Conference PaperPDF Available

Evaluation and optimization of robustness in the IEEE 802.15.4a standard

Authors:
  • Institute for High Perfomance microelectronics and TU-Cottbus

Abstract and Figures

In this paper, we introduce ways to improve the robustness by lowering the packet loss rates of transmissions using ultra-wideband impulse radio (UWB-IR). It has been shown that the packet loss rate caused by erroneous synchronization can be tremendously decreased compared to the preamble specified in the standard IEEE 802.15.4a by our approach. These improvements are particularly intended for low-power, low-complexity transceivers operating in environments with harsh multi path propagation and high noise levels, such as industrial control. These improvements include novel ways for receiver implementation to reduce detection errors for low-power energy detection receivers with slow sampling rates. We also introduce a modified preamble that significantly reduces packet loss caused by failed preamble synchronizations We evaluate our improvements by simulation. Our approach to receiver implementation enables receiver to achieve the same packet loss rate at a signal-to-noise ratio (SNR) 10 dB lower than traditional receiver designs.
Content may be subject to copyright.
1
Evaluation and Optimization of Robustness in the
IEEE 802.15.4a Standard
Johannes Hund, Sonom Olonbayar, Rolf Kraemer, Chris Schwingenschlögl
{johannes.hund.ext, chris.schwingenschloegl}@siemens.com, {sonom,kraemer}@ihp-microelectronics.com
Abstract—In this paper, we introduce ways to improve the
robustness by lowering the packet loss rates of transmissions
using ultra-wideband impulse radio (UWB-IR). It has been shown
that the packet loss rate caused by erroneous synchronization can
be tremendously decreased compared to the preamble specified in
the standard IEEE 802.15.4a by our approach. These improve-
ments are particularly intended for low-power, low-complexity
transceivers operating in environments with harsh multi path
propagation and high noise levels, such as industrial control.
These improvements include novel ways for receiver implemen-
tation to reduce detection errors for low-power energy detection
receivers with slow sampling rates. We also introduce a modified
preamble that significantly reduces packet loss caused by failed
preamble synchronizations We evaluate our improvements by
simulation. Our approach to receiver implementation enables
receiver to achieve the same packet loss rate at a signal-to-noise
ratio (SNR) 10 dB lower than traditional receiver designs.
Index Terms—Ultrawideband, Impulse Radio, Preamble
I. INTRODUCTION
ULTRAWIDEBAND radio communication using impulse
radio (UWB-IR) is a promising standard for low-latency
and highly robust communications. It is standardized in the
IEEE 802.15.4a standard.
The IEEE.802.15.4a standard is primarily designed for
wireless low rate and low power communications and it
can additionally support ranging. The standard is based on
the impulse radio ultra-wideband (IR-UWB) communication
where very short pulses measured in 2 ns are transmitted.
The bandwidth is very high, exceeding 500 MHz. Due to the
short pulses, IR-UWB is able to offer localization accuracy
around few cm. As a consequence, IR-UWB has potential
to enable a large number of applications. A combination of
pulse position modulation (PPM) and binary phase shift keying
(BPSK) schemes is used for mapping data bits to pulses. The
symbol duration in our case is around 1024 ns within which
two bits are transmitted. The one is contained in the position
of a burst consisting of 16 chips and the other one is carried
by the polarity of the burst. A burst with the duration of 32
ns is placed either on the first half or on he second half of a
symbol depending on the data bit. See fig. 1 as an example.
Finally, Gaussian shaped pulses of less than 2 ns duration
are up-converted to one of the center frequencies in the range
of (3.4 -9.4) GHz and emitted with the power spectral density
of -41.3 dBm/MHz.
The data signal is first channel coded with a Reed Solomon
encoder introducing a redundancy of 48 bits at the end of each
330 bits for burst error correction and detection purposes. The
half rate systematic convolutional coder with the constraint
Preamble Data
Guard
Guard
32 possible burst positions (4 chips of 2ns each)
BPPM (+1) BPPM (1)
Di
...
Figure 1. Data symbol in the IEEE 802.15.4a standard[1]
length of K=3 is followed for improving an error performance
of the system. A preamble is provided to each data packet
for a synchronisation as well as for gain setting. A preamble
symbol is constructed from a ternary sequence with the length
of 31. 15 or 63 zeros are inserted between the elements of
it and it is repeated 16, 64, 1024 or 4096 times forming a
preamble. At the end of a preamble, some symbols known as
start frame delimiter (SFD) are transmitted which indicate the
end of preamble.
PHY header (PHR) with 19 bits is introduced to inform the
receiver packet length, data rate and preamble duration and
some check bits. More detailed information about the standard
is given in [2, 3].
In this paper, we discuss methods to extend the standard
in order to improve the robustness of the reception. The
remainder of the paper is structured as follows:
In section II, we explain our proposed methods. Section III
describes the simulator we use to prove our methods, the
experiments in detail and the results are described in section
IV. In section V, we summarize our work and show future
work.
II. IMPROVEMENTS TO THE STANDARD
We propose methods to improve the robustness of commu-
nication based on the IEEE 802.15.4a standard [2], especially
on low-power receivers that work with long integration time
[4]. These methods are not only improved ways to implement
a low-power, low-complexity, non-coherent receiver based
on energy-detection, but also a proprietary addition to the
standard to improve reception for such a receiver. The target
application is an industrial automation application with short
packet lengths and demanding requirements to both latency
2
and the so-called robustness, which means that errors rarely
occur and are properly detected if they occur.
A. Burst Preamble
Analysis of industrial communication systems [5] and our
previous simulations [6] have revealed a vulnerability of the
preamble. The standard defines the preamble to consist of
two parts: the synchronization header (SHR) and the start
of frame delimiter (SFD). The purpose of the preamble is
to synchronize the receiver and to set the AGC threshold
of the receiver. It consists of one of 6 ternary sequences
with very good autocorrelation properties. These sequences
are transmitted as sequences of single pulses.
AGC setting is easier when the pulses in the preamble
have about the same amplitude as in the data section. The
data section transmits not single pulses, but bursts of several
pulses with long empty intervals between them. Therefore,
the repetition frequency of the preamble pulses, referred to
as pulse repetition frequency (PRF), must be lower than the
PRF within the bursts. Therefore, a mean PRF for the data
section is calculated, by dividing the number of pulses per
burst through the duration of a symbol.
mprf =NCP B
TSY M
(1)
To conform to regulations, pulses in the preamble must be
sent in this mean PRF. Therefore the sequence is spread by
inserting a number of zeros (L) between each symbol, using
a Kroneker product. See Figure 2 for an example preamble.
SYNC (64 symbols of 4µs each) SFD (8 symbols)
Preamble Data
Code Symbol (64 chips of 2ns)
00
00
0
0
0000 SiSi
Si
Si
Si
SiSi
Si
C0C1C30 ...
...
...
...
...
Figure 2. Preamble in the IEEE standard[1]
This parameter L is determined by equation (2). Since 15
of the 31 symbols of the ternary sequence are zeros, only the
16 nonzero values are counted as pulses. By simplifying the
equation, its obvious that the preamble spreading factor L is
only dependent on the number of hopping positions used in
the data modulation.
L=1
mprf ·Tchip
·16
31 1=2·NHops 1(2)
The spread sequence is referred to as the preamble symbol.
This symbol is repeated several times (16,64,1024 or 4096)
to make it possible for the receiver to synchronize to the
transmission. The following SFD consists of a length of 8
ternary code, spread by preamble symbols.
A larger number of preamble repetitions enables a more
robust communication, as simulation results confirm that far
more packets are lost through a faulty synchronization than
through bit errors in the data field. However, the preamble
causes a large overhead, increasing the transmission time of
each packet and therefore the latency of the communication
system.
An energy detection receiver will synchronize to the pream-
ble by integrating over a certain time interval and then
correlate the resulting energy level with a generated reference
preamble symbol. In noisy environments, a receiver cannot
differentiate between a single pulse and noise, especially
when using longer integration times. Multipath propagation
makes the situation even worse. We therefore propose to send
bursts of pulses in the preamble instead of single pulses. This
increases the energy collected by the integrator and therefore
increases the processing gain. However, to keep the mean PRF,
the pauses between bursts must also be enlarged by the same
factor. We will hereafter call this variable pburst. Figure 3
depicts the difference between a preamble transmission using
the standard (1) and our proposed burst preamble (2) with a
pburst of 4.
Figure 3. Example of burst preamble with factor 4
For each code symbol in the ternary preamble code, 4 pulses
are sent instead of one and the L-spreading is also multiplied
by 4. This increases preamble recognition and synchronization
performance and significantly reduces the packet losses caused
due to synchronization errors.
On the other hand, each preamble symbol’s duration is
increased by factor 4, which increases latency. But due to the
improved preamble recognition, smaller numbers of preamble
repetitions are needed for successful synchronization. How-
ever, the duration of the SFD will be multiplied by pburst, as it
consists of 8 preamble symbols. This additional latency could
be balanced by using shorter SFD codes or fewer repetitions
of the preamble than the standard provides.
B. Sign recovery
An UWB-IR energy detection receiver for IEEE 802.15.4a
usually uses a gilbert cell in the analog frontend. This element
squares the voltage and thus amplifies pulses (rsp. samples
3
containing pulses) above the noise floor. The downside is that
the sign (polarity) of the voltage is lost. Some pulse shapes
have an integral not equal to zero, such as root-raised cosine
pulses or Gaussian doublets. For these pulse shapes, the phase
of the pulse can be determined by the sign of the integral
of that pulse. Therefore the polarity of a sample containing
a pulse or pulse burst indicates the phase of the contained
pulse(s). The ternary preamble code mentioned above has
very good autocorrelation properties. These properties degrade
if the sign information is not recovered. Summing up, we
can see that in energy-detection receivers using gilbert cells,
the polarity and thus the sign of preamble code samples is
lost, which implies a loss of a loss of 20% of the preamble
information. This reduces the quality of correlation results and
therefore reduces successful synchronizations. Our method is
to preserve the sign by adding a conditional negator after the
ADC, which inverts the ADC value according to the polarity
of the sample before the gilbert cell. This can preserve the
sign to improve preamble recognition.
C. Data
In the data section, the PHY header (PHR) and payload
bits are transferred using a dual modulation with binary
phase shift keying (BPSK) and burst position modulation
(BPM). The BPSK modulation can only be demodulated by
a coherent receiver, it is used to transfer the parity bits of an
outer systematic convolutional code. Non-coherent receivers
can demodulate only the BPM bit, it is used for the data
with an inner Reed-Salomon(55,63) code. Traditional BPM
demodulation integrates samples to the length of a burst. Then
the energy level that correspond the current hopping position
of the time hopping sequence in the first and second half of
the symbol are compared. If the sum of the samples in the
hopping position of the first half of the symbol is larger, the
symbol is interpreted as a "1", otherwise as a "0". The method
we propose does not only sum up the samples corresponding to
the hopping position, but also the two adjacent samples to the
burst. This will collect more energy for each hopping position
and will achieve suboptimal synchronized transmissions. The
two adjacent samples do not count fully, but bit-shifted so only
the half or quarter value will be added. Otherwise multi-user
interference and noise will start to distort the reception. This
can improve reception if the preamble synchronization is not
perfect. This method also improves reception in the data part
if the channel impulse response is rather long, as the energy
of one pulse will exceed the length of one hopping position.
III. SIMULATION
We implemented a simulator using MATLAB. This sim-
ulator implements a transmitter, receiver and channel of a
transmission using IEEE 802.15.4a. It is highly configurable
though and does permit modes not foreseen by the standard.
The configuration parameters include data length, channel im-
pulse response model number [7], preamble repetitions, chips
per burst, receiver sampling frequency, number of hopping
positions and AWGN with a variable SNR. We extended this
simulator so we could simulate the above mentioned additions.
The simulations were then done with a standard-compatible
mode (16 chips per burst, 8 hopping positions, 62.4 MHz
sampling rate receiver with an integrating circuit over 16 ns)
IV. RESULTS
Our simulations proved that our methods improve robust-
ness . We simulated three different cases, explained below, to
show the individual effect of each modification.
A. Burst preambles
Burst preambles were implemented as an additional pa-
rameter, 1000 test runs with the values 1,2 and 4 for pburst
(additional pulses in the preamble) were carried out in an
industrial LOS channel (IEEE channel model 7) with S/N
ratios from -5 to -15 dB. The basic preamble symbol repetition
was 64 times. For simulations using pburst, we divided the
number of repetitions by the value of pburst. So the preamble
with a pburst of 4 consisted of 16 repetitions of the prolonged
preamble.
−15 −14 −13 −12 −11 −10 −9 −8 −7 −6 −5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
SNR [dB]
loss rate
Preamble burst length
1
2
4
Figure 4. Packet losses with different length of preamble bursts
Figure 4 shows the resulting packet loss rate that was caused
by failed preamble synchronizations. With a higher preamble
burst, the packet error rate is very low, approaching to zero.
Figure 4 suggests that the packet loss rate is reduced to
almost 0% for pburst=2 compared to 60% for pburst=1 (at
a SNR of -10 dB). Pburst=4 nearly achieves 100% successful
synchronization even for SNR=-15 dB.
B. Sign recovery
The ability of the receiver to recover the sign was imple-
mented optionally and 1000 test runs were made with and
without sign recovery. We used 64 preamble repetitions and
counted packet losses due to failed preamble synchronizations.
The results in figure 5 show that preamble code sign
recovery offers the same loss rate at 9 dB less SNR compared
to the receiver with no sign recovery.
4
−12 −10 −8 −6 −4 −2 02468
0
10
20
30
40
50
60
70
80
90
100
SNR [dB]
loss rate
failed preamble recognitions
sign−aware
sign−agnostic
9dB
Figure 5. Packet losses with and without sign recovery
C. Imperfect synchronization
Overlapping samples were also implemented optionally and
1000 test runs were performed with 41 byte packets. To
negate the effect of failed preamble synchronizations, the
test runs were made with perfect synchronization with an
artificial jitter in a uniform distribution of {−2,2}samples
on the synchronization. This simulates the case where the
receiver manages to detect the preamble but only synchronizes
inaccurately. Figure 6 shows the lowered bit-error rate for
−20 −15 −10 −5 0 5 10
10−5
10−4
10−3
10−2
10−1
SNR [dB]
BER
with overlap
without overlap
Figure 6. Bit error rates with and without overlapping samples
imperfect synchronization when using overlapping samples
compared to traditional demodulation.
V. C ONCLUSION
In this paper we introduced additions to the standard to
improve the robustness of IEEE 802.15.4a transmissions with
a low-power, low-complexity energy detection receiver. Two
of these approaches (overlapping samples and preamble code
sign restoration) can be satisfied implementing a receiver
properly and therefore also improve receptions of standard-
conform transmissions. The third improvement is a modifi-
cation of the standard preamble transmission that improves
preamble recognition for low-power energy-detection receivers
with long integration times. Our simulation showed that these
improvements significantly enhance preamble synchronization,
even for channels with heavy multi path propagation and low
SNR. Moreover, even in the case of imperfect synchronization,
bit errors are reduced. Future work might include cross-
layer methods to determine at the sender whether to use
burst preambles or traditional preambles for a transmission,
according to the receiver architecture.
REFERENCES
[1] M. Flury, R. Merz, J. Y. L. Boudec, and J. Zory, “Per-
formance evaluation of an IEEE 802.15.4a physical layer
with energy detection and multi-user interference,” in
Ultra-Wideband, 2007. ICUWB 2007. IEEE International
Conference on, Singapore„ Sep. 2007, pp. 663–668.
[2] (2007) IEEE std 802.15.4a-2007 (amendment to
IEEE std 802.15.4-2006). [Online]. Available:
http://standards.ieee.org/getieee802/download/802.15.4a-
2007.pdf
[3] L. D. Nardis and M. G. D. Benedetto, “Overview of the
IEEE 802.15.4/4a standards for low data rate wireless
personal data networks,” in Positioning, Navigation and
Communication, 2007. WPNC ’07. 4th Workshop on,
Hannover„ Mar. 2007, pp. 285–289.
[4] S. Olonbayar, D. Kreiser, and R. Kraemer, “Performance
and design of IR-UWB transceiver baseband for wireless
sensors,” in Ultra-Wideband, 2009. ICUWB 2009. IEEE
International Conference on, Vancouver, BC, Sep. 2009,
pp. 809–813.
[5] A. Vedral, “Digitale Analyse, Leistungsbewertung und
generative Modellierung von WPAN-Verbindungen unter
industriellen Ausbreitungsbedingungen,” Dr.-Ing, Fakultät
für Mathematik, Naturwissenschaften und Informatik
der Brandenburgischen Technischen Universität Cottbus,
2008.
[6] S. Olonbayar, G. Fischer, and R. Kraemer, “Synchro-
nisation performance of wireless sensor networks,” in
Ultra-Wideband, 2008. ICUWB 2008. IEEE International
Conference on, vol. 2, Hannover, Sep. 2008, pp. 59–62.
[7] A. F. Molisch, K. Balakrishnan, D. Cassioli, C.-C. Chong,
S. Emami, A. Fort, J. Karedal, J. Kunisch, H. Schantz,
U. Schuster, and K. Siwiak. (2005, April) IEEE 802.15.4a
channel model - final report. online. [Online]. Avail-
able: http://grouper.ieee.org/groups/802/15/pub/04/15-04-
0662-02-004a-channel-model-final-report-r1.pdf
Conference Paper
Ultra-Wideband Impulse Radio is a promising technology for industrial automation applications due to its inherent multipath robustness and coexistence features. The ultra-wideband impulse radio PHY standardized in IEEE 802.15.4, however, is optimized for low duty cycles and not optimized for real-time communication. Especially the long preambles that are needed for synchronization cause a large overhead that might not be necessary in systems with a high duty cycle. We propose an efficient synchronization scheme, which is a cross-layer improvement and which combines this PHY layer with a MAC layer for wireless real-time communication in low latency deterministic networks based on draft IEEE 802.15.4e. This scheme splits synchronization into a network-wide frame synchronization via a broadcast beacon with a standard preamble and bit-level synchronization using just very short preambles per data frame. To achieve this, we exploit the short communication cycles and the centralized communication flow common in factory automation networks. We present and discuss simulation results of our proposed scheme that verify its higher efficiency.
Conference Paper
Using UWB for wireless short range, low rate communication is attracting growing interest due to its low power consumption and very high bandwidth. Regarding these qualities, it is comprehensible that automation industry's interest to use ultra wideband based on wireless sensor nodes increases. Those shall replace wired systems in order to save energy and other costs without reducing the reliability of the whole system. For this reason, optimizations of the standard IEEE.802.15.4a were investigated in this paper with particular emphasis on latency, data rate, energy efficiency, and reliability. This paper demonstrates the optimization potential of the official standard IEEE.802.15.4a. Different methods for optimization are investigated and combined.
Conference Paper
Using UWB for wireless short range, low rate communication is attracting growing interest due to its low power consumption and very high bandwidth. Moreover, it is able to offer accurate localization in the range of few centimeters. Regarding these qualities, it is desirable to design an IR-UWB transceiver which consumes as little power as possible when it is applied for battery driven wireless sensors. For this reason an IRUWB transceiver based on the standard IEEE.802.15.4a was investigated in this paper with particular emphasis on energy efficiency, data rate, latency, synchronization and preamble detection. This paper demonstrates the optimization potential of the official standard IEEE.802.15.4a. Different methods for optimization are investigated and combined.
Article
Using UWB for wireless short range, low rate communication is attracting growing interest due to its low power consumption and very high bandwidth. Moreover, it is able to offer accurate localization in the range of few centimeters. Considering these qualities, it is desirable to design an IR-UWB transceiver which consumes as small power as possible when it is applied for battery driven wireless sensors. For this reason, an IR-UWB transceiver based on the standard IEEE.802.15.4a was investigated in this paper with particular emphasis on its implementation, power consumption and area need for both FPGA and ASIC solutions. This paper presents results obtained with real hardware.
Article
Ultra-Wideband Impulse Radio is a promising technology for industrial automation applications because of its inherent multipath robustness and coexistence features. In our efforts to deploy UWB-IR for industrial automation, we use an adapted version of the IEEE 802.15.4a PHY. In this paper, we present a way to improve robustness for low-latency sensor-actor networks with short cyclic packets of process data, as found in industrial control applications. Due to the low cycle times of the wireless network, which can be shorter than the application cycle, soft-bit combinations across multiple repetitions of a packet can be used to collect enough information to recover packets that would not be decodable by a single transmission. To determine if a packet is a repetition or contains different data than the last transmission, a change indicator is inserted into the packet by the sender. The findings are proven by simulation against IEEE industrial channel models and show a significant improvement.
Article
Full-text available
This is a discussion document for the IEEE document of the IEEE 802.15.4a channel modeling subgroup. It provides models for the following frequency ranges and environments: for UWB channels dovering the frequency range from 2 to 10 GHz, it covers indoor residential, indoor office, industrial, outdoor, and open outdoor environments (usually with a distinction between LOS and NLOS properties). For the frequency range from 2 to 6 GHz, it gives a model for body area networks. For the frequency range from 100 to 900 MHz, it gives a model for indoor office-type environments. Finally, for a 1MHz carrier frequency, a narrowband model is given. The document also provides MATLAB programs and numerical values for 100 impulse response realizations in each environment.
Conference Paper
Full-text available
The IEEE 802.15.4 standard provides a framework for low data rate communications systems, typically sensor networks. The 802.15.4a revision introduces new options for the physical layer, in order to support higher data rates and accurate ranging capability, enabling new applications based on information on distance and positions of the devices in the network. In this paper the differences among physical layers of 802.15.4 vs. 802.15.4a are briefly reviewed. Next, the MAC layer of 802.15.4 vs. 802.15.4a are reviewed and compared. Device functionalities, network topologies as well as access strategies in both standards are described, and the impact of the new physical layer features on MAC and higher layers are discussed, with particular attention to the ranging scheme adopted in the new revision of the standard.
Conference Paper
Full-text available
We evaluate the performance of an IEEE 802.15.4a ultra-wide band (UWB) physical layer, with an energy-detection receiver, in the presence of multi-user interference (MUI). A complete packet based system is considered. We take into account packet detection and timing acquisition, the estimation of the power delay profile of the channel, and the recovery of the encoded payload. Energy detectors are known to have a low implementation complexity and to allow for avoiding the complex channel estimation needed by a Rake receiver. However, our results show that MUI severely degrades the performance of the energy detection receiver, even at low traffic rate. We demonstrate that using an IEEE 802.15.4a compliant energy detection receiver significantly diminishes one of the most appealing benefits of UWB, namely its robustness to MUI and thus the possibility to allow parallel transmissions. We further find that timing acquisition and data decoding both equally suffer from MUI.
Conference Paper
Using UWB for wireless short range, low rate communication has been attracting growing interest due to its low power consumption and very high bandwidth. Moreover, it is able to offer accurate localization in the range of few centimeters. Recognising these qualities, it is desirable to design IR-UWB transceiver which can draw minimum possible power when it is applied for battery driven wireless sensors. For this reason, in this paper, baseband design and performance of IR-UWB transceiver based on the standard IEEE.802.15.4a was investigated with particular emphasis on reducing power consumption. Synchronization algorithm that achieves 16 ns is presented and its performance is very promising offering nearly 100% synchronization for as low SNR as 8 dB. Different resolutions of analogue to digital converter (ADC) are investigated to find out the optimum with respect to power consumption and performance. 4 bit ADC was found to be the most optimal for the sample rate of 62.4 MHz. BER performance of pulse position modulation was evaluated under realistic channel conditions with multipath components.
Conference Paper
Ultra Wideband (UWB) has been gaining growing research and industry interests in many different areas of wireless communications. Among them using UWB as a radio interface for wireless sensor network is of high interest due to its precise localisation and low power consuming capabilities. In this paper we carry out quantitative analysis into the synchronisation performance of UWB when it operates under usual wireless channel environments. Simulation results confirm that coherent receiver needs around 10 dB less signal to noise ratio to achieve synchronisation compared to the noncoherent detection. Higher spreading factors lead to noticeable improvement for both coherent and noncoherent detection.
Digitale Analyse, Leistungsbewertung und generative Modellierung von WPAN-Verbindungen unter industriellen Ausbreitungsbedingungen
  • A Vedral
A. Vedral, "Digitale Analyse, Leistungsbewertung und generative Modellierung von WPAN-Verbindungen unter industriellen Ausbreitungsbedingungen," Dr.-Ing, Fakultät für Mathematik, Naturwissenschaften und Informatik der Brandenburgischen Technischen Universität Cottbus, 2008.
Available: http://grouper.ieee.org/groups
  • A F Molisch
  • K Balakrishnan
  • D Cassioli
  • C.-C Chong
  • S Emami
  • A Fort
  • J Karedal
  • J Kunisch
  • H Schantz
  • U Schuster
  • K Siwiak
A. F. Molisch, K. Balakrishnan, D. Cassioli, C.-C. Chong, S. Emami, A. Fort, J. Karedal, J. Kunisch, H. Schantz, U. Schuster, and K. Siwiak. (2005, April) IEEE 802.15.4a channel model -final report. online. [Online]. Available: http://grouper.ieee.org/groups/802/15/pub/04/15-04- 0662-02-004a-channel-model-final-report-r1.pdf
Performance evaluation of an IEEE 802.15.4a physical layer with energy detection and multi-user interference
  • M Flury
  • R Merz
  • J Y L Boudec
  • J Zory
M. Flury, R. Merz, J. Y. L. Boudec, and J. Zory, "Performance evaluation of an IEEE 802.15.4a physical layer with energy detection and multi-user interference," in Ultra-Wideband, 2007. ICUWB 2007. IEEE International Conference on, Singapore" Sep. 2007, pp. 663-668. [2] (2007) IEEE std 802.15.4a-2007 (amendment to IEEE std 802.15.4-2006).