NUMERICAL ANALYSIS AND DISCUSSION
In order to evaluate performance of our proposed scheme with the existing
standard, we used the evaluation environment as in [21]. In evaluation, we
consider PAN scenario where multiple nodes with varying data traffic from
10 to 180 Bytes is considered. The effects of different values of BO and
SO parameters on performance of CFP are analyzed. For better comparison
of our proposed scheme with the original standard, we have analyzed our
proposed superframe structure in different prospects including delay calcula-
tions, link utilization and slot(s) allocation to GTS requesting nodes. Rest of
the parameters are summarized in Table 2.
4.1 Delay Analysis
Figures 4,5,6 show the average delay for 868MHz, 915MHz and 2400Mhz
frequency bands, respectively. The delay is evaluated for varying BO against

162 AHMAD NASEEM ALVI et al.
Parameter Value
Network Area 100m x 100m
PAN coordinator’s position (0, 0)
Number of nodes in a PAN 14
Data requested by each node 10 .. 180 (bytes)
Frequency Bands 868, 915 and 2400 MHz
BO 0 .. 10
SO 0 le SO le BO
868MHz = 50μs
Transmission Time/Bit 915MHz = 25μs
2400MHz = 4μs
868MHz = 48 ∗2SO ms
Superframe Duration 915MHz = 24 ∗2SO ms
2400MHz = 3.84 ∗2SO ms
868MHz = 3 ∗2SO ms
GTS Duration 915MHz = 1.5∗2SO ms
2400MHz = 0.24 ∗2SO ms
TABLE 2
Numerical Evaluation Parameters
0 1 2 3 4 5 6 7 8 9 10
0
10
20
30
40
50
60
Beacon Order (BO)
Avg. Delay (s)
Orig. Standard
Proposed
1 2 3 4
0.2
0.4
0.6
0.8
1
1.2
FIGURE 4
Average delay ηBO vs BO where 0 ≤BO ≤10 and 0 ≤SO ≤BO in 868MHz.
0 1 2 3 4 5 6 7 8 9 10
0
5
10
15
20
25
30
Beacon Order (BO)
Avg. Delay (s)
Orig. Standard
Proposed
1 2 3 4
0.2
0.4
0.6
Orig. Standard
Proposed
FIGURE 5
Average delay ηBO vs BO where 0 ≤BO ≤10 and 0 ≤SO ≤BO in 915MHz

NEW SUPERFRAME STRUCTURE WITH MINIMUM DELAY AND MAXIMUM CFP LINK 163
0 1 2 3 4 5 6 7 8 9 10
0
2
4
6
8
10
12
14
16
18
20
Beacon Order (BO)
Avg. Delay (s)
Orig. Standard
Proposed
1234
0.1
0.2
0.3
FIGURE 6
Average delay ηBO vs BO where 0 ≤BO ≤10 and 0 ≤SO ≤BO in 2400MHz
all possible values of SO, as follows:
ηBO =BO
SO=0DpmaxSO
BO where BO =0..10 (22)
The graphs show that average delay of IEEE 802.15.4 and the proposed
Superframe structure rises with increase in BO. It is also observed from
the results that the average delay increases gradually as frequency band
decreases, however ratio of the delay between standard and proposed
schemes for all BO values remain almost similar.
The average delay is computed for the data packet of 125 bytes. Result
shows that average delay of the network in proposed scheme is exceptionally
better than the original standard. This improvement is only because node(s)
do(es) not wait for the whole CAP period after successful assignment of the
GTS in our proposed scheme. The detailed results of delay for fixed BO and
varying SO are also analyzed. Figures 7, 8 and 9 show the delay for 868MHz,
915MHz and 2400MHz frequency bands. Each figure comprises of four parts
(a),(b),(c) and (d) that depict the delay for varying SO and the fixed value of
BO that is 10, 8, 6, and 4, respectively. It is obvious from this detailed delay
analysis that our scheme has comparatively very less delay than the original
standard in all three frequency bands.
Figures 10, 11 and 12 represent delay difference between average delay
experienced by proposed and standard Superframe structures for varying val-
ues of BO, ranges from 0 to 10. This average delay is calculated by sum-
ming up the delay for each SO value and then dividing it by the number
of SO in that range. Difference between the delay experienced by the origi-
nal standard and the proposed scheme is clearly evident from these figures.

164 AHMAD NASEEM ALVI et al.
0 1 2 3 4 5 6 7 8 9 10
40
60
80
100
Superframe Order (SO)
Delay, (s)
Orig. Standard, BO=10
Proposed, BO=10
0 1 2 3 4 5 6 7 8
10
15
20
25
Superframe Order (SO)
Delay, (s)
Orig. Standard, BO=8
Proposed, BO=8
0 1 2 3 4 5 6
3
4
5
6
7
Su
p
erframe Order (SO)
Delay, (s)
Orig. Standard, BO=6
Proposed, BO=6
0 1 2 3 4
0.8
1
1.2
1.4
1.6
Su
p
erframe Order
(
SO
)
Delay, (s)
Orig. Standard, BO=4
Proposed, BO=4
(a)
(c) (d)
(b)
FIGURE 7
Average delay vs varying SO in 868MHz
0 1 2 3 4 5 6 7 8 9 10
20
30
40
50
Superframe Order (SO)
Delay, (s)
Orig. Standard, BO=10
Proposed, BO=10
0 1 2 3 4 5 6 7 8
6
8
10
12
14
Superframe Order (SO)
Delay, (s)
Orig. Standard, BO=8
Proposed, BO=8
0 1 2 3 4 5 6
1.5
2
2.5
3
3.5
Su
p
erframe Order
(
SO
)
Delay, (s)
Orig. Standard, BO=6
Proposed, BO=6
0 1 2 3 4
0.4
0.5
0.6
0.7
0.8
Su
p
erframe Order
(
SO
)
Delay, (s)
Orig. Standard, BO=4
Proposed, BO=4
(a)
(c)
(b)
(d)
FIGURE 8
Average delay vs varying SO in 915MHz
0 1 2 3 4 5 6 7 8 9 10
15
20
25
30
35
Superframe Order (SO)
Delay, (s)
Orig. Standard, BO=10
Proposed, BO=10
0 1 2 3 4 5 6 7 8
4
5
6
7
8
Superframe Order (SO)
Delay, (s)
Orig. Standard, BO=8
Proposed, BO=8
0 1 2 3 4 5 6
1
1.2
1.4
1.6
1.8
2
Superframe Order (SO)
Delay, (s)
Orig. Standard, BO=6
Proposed, BO=6
0 1 2 3 4
0.25
0.3
0.35
0.4
0.45
0.5
Superframe Order (SO)
Delay, (s)
Orig. Standard, BO=4
Proposed, BO=4
(a)
(c) (d)
(b)
FIGURE 9
Average delay vs varying SO in 2400MHz

NEW SUPERFRAME STRUCTURE WITH MINIMUM DELAY AND MAXIMUM CFP LINK 165
012345678910
0
1
2
3
4
5
6
7
8
9
10
0
10
20
30
40
50
Delay Dierence, Δ, (s)
Beacon Order (BO) Superframe Order (SO)
FIGURE 10
Difference in average delay for varying BO in 868MHz
012345678910
0
1
2
3
4
5
6
7
8
9
10
0
10
20
30
Delay Dierence, Δ, (s)
Beacon Order (BO) Superframe Order (SO)
FIGURE 11
Difference in average delay for varying BO in 915MHz
012345678910
0
1
2
3
4
5
6
7
8
9
10
0
5
10
15
20
Delay Dierence, Δ, (s)
Beacon Order (BO) Superframe Order (SO)
FIGURE 12
Difference in average delay for varying BO in 2400MHz

166 AHMAD NASEEM ALVI et al.
50 100 150
0
20
40
60
80
100
GTS Utilization (%)
Orig. Standard, SO=0
Proposed, SO=0
50 100 150
0
20
40
60
80
100
Orig. Standard, SO=2
Proposed, SO=2
50 100 150
0
20
40
60
80
100
Orig. Standard, SO=4
Proposed, SO=4
Data Requested by Each Node (Bytes)
FIGURE 13
Link Utilization of nodes vs varying data request in 868MHz
It has also been observed from the results that these differences in delay
increase as BO range increases. This is due to the fact that larger value of
BO causes increase in slot duration and consequently node/s has/have to wait
for longer duration during CAP in the current standard, whereas the same is
avoided in the proposed scheme by moving GTS immediately after beacon
frame.
4.2 Link Utilization
Link utilization is calculated by finding the total usage of available GTS space
against total allocated GTS in the superframe. Figures 13, 14 and 15 show
the link utilization against varying data traffic for different values of SO for
868MHz, 915Mhz and 2400MHz frequency bands respectively. There are
14 nodes and their request data size varies from 10 to 180 bytes per node.
The results show that link utilization improves in the proposed standard
for different values of SO. This increase in link utilization in the proposed
scheme is due to the allocation of smaller time slots of CFP to each GTS
50 100 150
0
20
40
60
80
100
GTS Utilization (%)
Orig. Standard, SO=0
Proposed, SO=0
50 100 150
0
20
40
60
80
100
Orig. Standard, SO=2
Proposed, SO=2
50 100 150
0
20
40
60
80
100
Orig. Standard, SO=4
Proposed, SO=4
Data Requested by Each Node (Bytes)
FIGURE 14
Link Utilization of nodes vs varying data request in 915MHz

NEW SUPERFRAME STRUCTURE WITH MINIMUM DELAY AND MAXIMUM CFP LINK 167
50 100 150
10
20
30
40
50
60
70
80
90
100
GTS Utilization (%)
Orig. Standard, SO=0
Proposed, SO=0
50 100 150
0
20
40
60
80
100
Orig. Standard, SO=2
Proposed, SO=2
50 100 150
0
20
40
60
80
100
Orig. Standard, SO=4
Proposed, SO=4
Data Requested by Each Node (Bytes)
FIGURE 15
Link utilization of nodes vs varying data requests in 2400MHz
requesting node, which avoids link wastage. It has been noticed that link uti-
lization increases at an average of 6.9%, 6.4% and 29.3% for both 868MHz
and 915MHz frequency bands and 6.5%, 29.3% and 100% for 2400MHz fre-
quency band for SO of 0,2 and 4 respectively.
4.3 GTS allocation nodes
It has been observed that increase in SO value causes longer slot duration and
each slot that can accommodate large data traffic. In our proposed model, 14
CFP slots of shorter duration can manage sufficient amount of data and in
response can entertain maximum number of nodes. Figures 16, 17 and 18
show assignment of GTS to nodes during 868MHz, 915Mhz and 2400MHz
frequency bands. It is evident from the results that proposed model can assign
GTSs to large number of nodes as of the original standard. As SO increases,
each node is assigned one slot and hence maximum of 14 nodes can be enter-
tained as of the original standard where only 7 nodes can be entertained for
the same data traffic for each node. Similarly, for the smaller values of SO
20 40 60 80 100 120 140 160 180
0
5
10
15
Data Requested by each Node (Bytes)
No. of Nodes Assigned GTS Slot(s)
Orig. Standard, SO=0
Orig. Standard, SO=2
Orig. Standard, SO=4
Proposed, SO=0
Proposed, SO=2
Proposed, SO=4
FIGURE 16
No. of GTS assigned vs varying data request in 868MHz

168 AHMAD NASEEM ALVI et al.
20 40 60 80 100 120 140 160 180
0
5
10
15
No. of Nodes Assigned GTS Slot(s)
Data Requested by each Node (Bytes)
Orig. Standard, SO=0
Orig. Standard, SO=2
Orig. Standard, SO=4
Proposed, SO=0
Proposed, SO=2
Proposed, SO=4
FIGURE 17
No. of GTS assigned vs varying data request in 915MHz
20 40 60 80 100 120 140 160 180
0
5
10
15
Data Requested by each Node
No. of Nodes Assigned GTS Slot(s)
Orig. Standard, SO=0
Orig. Standard, SO=2
Orig. Standard, SO=4
Proposed, SO=0
Proposed, SO=2
Proposed, SO=4
FIGURE 18
No. of GTS assigned vs varying data request in 2400MHz
and large volume of data requests from the nodes, our proposed scheme enter-
tains equal or more number of nodes compared to the original standard in all
frequency bands.
It is evident from the simulation results that for any scenario, the average
delay of the proposed scheme is less than the original standard for all fre-
quency bands. However, following are the scenarios where proposed scheme
fully utilizes the CFP link duration. For 868/915MHz, the CFP link will be
fully utilized by the proposed scheme if the data requested by each node is
equal to eq.(23). Similarly, the 100% CFP link is utilized for the 2.4GHz
band when every node requests the data (in bits) as in eq. (24). It is also evi-
dent from Figure 19 that proposed scheme utilizes the 100% of the CFP for
the given scenario. However, the original standard fails to achieve the same,

NEW SUPERFRAME STRUCTURE WITH MINIMUM DELAY AND MAXIMUM CFP LINK 169
200 400 600 800 1000 1200 1400 1600
50
60
70
80
90
100
GTS Link Utilization (%)
0 1000 2000 3000 4000 5000 6000 7000
50
60
70
80
90
100
Proposed, SO=2
Orig. Standard, SO−2
Proposed, SO=2
Orig. Standard, SO=2
IEEE 802.15.4
868 MHz and 915 MHz
IEEE 802.15.4
2.4 GHz
Data Requested (Bytes)
FIGURE 19
CFP or GTS link utilization versus data requested by nodes.
which validates the claim of the proposed scheme.
No. of Slots in CFP

i=0
30 ×2SO ×iwhere 0 ≤SO ≤BO (23)
No. of Slots in CFP

i=0
120 ×2SO ×iwhere 0 ≤SO ≤BO (24)
5 CONCLUSION
In this paper we proposed an efficient superframe structure for IEEE802.15.4
standard. The Beacon frame format along with all suitable parameters is also
proposed to prove that the our superframe structure is backward compati-
ble with the current standard with very minor modifications. The analytical
results show that this superframe format improves delay, accommodates more
number of nodes and better utilizes the CFP slots compared to the original
802.15.4 standard in all three frequency bands.
ACKNOWLEDGMENTS
This research was supported by the MSIP (Ministry of Science, ICT and
Future Planning), Korea, under the ITRC (Information Technology Research
Center) (IITP-2016-H8601-16-1002) supervised by the IITP (Institute for
Information & communications Technology Promotion).

170 AHMAD NASEEM ALVI et al.
REFERENCES
[1] Ahmed, S.H., Kim, G., and Kim, D. (2013). “Cyber-Physical System: Architecture, Appli-
cations and Research Challenges,” IFIP Wireless Days, Valencia, Spain, Nov.
[2] Moravejosharieh, A., and Lloret, J. (May 2016). “A Survey of IEEE 802.15.4 Effective
System Parameters for Wireless Body Sensor Networks,” International Journal of Com-
munication Systems, 29(7), pp. 1269–1292.
[3] Zheng, J., and Lee, M.J. (June 2004). “Will IEEE 802.15.4 Make Ubiquitous Network-
ing a Reality?: A Discussion on a Potential Low Power, Low Bit Rate Standard,” IEEE
Communications Magazine, 42(6), pp. 140–146.
[4] Pollin, S., Ergen, M., Ergen, S., Bougard, B., Der Perre, L., Moerman, I., Bahai, A.,
Varaiya, P., and Catthoor, F. (September 2008). “Performance analysis of slotted carrier
sense ieee 802.15.4 medium access layer,” IEEE Transactions on In Wireless Communica-
tions, 7(9), pp. 3359–3371.
[5] Alvi, A.N., Bouk, S.H., Ahmed, S.H., and Yaqub, M.A. (2016). “Influence of Backoff
Period in slotted CSMA/CA of IEEE 802.15.4,” IFIP International Conference on Wired
and Wireless Internet Communications (WWIC), Greece, pp 40–51.
[6] Valero, M., Bourgeois, A., and Beyah, R. (May 2010). “DEEP: A Deployable Energy Effi-
cient 802.15.4 MAC Protocol for Sensor Networks,” IEEE International Conference on
Communications (ICC), 2010, pp. 1–6.
[7] Xiaoyun, L., and Bleakley, C.J. and Wojciech, B. (2012). “Enhanced Beacon-Enabled
Mode for improved IEEE 802.15.4 low data rate performance,” Journal of Wireless Net-
works, Springer US, 18(1), pp. 59–74.
[8] Kojima, F., and Harada, H. (May 2010). “Long-Lived Smart Utility Network Management
Using Modified IEEE 802.15.4 MAC,” IEEE International Conference on Communica-
tions Workshops (ICC), 1(5), pp. 23–27.
[9] Alvi, A.N., Bouk, S.H., Ahmed, S.H., Yaqub, M.A., Javaid, N., and Kim, D. (June 2015).
“Enhanced TDMA based MAC protocol for adaptive data control in wireless sensor net-
works,” in Journal of Communications and Networks, 17(3), pp. 247–255.
[10] Hiep, P.T. (2016). “Spatial Reuse Superframe for High Throughput Cluster-Based WBAN
with CSMA/CA,” Ad Hoc and Sensor Wireless Networks, 31(1-4), pp. 69–87.
[11] Mori, K., Naito, K., and Kobayashi, H. (May 2011). “Distributed backoff mechanism for
traffic adaptive active period control in cluster-based IEEE 802.15.4 WSNs”, IEEE 73rd
Vehicular Technology Conference (VTC Spring), pp. 1–5.
[12] Chen, J., Ferreira, L.L., and Tovar, E. (Sept. 2011). “An explicit GTS allocation algo-
rithm for IEEE 802.15.4,” IEEE 16th Conference on Emerging Technologies and Factory
Automation (ETFA), pp. 1–8.
[13] Choi, W.C., and Lee, S.M. (2012). “Novel GTS Mechanism for Reliable Multihop Trans-
mission in the IEEE 802.15.4 Network,” International Journal of Distributed Sensor Net-
works, 2012, Article ID 796426, 10 pages.
[14] Le, N.T., Choi, S., and Jang, Y.M. (Nov. 2012). “A New QoS Resource Allocation Scheme
Using GTS for WPANs,” Journal of Wireless Personal Communication, 67(1), pp. 25–45.
[15] Xia, F., Hao, R., Li, J., Xiong, N., Yang, L.T., and Zhang, Y. (Nov. 2013). “Adaptive GTS
allocation in IEEE 802.15.4 for real-time wireless sensor networks,” Journal of Systems
Architecture, 59(10), Part D, pp. 1231–1242.
[16] Moravejosharieh, A., and Lloret, J. (October 2016). “Performance Evaluation of
Co-located IEEE 802.15.4-based Wireless Body Sensor, Networks,” Annals of
Telecommunications-Annales des Telecommunications, 71(9), pp 425–440.

NEW SUPERFRAME STRUCTURE WITH MINIMUM DELAY AND MAXIMUM CFP LINK 171
[17] Huang, P., Xiao, L., Soltani, S., Mutka, M.W., and Xi, N. (First Quarter 2013). “The Evo-
lution of MAC Protocols in Wireless Sensor Networks: A Survey,” IEEE Communications
Surveys & Tutorials, 15(1), pp. 101–120.
[18] Rao, S., Keshri, S., Gangwar, D., Sundar, P., and Geetha, V. (April 2013). “A survey and
comparison of GTS allocation and scheduling algorithms in IEEE 802.15.4 wireless sen-
sor networks,” In Proceedings of the IEEE Conference on Information & Communication
Technologies (ICT) 2013, pp. 98, 103.
[19] Huang, Y.-K., Pang, A.-C., and Hung, H.-N. (May 2008). “An Adaptive GTS Alloca-
tion Scheme for IEEE 802.15.4,” IEEE Transactions on Parallel and Distributed Systems,
19(5), pp. 641–651.
[20] Yang, L., and Zeng, S. (Oct. 2012). “A new GTS allocation schemes for IEEE 802.15.4,”
In Proceedings of the 5th International Conference on Biomedical Engineering and Infor-
matics (BMEI), 2012, pp. 1398–1401.
[21] Ranjit, J.S., and Shin, S. (2013). “A Modified IEEE 802.15.4 Superframe Structure for
Guaranteed Emergency Handling in Wireless Body Area Network,” Network Protocols
and Algorithms, 5(2), pp. 1–15.