See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/235676934

Experimental Assessment of Wireless Coexistence for 802.15.4 in the Presence

of 802.11g/n

Conference Paper · August 2012

DOI: 10.1109/ISEMC.2012.6351685

CITATIONS READS
32 362

3 authors:

Nickolas LaSorte Samer Rajab

NTIA Honda R&D Americas Inc.
28 PUBLICATIONS 231 CITATIONS 36 PUBLICATIONS 293 CITATIONS

SEE PROFILE SEE PROFILE

Hazem Refai
University of Oklahoma
267 PUBLICATIONS 3,405 CITATIONS

SEE PROFILE

All content following this page was uploaded by Nickolas LaSorte on 27 May 2014.

The user has requested enhancement of the downloaded file.

 Experimental Assessment of Wireless Coexistence
for 802.15.4 in the Presence of 802.11g/n
Nickolas J. LaSorte, Samer A. Rajab, Hazem H. Refai
Electrical and Computer Engineering, University of Oklahoma
Tulsa, OK, USA
[email protected]
[email protected]
[email protected]

Abstract— Wireless coexistence is a growing concern, given evaluating interference among wireless services.
the ubiquity of wireless technology. Although IEEE Standards In the literature, a notable problem exists: wireless networks
have started to address this problem in an analytical framework, are purported to emulate real-world environments; however,
a standard experimental setup and process to evaluate wireless
no standard methodology is utilized [24, 30, 32, 33, 34, 45,
coexistence is lacking. Literature that reports experimental
assessment of wireless coexistence places little emphasis on 46]. Line-of-sight (LOS) is most often used, and wireless
separation distance of wireless nodes under test or the spectrum transmission characteristics and separation distances among
occupancy of the interfering network, making comparisons wireless nodes are dissimilar. For these reasons, it is difficult
difficult. This paper provides an extensive literature survey of to compare results.
802.15.4 and 802.11 b/g/n wireless coexistence and demonstrates
that in a higher wireless channel occupancy environment, ZigBee A. Contribution
coexists with 802.11n better than with 802.11g. A reproducible, Coexistence among wireless devices is dependent on three
versatile, and practical test setup is presented to serve as a main factors: 1) frequency, 2) space, and 3) time. The key to
starting point toward establishing standard practice for wireless
achieving coexistence lies in the ability to control at least one
coexistence testing of wireless systems in general and wireless
medical devices in particular. Experimental evaluations of the three aforementioned factors. Coexistence is possible
demonstrated consistency with results reported in the literature. given one of the three following conditions: 1) Adequate
frequency separation between wireless networks; 2) Sufficient
I. INTRODUCTION distance between wireless networks, effectively decreasing the
Wireless coexistence of communication services is a signal-to-interference ratio (SIR) in each; and/or 3) Relatively
growing concern. The early 802.16.2-2001 standard [1] low overall occupancy of the wireless channel. Taking these
specified guidelines and deployment practices for minimizing three factors into consideration, we have developed an
interference among fixed broadband wireless access systems experimental coexistence test protocol that is practical,
and covered frequencies from 10-66 GHz. The standard was versatile, and reproducible.
later superseded by IEEE 802.16.2-2004 [2] and then IEEE The main contribution of this paper is the validation of a
802.15.2-2003 [3], which recommended coexistence practices reproducible non-light-of-sight (NLOS) setup to test
among personal area networks and other selected wireless coexistence of 802.15.4 alongside 802.11g and 802.11n. We
devices operating in unlicensed frequency bands. IEEE demonstrate that in a higher wireless channel occupancy
802.15.4-2003[4] followed and listed suggested factors for environment, ZigBee coexistence with 802.11n is superior to
ZigBee coexistence that can be found in Annex E. IEEE 802.11g. A discussion of coexistence factors will be illustrated
1900.2-2008 [5] recommended interference analysis for utilizing IEEE 802.15.4 (ZigBee) and 802.11b/g/n. Relevant
wireless systems and offered an exhaustive list of coexistence literature discussed in this paper is limited to these standards.
factors in both the physical and medium access control layers. Experimental results in the NLOS test setup are shown
The standard also suggested a structure for a coexistence consistent with trends found in the literature. The NLOS test
report. setup provides a practical, versatile, and reproducible test
The IEEE 802.19 Wireless Coexistence Working Group [6] method as a starting point towards determining a coexistence
is a technical advisory group created in the wake of 802.15.2 standard.
success. Members act as a coexistence advisory committee for
all IEEE 802 standards. The primary focus, however, is on B. Literature Review
IEEE 802 standards operating in unlicensed bands. The group Early efforts to study wireless interference in the 2.4 GHz
is currently concentrating on practice methods to assess band commenced in the latter part of the 1990s. In 1997
wireless network coexistence. Although each standard [1-6] Kamerman and Erkocevic [7] investigated interference caused
adequately outlines analytical guidelines for determining by microwave ovens operating in the vicinity of a wireless
coexistence, there is no widely used experimental protocol for local area network (WLAN) and then presented requirements
on the signal-to-noise ratio (SNR). In the following two-year

978-1-4673-2060-3/12/$31.00 ©2012 IEEE 473

period, a handful of papers addressed possible interference
between 802.11 and Bluetooth operating in the 2.4 GHz
Industrial, Scientific and Medical (ISM) band [8-16]. The SIR
threshold of an 802.11b receiver decoding a signal was found 802.11g/n

empirically [11, 13], and this provided a defined interference

factor.
In 2001, several papers reported analytical and experimental 4m
ZigBee
evaluation of the interference between 802.11b and Bluetooth. Node

Soltanian et al. [17] and Howitt et al. [18] were the first
1m
among them to mention that the effect of adjacent channel
interference on bit-error rate strongly depends on the
802.11g/n
frequency offset between useful and interfering carriers.
Howitt [19] studied Bluetooth network performance in the
11 m
presence of an interfering 802.11b network in 2002. His
methodology consisted of a three-step process: 1) characterize
802.11b interference under static condition; 2) characterize
Bluetooth network performance when collocated with a single
802.11b source; and 3) characterize Bluetooth network
performance when operating in an arbitrary 802.11b
environment. In 2003, IEEE published a recommended ZigBee
practice [3] reporting the development of simulation and Node

analytical models addressing coexistence problems between

802.15.4 and 802.11b wireless networks. Later that year, Fig. 1. NLOS Test Setup
Howitt el al. [20] was one of the first papers to report the TABLE I
impact of an 802.15.4 network on an 802.11b network. INITIAL AND MINIMUM DISTANCES REQUIRED BY ANSI C63-18
In the years that followed, wireless coexistence literature
Transmit Power INITIAL DISTANCE Minimum Distance
focusing on 802.15.4 can be categorized into two types—
those looking at the physical layer and those looking at the <600 mW 1m 0.25 m
medium access control (MAC) layer. Literature pertaining to 600 mW – 2 W 2m 0.5 m
2W–8W 3m 1m
the physical layer can be further broken down into three
sections: frequency [21-31], distance [24-43], and time [21, 22,
32-38, 44-46]. Literature regarding the MAC layer is signal and interference power [49]. Instead, SIR perceived by
primarily concerned with the clear channel assessment the receiver is a more robust metric for determining packet
threshold (CCA) [21, 27, 40, 47, 48] and the size of 802.15.4 decoding success rate.
packets [33-36]. IEEE 802.15.4 [50] specifies that a compliant device shall
The remainder of this paper is organized as follows. be capable of achieving a sensitivity of -85 dBm or better.
Section II presents an NLOS test protocol and then describes For the test reported in this paper, wireless medical devices
the equipment used during testing. Section III presents the were separated so that RSS at the receiver was at a minimum
experimental results. and packet error rate (PER) was maintained at 0%. The
receiver was placed on the outer edge of the transmitter’s cell
II. EXPERIMENTAL SETUP area, creating a worst-case scenario. To achieve a symbol
A. Non-Line-of-Sight Test Setup error rate of 10e-5 for quadrature phase shift keying (QPSK)
modulation, a 13.5 dB SNR must be maintained [51]. Taking
NLOS testing was performed without an anechoic chamber
these requirements into account and compensating for the
and could, therefore, be considered real world. During the
compliance testing phase for medical devices, it is possible transmitter’s power amplifier variation, the RSS measured at
that the transmit power of the wireless medical device might the medical device receiver is suggested to be -70 dBm.
not be adjustable. To simulate changes in transmit power, Coding gains and forward error corrections (FEC) are ignored
adjustment should be made to the path loss of the wireless for the ZigBee wireless devices in the development boards
signal. The test setup illustrated in Figure 1 incorporates SIR that were used (see Section IIB). However, it should be noted
and was utilized to control the received signal strength (RSS) that different coding gains and FEC in wireless transceivers
for a medical device under test. will alter PER. RSS standard deviation was 0.4493, and
An ambient scan was first conducted at each point where a median absolute deviation was 0.2640. Toscano [49] showed
wireless medical device under test was placed to identify that RSS for IEEE 802.15.4 is directly related to distance and
background noise in the 2.4 GHz ISM band. Previous testing is also stable, e.g., variation coefficient was below 2% for
showed that RSS is a poor indicator of link quality in noisy nearly all measurements. ZigBee wireless network
environments, as the receiver cannot distinguish between performance was tested to validate 0% PER.

474
 Once parameters were established, ZigBee nodes were neglected [53-55].
subjected to an interfering network (802.11g). Although the For ZigBee transmissions, the wireless network is capable
interfering network was limited to only two terminals, the of working in a reverberation chamber and is only seriously
configuration is considered general. Because the 802.11g limited for a value of Q-factor above 5000—a figure greater
network utilizes a carrier sense multiple access with collision than one typically found outside a reverberation chamber [57].
avoidance (CSMA/CA) mechanism, only one wireless node ZigBee PER is below 1% with a Q factor of 1000. Johnson et
can transmit at any given time for a given channel. The al. [58] measured Q in harsh military-related environments
(Zigbee) wireless medical device receiver and transmitter and found no Q factor greater than 1000. Theoretically, a 1000
were evaluated separately—each exposed to one or multiple Q-factor should not be present in any practical environment
interfering wireless networks. Various interference wherein a wireless medical device operates. Thus, in the
phenomena occurred, depending on whether or not the NLOS ZigBee network test setup the delay spread of the
interfering wireless network was in proximity to the ZigBee wireless network will not cause inter-symbol
transmitter or the receiver of the wireless network under test. interference and can be neglected as a contributor to bit error.
A receiver surrounded by one or more interfering networks To increase reproducibility, wireless channel delay spread can
experienced increased packet collision at the receiver, i.e., the be measured with a vector network analyzer at each lab
hidden terminal effect. In contrast, when a transmitter was performing wireless coexistence testing. Future work aims at
surrounded by one or more interfering networks, channel normalizing test results between different test setup
utilization decreased, i.e., the exposed terminal effect. configurations.
All wireless nodes were positioned on wooden tables at a Dimensions of the test setup are arbitrary. Emphasis is
height of 1 m from the ground. The separation distance placed on the average RSS at the wireless medical device
between the interfering wireless network and the wireless under test and also on isolating the effects of the interfering
medical device under test was also 1 m—a distance suggested wireless network on the transmitter and receiver. To do this,
by ANSI C63-18 [52]. Initial and minimum distances, shown LOS between the interfering wireless network and the
in Table 1, are based on the transmitting power of the wireless medical device under test is required. NLOS is
interfering wireless network. The measured power was required between wireless medical devices, as well as between
-60 dBm of the interfering network (802.11g) at the point of the interfering wireless network and the wireless medical
the wireless medical device under test. It should be noted that device furthest away.
the initial distance between the wireless medical device under
test and the interfering network remained at 1 m; experiments B. Wireless Network Equipment
testing different separation distances were not conducted. 1) 802.15.4 Wireless Network
Power-auto-leveling algorithms were disabled in all wireless Wireless medical devices were simulated using 802.15.4
networks during testing, and the interfering wireless networks (ZigBee) development test boards (CC2530) manufactured by
were set to their maximum transmission powers, creating a Texas Instruments. To generate ZigBee packets, TI SmartRF
worst-case scenario. Studio 7 for ZigBee development boards were used. ZigBee
Testing NLOS configurations outside an anechoic chamber wireless devices were set to channel 22 (2.460 GHz) for all
raises the possibility of collateral phenomena, including tests, and clear channel assessment (CCA) threshold was set to
reflections from nearby structures that cause multipath. To a maximum of -77 dBm. The literature shows that as the CCA
account for this, multiple test arrangements of the transmitter threshold increases, PER decreases [21, 27, 40, 47, 48].
and receiver terminals can be considered. It is suggested that Packet size was set to the development board minimum—a
the wireless device under test be placed in the middle of the 15-byte length. The literature notes that as the packet size of
room to allow for extra separation distance between the device the wireless network under test decreases, the probability of
and the interfering network. interference also decreases [33-36]. ZigBee transmission
For wireless communication in a multipath environment, parameters were set to create the best-case scenario and avoid
excessive time delay spread is known to cause bit error as interference exaggeration caused between 802.15.4 and
inter-symbol interference. Because a considerable time delay 802.11g. The 802.15.4 channels and center frequencies are
spread cannot be neglected, transmitted signals will suffer listed in Table II.
from frequency selective fading. Inter-symbol interference is,
therefore, the dominant factor causing irreducible bit error rate 2) 802.11g/n Wireless Network
[53]. However, if the transmitter symbol rate is lower than the An interfering network was implemented using 802.11g
coherent bandwidth, the adverse effect of channel time delay network developmental boards from ADI Pronghorn Metro
spread on the received signal can be neglected. In this case, SBC and 802.11g cards from Ubiquiti SR2. The 802.11n cards
multipath propagation causes only transmitted signal fading, were Ubiquiti SR71. The spectrum profiles of the 802.11g and
and Gaussian noise becomes the dominant factor causing bit 802.11n transmitters operating on channel 11 are shown in
error [54]. Literature has been published to quantitatively Figures 2 and 3. Wireless nodes were fully programmable in
determine the ratio at which time delay spread can be transmission power, radio channel, and modulation, among

475
other factors. Nodes were also able to execute programming TABLE II
scripts, allowing complete control of the testing procedure. In 802.15.4 CHANNELS AND CENTER FREQUENCIES
addition to transmitter/receiver pairs, traffic probes, spectrum
analyzers, and packet sniffers were used to monitor wireless Channel Center Frequency
communication. Power-auto-leveling algorithms were 11 2.405 GHz
disabled in the wireless nodes, and the maximum transmission 12 2.410 GHz
power was set to produce a worst-case scenario. 13 2.415 GHz
Iperf software was used to generate traffic on the 802.11g 14 2.420 GHz
wireless network and to test the network data rate. Iperf 15 2.425 GHz
ensured that a constant data stream was broadcasted over the 16 2.430 GHz
wireless network. Given a server on one node and clients on 17 2.435 GHz
either one or multiple nodes, Iperf configuration allows users 18 2.440 GHz
to exchange various traffic rates between terminals in a single 19 2.445 GHz
network. The software supports both transmission control 20 2.450 GHz
protocol (TCP) and user datagram protocol (UDP) traffic 21 2.455 GHz
exchange and gives periodic reports relative to bandwidth and 22 2.460 GHz
data sent/received on both server and client. A number of 23 2.465 GHz
connection options can be set on the server/client, including 24 2.470 GHz
buffer length, window and segment size for TCP, and buffer 25 2.475 GHz
and packet sizes for UDP. The 802.11g network utilized all 11 26 2.480 GHz
channels during testing, as shown in Table III. During TABLE III
coexistence testing, a National Instruments USRP-2921 was 802.11G/N CHANNELS AND CENTER FREQUENCIES
used to measure the duty cycle of the interfering network.
Channel Center Frequency
III. EXPERIMENTAL RESULTS
1 2.412 GHz
Frequency separation between two wireless devices is a 2 2.417 GHz
major factor in wireless coexistence. Research in this area is 3 2.422 GHz
separated into two groups: adjacent channel interference and 4 2.427 GHz
co-channel interference. For co-channel interference analysis, 5 2.432 GHz
SIR and channel occupancy parameters have been studied. 6 2.437 GHz
Experimental tests were carried out using the NLOS test setup. 7 2.442 GHz
8 2.447 GHz
A. Adjacent Channel Interference
9 2.452 GHz
Adjacent channel interference is caused by extraneous 10 2.457 GHz
power from an adjacent channel and due to either inadequate 11 2.462 GHz
filtering or improper frequency tuning. The work in [27, 49]
shows that interference is a result of spurious emissions
caused by surrounding ZigBee nodes employing Offset-
QPSK. Adjacent channel interference has also been studied
for 802.15.4 and 802.11b/g [21-31].
In the NLOS test setup, the 802.11g/n nodes transferred
data at a maximum, sustainable rate. The 802.11g/n nodes
operated on channels 1-11 to test for interference to ZigBee
nodes operating on channel 22 (2.460 GHz). The 802.11g/n
channels and frequencies are listed in Table III.
Figure 4 shows that the ZigBee packet error rate increases
as the frequency separation between the 802.15.4 and the Fig. 2. Spectrum profile of the 802.11g transmitter on channel 11.
802.11g networks decreases. PER was the highest when the
interfering network converged to the same frequency as the
ZigBee network (Channel 11). The figure also illustrates the
PER when the ZigBee receiver and transmitter were subjected
to the 802.11g interfering network. The ZigBee transmitter is
affected less by the side-lobes of the interfering network. In
this case, the ZigBee transmitter sensed when the wireless
channel was free and subsequently sent packets. This was not
the case when the transmitter was further away from the
Fig. 3. Spectrum profile of the 802.11n transmitter on channel 11.

476
interfering network. Instead, the transmitter sensed the
channel was free when it might have actually been busy.
Figure 4 also demonstrates that when CSMA in the ZigBee
transmitter was disabled, maximum PER improved to 58%.
This is in contrast to a maximum PER of 97% with CSMA
enabled. When CSMA-enabled ZigBee sensed a busy channel
and yet was unable to transmit, the ZigBee buffer overflowed,
resulting in additional packet loss. With CSMA disabled, the
ZigBee transmitter sent the packets regardless, causing an
increase in packet collisions at the ZigBee receiver located
furthest away. However, some packets were demodulated due
to the QPSK modulation used. This could suggest novel
modifications to the CSMA protocol for improving
coexistence. Our results suggest that careful analysis of the
CSMA configuration is needed to optimize network
coexistence.
Fig. 4. ZigBee receiver and transmitter subjected to 802.11g channels 1-11.
Figure 5 shows that ZigBee packet error rate for the
receiver and transmitter increases as the frequency separation
between the 802.15.4 and the 802.11n networks decreases.
However, it appears that adjacent channel interference to the
ZigBee receiver is not as severe when the interfering network
is an 802.11n as when it is 802.11g. One factor is that the side
lobes of the 802.11n spectrum are lower in power than the
spurious side lobes of 802.11g. The CSMA was again
disabled in the ZigBee transmitter when subjected to the
802.11n interfering network. It should be noted that when
CSMA was disabled, the PER was around 70%, which was
higher than the PER when subjected to 802.11g. This is
partially due to the fact that 802.11n has a 68% duty cycle,
compared to the 48% duty cycle of 802.11g.
B. Co-Channel Interference
Literature for co-channel interference categories are based Fig. 5: ZigBee receiver and transmitter subjected to 802.11n channels 1-11.
on SIR and channel occupancy. Literature discussing SIR [24-
43] mentions several factors, including the distance between
the interfering network and the wireless node under test; clear
channel assessment threshold of the network under test; and
receiver sensitivity. Literature focusing on channel occupancy
[21, 22, 32-38, 44-46] also notes that as packet size of the
wireless network under test increases, the probability of
interference increases [33-36].
In the NLOS test setup, the 802.11g/n nodes occupied
channel 11 (2.462 GHz) and the ZigBee nodes occupied
channel 22 (2.460 GHz). The 802.11g/n nodes exchanged data
at various rates, creating different channel occupancy rates.
Figure 6 shows that ZigBee PER increased as the channel
occupancy increased due to increasing throughput of the
802.11g wireless network. Throughput was increased until
PER was greater than 90%. Figure 6 also illustrates PER when Fig. 6: ZigBee receiver subjected to an increasing throughput of the 802.11g
the ZigBee receiver and transmitter were subjected to the that operates on the same channel.
802.11g interfering network. To cause a PER greater than CSMA was then disabled in the ZigBee transmitter. The
90%, the transmitter needed to be exposed to a higher 802.11g wireless network throughput was tested from 0 to 19
throughput. This is due to the fact that the ZigBee transmitter MB/s and found unreliable to transmit beyond 19 MB/s; thus,
senses when the channel is clear when it is placed near the limiting throughput testing. At 16 MB/s, ZigBee PER without
interfering network (as opposed to when furthest away). CSMA was 23%; ZigBee PER with CSMA was 96%.

477
 networks. Of note is that at higher duty cycles for 802.11n, the
ZigBee network is able to send packets with a 70% duty cycle
as opposed to when it is exposed to the 40% duty cycle of the
802.11g interfering network. In a higher wireless channel
occupancy environment, ZigBee coexistence with 802.11n is
better than with 802.11g.

IV. CONCLUSION
Reproducible NLOS test parameters were determined for
testing the wireless coexistence of 802.15.4 with 802.11g and
802.11n networks. The purpose of the test protocol was to
gain helpful information to efficiently tackle coexistence
problems between heterogeneous networks while optimizing
their deployment in real-life conditions.

REFERENCES
Fig. 7: ZigBee receiver subjected to an increasing throughput of the 802.11n [1] IEEE Recommended Practice for Local and Metropolitan Area Networks
that operates on the same channel. Coexistence of Fixed Broadband Wireless Access Systems, IEEE Standard
802.16.2-2001.
[2] IEEE Recommended Practice for Local and Metropolitan Area Networks
Coexistence of Fixed Broadband Wireless Access Systems, IEEE Standard
802.16.2-2004.
[3] IEEE Recommended Practice for Information Technology—
Telecommunications and Information Exchange Between Systems—Local
and Metropolitan Area Networks—Specific Requirements Part 15.2:
Coexistence of Wireless Personal Area Networks With Other Wireless
Devices Operating in Unlicensed Frequency Bands, IEEE Standard
802.15.2-2003.
[4] IEEE Standard for Telecommunications and Information Exchange Between
Systems—LAN/MAN Specific Requirements—Part 15: Wireless Medium
Access Control and Physical Layer Specifications for Low rate Wireless
Personal Area Networks, IEEE Standard 802.15.4-2003.
[5] IEEE Recommended Practice for the Analysis of In-Band and Adjacent band
Interference and Coexistence Between Radio Systems, IEEE Standard
1900.2-2008.
[6] http://ieee802.org/19/
[7] A. Kamerman and N. Erkocevic, Microwave oven interference on wireless
LANs operating in the 2.4 GHz ISM band, in: Proceedings of the 8th IEEE
International Symposium on Personal, Indoor and Mobile Radio
Communications, vol. 3, pp. 1221–1227, 1997.
Fig. 8: Packet error rate of ZigBee when exposed to an 802.11g and 802.11n [8] S. Shellhammer, Packet error rate of an IEEE 802.11 WLAN in the presence
interfering network compared with duty cycle. of Bluetooth, IEEE P802.15 Working Group Contribution, IEEE P802.15-
00/133r0 (May 2000).
Figure 7 shows the PER of the ZigBee receiver and [9] G. Ennis, Impact of Bluetooth on 802.11 direct sequence, IEEE P802.11
Working Group Contribution, IEEE P802.11-98/319 (September 1998).
transmitter exposed to the 802.11n interfering network [10] J. Zyren, Reliability of IEEE 802.11 WLANs in presence of Bluetooth
transmitting data from 1-54 MB/s. PER was similar in each radios, IEEE P802.11 Working Group Contribution, IEEE P802.15- 99/073r0
case when the receiver and transmitter were exposed to the (September 1999).
[11] A. Kamerman, Coexistence between Bluetooth and IEEE 802.11 CCK:
interfering network. The ZigBee receiver and transmitter had Solutions to avoid mutual interference, IEEE P802.11 Working Group
similar PER when exposed to the 802.11n interfering network, Contribution, IEEE P802.11-00/162r0 (July 2000).
[12] C. F. Chiasserini and R. R. Rao, “Performance of IEEE802.11 WLANs in
as opposed to when they were exposed to the 802.11g Bluetooth environment,” in Proc. IEEE Wireless Communications and
interfering network. Again, CSMA was disabled in the ZigBee Networking Conf., vol. 1, IL, Sept. 2000, pp. 94–99.
transmitter and it was exposed to the 802.11n interfering [13] J. Zyren, “Extension of Bluetooth and 802.11 direct sequence interference
model,” IEEE 802.11-98/378, 1998.
network. The trend that a disabled CSMA ZigBee transmitter [14] J. C. Haartsen and S. Zürbes, Bluetooth Voice and Data Performance in
will produce a lower PER was confirmed again. Of note is that 802.11 DS WLAN Environment: Ericsson Sig Publication, 1999.
[15] K. Takaya, Y. Maeda, and N. Kuwabara, “Experimental and theoretical
an 802.11n with 93% duty cycle will produce a 79% PER evaluation of interference characteristics between 2.4-GHz ISM-band
when CSMA is disabled in the ZigBee transmitter. wireless LANs,” in Proc. IEEE Int. Symp. Electromagn. Compat., Aug. 24–
Figure 8 compares the PER of the ZigBee receiver and 28, 1998, vol. 1, pp. 80–85.
[16] S. Unawong, S. Miyamoto, N. Morinaga, “Techniques to Improve the
transmitter when subjected to the 802.11g and 802.11n Performance of Wireless LAN under ISM Interference Environments,” in
interfering networks. The x-axis is the measured duty cycle of Proc. IEEE APCC/OECC, Oct. 18-22, 1999, vol 1, pp. 802-805.
each interfering network at different throughputs. For duty [17] A. Soltanian and R. E. Van Dyck, “Performance of the Bluetooth system in
fading dispersive channels and interference,” in Proc. IEEE Global
cycles less than 30%, the PER is similar. However, when the Telecommunications Conf., vol. 6. San Antonio, TX, Nov. 2001, pp. 3499–
PER of the ZigBee network is >60%, there is divergent 3503.
behaviour between the 802.11g and 802.11n interfering

478
 [18] I. Howitt, V. Mitter, and J. Guitierrez, “Empirical study for IEEE802.11 and [39] S. Y. Shin, H. S. Park, S. Choi, and W. H. Kwon, “Packet error rate analysis
Bluetooth interoperability,” in Proc. Vehicular Technology Conf., vol. 2. of ZigBee under WLAN and Bluetooth interferences,” IEEE Trans. Wireless
Rhodes, Greece, May 2001, pp. 1109–1113. Commun., vol. 6, no. 8, pp. 2825–2830, Aug. 2007.
[19] I. Howitt, “Bluetooth Performance in the Presence of 802.11b WLAN,” [40] M. Zeghdoud, P. Cordier, M. Terre, “Impact of Clear Channel Assessment
IEEE Transactions on Vehicular Technology, vol. 51, No. 6, pp. 1640-1651, Mode on the Performance of ZigBee Operating in a WiFi Environment,” in
November 2002. Proceedings IEEE Operator Assisted Community Networks 2006, Sept.
[20] I. Howitt and J. Gutierrez, “IEEE 802.15.4 Low Rate-Wireless Personal Area 2006, pp. 1-8.
Network Coexistence Issues,” Wireless Commun. Networking. WCNC 2003, [41] S. Myers, S. Megerian, S. Banerjee, M. Potkonjak, “Experimental
vol. 3, pp. 1481–1486, Mar. 2003. Investigation of IEEE 802.15.4 Transmission Power Control and Interference
[21] M. Petrova, L. Wu, P. Mahonen, and J. Riihijarvi, “Interference Minimization,” in Proceedings IEEE SECON 2007, June 2007, pp. 294-303.
measurements on performance degradation between colocated IEEE 802.11 [42] C. Sum, F. Kojima, H. Harada, “Coexistence of Homogeneous and
g/n and IEEE 802.15.4 networks,” in Proceedings of the 6th International Heterogeneous Systems for IEEE 802..15.4g Smart Utility Networks,” in
Conference on Networking (ICN ’07), Sainte-Luce, Martinique, 2007. Proceedings IEEE DySPAN 2011, May 2011, pp. 510-520.
[22] H. Khalell, C. Pastrone, F. Penna, M. Spirito, R. Garello, “Impact of Wi-Fi [43] N. Torabi, W. Wong, V. Leung, “A Robust Coexistence Scheme for IEEE
Traffic on the IEEE 802.15.4 Channels Occupation in Indoor Environments,” 802.15.4 Wireless Personal Area Networks,” in Proceedings IEEE CCNC
in Proceedings IEEE ICEAA 2009, Sept. 2009, pp. 1042-1045. 2011, Jan. 2011, pp. 1031-1035.
[23] M. Bertocco, G. Gamba, and A. Sona, “Assessment of Out-of-Channel [44] J. O. W. Chong, H. O. Y. Hwang, C. Y. Jung, and D. K. Sung, “Analysis of
Interference Effects on IEEE 802.15.4 Wireless Sensor Networks,” in Proc. throughput in a ZigBee network under the presence of WLAN interference,”
IEEE Instrumentation and Measurement Technology Conference, pp. 1712- in Proceedings of the International Symposium on Communications and
1717, May 2008. Information Technologies (ISCIT ’07), pp. 1166–1170, October 2007.
[24] A. Sikora and V. Groza, “Coexistence of IEEE 802.15.4 with other systems [45] S. Pollin, I. Tan, B. Hodge, C. Chun, A. Bahai, “Harmful Coexistence
in the 2.4 GHz-ISM-band,” in Proc. IEEE Conf. Instrum. Measure. Technol. Between 802.15.4 and 802.11: A Measurement-Based Study,” in
IMTC 2005, May 2005, vol. 3, pp. 1786–1791. Proceedings IEEE CrownCom 2008, May 2008, pp. 1-6.
[25] L. Angrisani, M. Bertocco, G. Gamba and A. Sona, “Effects of RSSI [46] B. Polepalli, W. Xie, D. Thangaraja, M. Goyal, H. Hosseini, Y. Bashir,
Impairments on IEEE 802.15.4 Wireless Devices Performance Suspectibility “Impact of IEEE 802.11n Operation on IEEE 802.15.4 Operation,” in
to Interference,” in Proc. IEEE Electromagnetic Compatibility Conference Proceedings IEEE WAINA 2009, May 2009, pp. 328-333.
EMC Europe, 2008, pp. 1-6. [47] M. Bertocco, G. Gamba, A. Sona, “Is CSMA/CA Really Efficient Against
[26] K. Subbu, I. Howitt, “Empirical Study of IEEE 802.15.4 Mutual Interference Interference in a Wireless Control System? An Experimental Answer,” in
Issues,” in Proceedings SoutheastCon 2007, March 2007, pp. 191-195. Proceedings IEEE EFTA 2008, Sept. 2008, pp. 885-892.
[27] L. Lo Bello and E. Toscano, “Coexistence Issues of Multiple Co-Located [48] B. Zhen, H. Li, S. Hara, R. Kohno, “Clear Channel Assessment in Integrated
IEEE 802.15.4/ZigBee Networks Running on Adjacent Radio Channels in Medical Environments,” EURASIP Journal on Wireless Communications
Industrial Environments,” IEEE Transactions on Industrial Informatics, vol. and Networking, 2008.
5, pp. 157–167, May 2009. [49] E. Toscano and L. Lo Bello, “Cross-channel interference in IEEE 802.15.4
[28] S. Han, S. Lee, S. Lee, Y. Kim, “Coexistence Performance Evaluation of networks,” Proc. IEEE WFCS 2008, May 2008, pp. 139–148.
IEEE 802.15.4 Under IEEE 802.11b Interference in Fading Channels,” in [50] IEEE Std 802.15.4-2006, “Part 15.4: Wireless Medium Access Control
Proceedings IEEE PIMCR 2007, Sept. 2007, pp. 1-5. (MAC) and Physical Layer (PHY) Specifications for Low-Rate Wireless
[29] S. Han, S. Lee, S. Lee, Y. Kim, “Outage Probability Analysis of WPAN Personal Area Networks (LR-WPANs),” 2006.
Under Coexistence Environments in Fading Channels,” in Proceedings IEEE [51] J. R. Barry and E. A Lee, Digital Communication: Third Edition. Springer,
ICOIN 2008, Jan. 2008, pp. 1-4. 2003.
[30] M. Petrova, J. Riihijarvi, P. Mahonen, S. Labella, “Performance Study of [52] Institute of Electrical and Electronics Engineers, Inc. American national
IEEE 802.15.4 Using Measurements and Simulations,” in Proceedings IEEE standard recommended practice for on-site ad hoc test method for estimating
WCNC 2006, April 2006, pp. 487-492. radiated electromagnetic immunity of medical devices to specific radio-
[31] R. Musaloui, A. Terzis, “Minimising the Effect of WiFi Interference in frequency transmitters. (standard C63.18). Piscataway, NJ: IEEE, 1997.
802.15.4 Wireless Sensor Networks,” International Journal of Sensor [53] C. I. Chuang, “The Effects of Time Delay Spread on Portable Radio
Networks, vol. 3, pp. 43-54, 2008. Communications Channels with Digital Modulation,” IEEE JSAC, vol. SAC-
[32] M. Bertocco, G. Gamba, A. Sona, S. Vitturi, “Performance Measurements of 5, no. 5, June 1987, pp. 879–89.
CSMA/CA-Based Wireless Sensor Networks for Industrial Applications,” in [54] Y. Chen and J. C. Chuang, “The Effects of Time Delay Spread on TCM in
Proceedings of IEEE IMTC 2007, May 2007, pp. 1-6. Portable Radio Environments,” Proc. IEEE Conference on Universal
[33] L. Angrisani, M. Bertocco, D. Fortin, and A. Sona, “Experimental study of Personal Communications, Nov. 1995, pp. 133–37.
coexistence issues between IEEE 802.11b and IEEE 802.15.4 wireless [55] M. Panitz, C. Christopoulos, P. Sewell, D. Hope, J. Dawson, and A. C.
networks,” IEEE Trans. Instrum. Meas., vol. 57, pp. 1514–1523, Aug. 2008. Marvin, “Modelling Wireless Communication in Highly-Multipath Low-
[34] L. Angrisani, M. Bertocco, D. Fortin and A. Sona, “Assessing Coexistence Loss Environments," Proc. International Symposium on Electromagnetic
Problems of IEEE 802.11b and IEEE 802.15.3 Wireless Network Through Compatibility, EMC Europe 2008, pp. 709-714.
Cross-Layer Measurements,” in Proc. IEEE IMTC 2007, May 2007, pp. 1-6. [56] B. Thoma, T. Rappaport, M Kietz, “Simulation of Bit Error Performance and
[35] M. Bertocco, G. Gamba, A. Sona and F. Tramarin, “Investigating Wireless Outage Probability of π/4 DQPSK in Frequency-Selective Indoor Radio
Networks Coexistence Issues Through An Interference Aware Simulator,” in Channels Using a Measurement-Based Channel Model,” Proc. IEEE
Proc. IEEE EFTA 2008, Sept. 2008, pp. 1153-1156. Globecom, Orlando, FL, Dec. 1992, pp.1825-1829.
[36] M. Bertocco, A. Sona, F. Tramarin, “Design of Experiments for the [57] D. Hope, J. Dawson, A. Marvin, M. Panitz, C. Christopoulos, P. Sewell,
Assessment of Coexistence Between Wireless Networks,” in Proceedings of “Assessing the Performance of ZigBee in a Reverberant Environment Using
the Instrumentation and Measurement Technology Conference, May 3-6, a Mode Stirred Chamber,” Proc. IEEE EMC, Aug 2008, pp. 1-6.
2010, pp. 928-932. [58] D. Johnson, M. Hatfield, M. Slocum, T. Loughry, A. Ondrejka, R. Johnk, and
[37] W. Yuan, X. Wang, J. Linnartz, “A Coexistence Model of IEEE 802.15.4 G. Freyer, “Phase II demonstration test of the electromagnetic reverberation
and IEEE 802.11b/g,” in Proceedings IEEE Communications and Vehicle characteristics of a large transport aircraft,” Naval Surface Warefare Center,
Technology 2007, Nov. 2007, pp. 1-5. Dahlgren, Virginia, USA, Tech. Rep. NWSCDD/TR-97/84, September 1997.
[38] M. Hossian, A. Mahmood, R. Jantti, “Channel Ranking Algorithms for
Cognitive Coexistence of IEEE 802.15.4,” in Proceedings IEEE Personal,
Indoor and Mobile Radio Communications 2009, Sept. 2009, pp. 112-116.

479

View publication stats